|
| Greg's diary |
| November 2019 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Friday, 1 November 2019 | Dereel | Images for 1 November 2019 |
| Top of page | ||
| next day | ||
| last day |
|
More garden work
|
Topic: gardening, opinion | Link here |
Mick Solly, the gardener, along this morning to tidy up the garden. He got round to trimming about half the edges of the lawn, and spent the last hour planting the lime tree we bought last week. There were problems with the wheelbarrow, and he had to remove a Valerian bush. Here the lime:
Somehow I have difficulty understanding why it took too long. But everybody tells me that I'm too impatient. And Mick won't be back for 2 weeks. Clearly I'm going to have to do something myself. Planted one more tomato, a black cherry tomato I think:
|
||||||||||||||||||||||||
I'm still concerned about the wind, but it's clear that I'm going to have to do something with the rest. Round the arches maybe?
|
||||||||||||||||||||||||
That's all we have to show for two years of planting. On the left of each arch are some sweet peas, which look better this year than previously. And the jasmine on the right is also acceptable for something planted this year (which it wasn't). That leaves the right side of the left arch. Tomatoes can't do much worse than what I have there now.
|
More wildflowers
|
Topic: gardening, opinion | Link here |
Not all flowers that grow wild are welcome. I've already noted the Disa bracteata (South African weed orchid), of which I got more photos today:
|
||||||||||||||||||||||||
And then there are the wild Watsonias, now also in flower:
|
||||||||||||||||||||||||
On the other hand, there are still more that I don't recognize. A couple of days I mistook this one for a Thelymitra pauciflora, though clearly it's not:
|
||||||||||||||||||||||||
The shakiness is indicative of the wind we had today. Hopefully I'll get better ones later, but who knows if they haven't already stopped blooming. I have little hope of seeing another Thelymitra rubra flower this year.
And then there's this one, which I should be able to identify some time:
|
||||||||||||||||||||||||
|
Mail pain, continued
|
Topic: technology, opinion | Link here |
There are really two parts of setting up email on the new server: sending mail (using Postfix (software) and receiving it (using not only Postfix, but also some POP3 client).
Receiving is critical because it's very easy to lose mail if there's some configuration issue. And maybe I should try using eureka as the lowest-order MX. If something goes wrong, I can redirect the MX elsewhere. In the meantime, though, how about getting POP3 working? I'm using fetchmail on eureka and qpopper on the server.
OK, try that. No joy:
=== grog@eureka (/dev/pts/50) ~/public_html/net 18 -> fetchmail -p pop3
fetchmail: Connection errors for this poll:
name 0: connection to www.lemis.com:pop3s [45.32.70.18/995] failed: Connection refused.
POP3 connection to www.lemis.com failed: Connection refused
fetchmail: Query status=2 (SOCKET)
fetchmail: Server certificate verification error: unable to get local issuer certificate
fetchmail: Broken certification chain at: /C=AU/ST=VIC/L=Dereel/O=LEMIS (SA) Pty Ltd/OU=Certificate Authority/CN=www.lemis.com/emailAddress=groggyhimself@lemis.com
fetchmail: This could mean that the server did not provide the intermediate CA's certificate(s), which is nothing fetchmail could do anything about. For details, please see the README.SSL-SERVER document that ships with fetchmail.
fetchmail: This could mean that the root CA's signing certificate is not in the trusted CA certificate location, or that c_rehash needs to be run on the certificate directory. For details, please see the documentation of --sslcertpath and --sslcertfile in the manual page.
fetchmail: Server certificate verification error: certificate not trusted
fetchmail: Server certificate verification error: unable to verify the first certificate
fetchmail: Warning: the connection is insecure, continuing anyways. (Better use --sslcertck!)
5 messages (5 seen) for groggyhimself at oldwww.lemis.com (822984 octets).
fetchmail: No mail for groggyhimself@gmail.com at pop.gmail.com
Isn't that clear? I can't even decide which messages relate to which server, beyond the fact that the last two clearly don't.
But inetd.conf is set correctly, I think. Oh, we need a configuration file? OK. On oldwww it's 717 lines long! But it's nearly all comments and empty lines. A quick grep shows that it's really only 2 lines:
set tls-private-key-file = /etc/mail/certs/pop3key.pem
set tls-server-cert-file = /etc/mail/certs/pop3cert.pem
The lack of those entries certainly explains some of the messages. OK, what's in /etc/mail/certs? None of above. OK, copy the certificate and key across from oldwww. Still no joy. Somehow all of this security stuff seems designed to cause maximum pain.
Back to look at when I set this stuff up, 10 years ago. I wrote a lot about it at the time, but not the information I need now. On the other hand, I did note a suggestion from Mads Martin Jørgensen to use dovecot instead of qpopper. At the time I didn't bother, because I had it working. Should I now investigate that? A line in the Wikipedia article suggests yes:
Qpopper is no longer maintained. The final version was 4.1.0 released in 2011.
|
Another grid power failure
|
Topic: Stones Road house, general | Link here |
Another potential grid power failure today at 14:59:54 or so. Was it real? It certainly wasn't longer than one second.
| Saturday, 2 November 2019 | Dereel | Images for 2 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Tracking Piccola
|
Topic: animals, opinion | Link here |
Our induction cooker has a polished dark glass surface, just what you need to show every possible mark. Today we found marks that we didn't expect:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At least it's unlikely that she'll get hurt on the cooktop.
|
Disa bracteata from close up
|
Topic: gardening, photography, opinion | Link here |
The Disa bracteata (South African Weed Orchid) in Stones Road is now flowering. High time to remove it before it spreads seeds. And, of course, an opportunity to take photos of the flowers themselves.
That's a challenge in itself: they're about 4 mm across. And I've decided that focus stacking and flash don't mix. Still, this is a completely static view, so long exposures should be OK. Here a couple, taken with the M.Zuiko Digital ED 60 mm f/2.8 Macro and 26 mm of extension tubes, for a frame size of 9 x 12 mm:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But what about diffraction? I really need to read up more about that. Today I took most of my photos at f/8, but I also tried two at f/11 and f/22. Here a comparison (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The most obvious difference is the prominence of the black spots in the background. Where do they come from? I had decided that they must be dust on the front of the lens, but at f/22 they're so sharp that I'm wondering if it isn't on the sensor. But after some examination, yes, there's dust on the surface of the front element of the lens.
|
Depth of field and focal length
|
Topic: photography, opinion | Link here |
I've known the factors that affect depth of field for decades:
Sometimes (all the time, if you ask a “full frame” afficionado) shallow depth of field is a Good Thing. That's why I'm considering buying a lens with f/0.95 aperture (apart from being ahead of the Nikon crowd). It seems that Voigtländer (a confusing way of spelling “Cosina”) are the best choice for the Micro Four Thirds system. But they make 4 different lenses of that aperture, with focal lengths 10.5 mm, 17 mm, 25 mm and 42.5 mm. Which should I choose? In principle it would make sense to take the 25 mm (“standard”) lens. But clearly I would get less depth of field with the 42.5 mm lens.
Or would I? There are other aspects in play here too: with a longer focal length lens at a specific distance, the angle of view is narrower. To get the same field of view, you need to go further back. (1) giveth, (2) taketh away.
Who wins? It would be nice to think that they cancel each other out. After all, it's plain triangulation. But before I spend a lot of time thinking it through, how about putting it through my depth of field program? So I did that and got:
=== grog@eureka (/dev/pts/48) ~/src/photography 107 -> ./dof 25 0.95 5
Lens focal length: 25 mm
Subject Focal plane Magnification Exposure Near Far Depth of
distance (m) distance (mm) comp (EV) limit (m) limit (m) field (m)
5.00 25.126 0.01 0.0 4.713 5.324 0.610
Field of view: 3.44 x 2.59 m=== grog@eureka (/dev/pts/48) ~/src/photography 109 -> ./dof 42.5 0.95 8.5
Lens focal length: 42.5 mm
Subject Focal plane Magnification Exposure Near Far Depth of
distance (m) distance (mm) comp (EV) limit (m) limit (m) field (m)
8.50 42.714 0.01 0.0 8.206 8.815 0.609
Field of view: 3.44 x 2.59 m
How about that, they really do cancel each other out. Focal length makes no difference at all between the depth of field at a specific aperture when taking the same field of view. Who would have thought that?
| Sunday, 3 November 2019 | Dereel | Images for 3 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye, Quora
|
Topic: technology, general, opinion | Link here |
Over the past three years I've spent quite a bit of time answering questions on Quora. Some were quite interesting and helped me learn things that I would otherwise never have considered.
But there have been problems. Their policies for downvoting are non-transparent and offensive. I grumbled about this earlier this year, and it has happened again since. Downvoting is instant. Reinstating takes an “appeal”, and that takes days. In every case so far, my replies have been reinstated, so the downvoting was clearly incorrect. But a little more even-handedness would go a long way.
In addition, over the course of time the signal to noise ratio has dropped. Quora is to blame for this: people get paid a (minimal) sum for every question that is answered. The result is a whole lot of inane questions. Yes, there's a procedure to report them, so when today I got lots of questions like Where can I clean a Canon camera?, I reported them, and mentioned it in the reply. That's painful enough, but then later all my replies were downvoted!
Quora Moderation collapsed your answer for violating a policy on Quora.
Your answer may need improvement
“May need improvement”! I effectively copied the report. Yes, of course this was probably done automatically, and may be reversed on appeal—still more work. But it makes me wonder, yet again, what I'm doing here. I'm reminded of the remarkably evocative French « de quoi je me mêle? », literally “what am I getting mixed up in?”, but in fact has an insulting undertone. Once again I think I'll stop Quora altogether.
Instead, spent most of the day catching up on other things.
|
Dovecot?
|
Topic: technology, opinion | Link here |
Now that qpopper no longer seems the way to go, started looking at setting up dovecot. That's no easier. Yes, there's a sample configuration file, but every line needs to be considered (and this time it's many more than 2), including all kinds of encryption to be done. Why? All this mail is transported in plain text via SMTP. Wouldn't it really make more sense to accept incoming SMTP on eureka? Why is this all so difficult?
The problem with SMTP on eureka is that it doesn't solve Chris Bahlo's problems. But then it doesn't need to. She has a .forward file that forwards to her Gmail account. That leaves David Yeardley, who was once on this server, and who is still collecting spam, but who hasn't accessed the account for over 5 years. And I never liked POP. I just need to sort out yet another series of configuration files.
| Monday, 4 November 2019 | Dereel | Images for 4 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
SMTP for Dereel
|
Topic: technology, opinion | Link here |
OK, I have given up the idea of POP for the moment. Let's deliver mail directly to eureka and bypass this POP stuff. This is a long article, so there's a TL;DR.
Why didn't I do this before? I did. It wasn't until I moved to Dereel and disappeared behind NAT that I had to give up direct delivery, because my external IP address keeps changing. But that's no longer the case: since I've been with Aussie Broadband, nearly 6 years, the external IP has only changed 3 or 4 times. So it makes sense to have the mail delivered directly.
OK, what do I need to change? Firewall rules, of course. Allow SMTP. What else? DNS: add an A record for mx0.lemis.com. On the old server (oldwww.lemis.com, alias mx1.lemis.com) tell it not to accept mail for lemis.com locally. In main.cf:
mydestination = $myhostname, localhost.$mydomain, localhost, narrawin.com,
- pasocentral.org, lemis.com, lagoon.lemis.com, w4.lemis.com
+ pasocentral.org
In the process, rediscovered another switch, soft_bounce:
# The soft_bounce parameter provides a limited safety net for
# testing. When soft_bounce is enabled, mail will remain queued that
# would otherwise bounce. This parameter disables locally-generated
# bounces, and prevents the SMTP server from rejecting mail permanently
# (by changing 5xx replies into 4xx replies). However, soft_bounce
# is no cure for address rewriting mistakes or mail routing mistakes.
#
soft_bounce = yes
Just what I want. And it soon proved its value:
Nov 4 01:59:40 oldwww postfix/smtpd[47327]: NOQUEUE: reject: RCPT from minnie.tuhs.org[45.79.103.53]: 454 4.7.1 <groggyhimself@lemis.com>: Relay access denied; from=<tuhs-bounces@minnie.tuhs.org> to=<groggyhimself@lemis.com> proto=ESMTP helo=<minnie.tuhs.org>
But why that? Ah, I have to say what I want to accept mail for. By default it's just for mydestination, and I had just removed lemis.com from that list. Instead:
@@ -296,7 +296,7 @@
# list this system as their primary or backup MX host. See the
# permit_mx_backup restriction description in postconf(5).
#
-#relay_domains = $mydestination
+relay_domains = $mydestination, lemis.com
OK, try again. That didn't quite work:
Nov 3 22:10:49 oldwww postfix/smtp[44859]: 78E521B72850: to=<groggyhimself@aussie-gw.lemis.com>, relay=aussie-gw.lemis.com[167.179.139.35]:25, delay=490, delays=486/0.01/0.89/3.3, dsn=4.3.5, status=deferred (host aussie-gw.lemis.com[167.179.139.35] said: 451 4.3.5 <groggyhimself@aussie-gw.lemis.com>: Recipient address rejected: Server configuration problem (in reply to RCPT TO command))
What's wrong there?
Nov 4 09:10:45 eureka postfix/smtpd[47836]: warning: hostname www.lemis.com does not resolve to address 208.86.226.86
Nov 4 09:10:45 eureka postfix/smtpd[47836]: connect from unknown[208.86.226.86]
Nov 4 09:10:46 eureka spawn[47928]: fatal: spawn_comand: execvp /usr/bin/perl: No such file or directory
Nov 4 09:10:49 eureka postfix/smtpd[47836]: warning: problem talking to server private/bld-policy: No error: 0
In passing, times on oldwww (and the other servers) are UTC, and times on eureka are local (UTC+11), so there's an 11 hour difference between the log entries.
What's that? A quick grep through the /usr/local/etc/postfix directory showed, in master.cf:
# Blacklist Daemon
bld-policy unix - n n - - spawn
user=nobody argv=/usr/bin/perl /usr/local/sbin/bld-pf_policy.pl
That's not one, but two problems, one a bug in the postfix port. FreeBSD doesn't use a /usr/bin/perl: it's /usr/local/bin/perl. But why was this being sent to a putative blacklist daemon? Something to do with the source, maybe? What's in mynetworks?
mynetworks = 192.109.197.0/24, 203.20.68.17
A bit of blame(1) shows that this line was last modified in revision 1.37 on 28 December 2006. There's nothing obvious in the commit message, but my diary shows that at the time I was having difficulty getting Apples to understand networking and DNS. 203.20.68.17 proves to be the IP address that Yana's Apple insisted on using. Clearly those days are long gone, so:
# Dereel oldwww w4
mynetworks = 192.109.197.0/24, 208.86.226.86, 45.32.70.18
OK, try again. This time, to be on the safe side, just send the mail locally:
Nov 4 09:21:26 eureka postfix/smtpd[49998]: connect from teevee.lemis.com[192.109.197.158]
Nov 4 09:21:26 eureka postfix/cleanup[50046]: 3ABB826358C: message-id=<20191103222126.32FB2897FC@teevee.lemis.com>
Nov 4 09:21:26 eureka postfix/qmgr[47278]: 3ABB826358C: from=<groggyhimself@lemis.com>, size=513, nrcpt=1 (queue active)
Nov 4 09:21:26 eureka postfix/smtp[50173]: 3ABB826358C: to=<groggyhimself@aussie-gw.lemis.com>, relay=none, delay=0.03, delays=0/0.02/0/0, dsn=5.4.6, status=bounced (mail for aussie-gw.lemis.com loops back to myself)
Huh? OK, I need to be aussie-gw.lemis.com as well. Add that to mydestination and try yet again. Finally some semblance of functionality.
Next, we have some messages hanging about on w4.lemis.com. With a correspondingly updated main.cf, I should be able to bounce them to mx0. Here unraveled to show individual messages:
Nov 4 15:48:34 eureka postfix/qmgr[83278]: AF2C126358E: from=<groggyhimself@w4.lemis.com>, size=1286, nrcpt=1 (queue active)
Nov 4 15:48:34 eureka postfix/local[18384]: AF2C126358E: to=<groggyhimself@lemis.com>, relay=local, delay=0.22, delays=0.2/0/0/0.03, dsn=2.0.0, status=sent (delivered to command: exec /usr/local/bin/procmail -t 2>>/home/grog/Mail/procmailerr || exit 75)
Nov 4 15:48:34 eureka postfix/qmgr[83278]: AF2C126358E: removed
Nov 4 15:48:34 eureka postfix/cleanup[18549]: BF05126358F: resent-message-id=<20191104044833.GB34461@w4.lemis.com>
Nov 4 15:48:34 eureka postfix/cleanup[18549]: BF05126358F: message-id=<201911040119.xA41JvdO031201@w4.lemis.com>
Nov 4 15:48:34 eureka postfix/qmgr[83278]: BF05126358F: from=<groggyhimself@w4.lemis.com>, size=1647, nrcpt=1 (queue active)
Nov 4 15:48:35 eureka postfix/local[18384]: BF05126358F: to=<groggyhimself@lemis.com>, relay=local, delay=0.23, delays=0.2/0/0/0.03, dsn=4.4.6, status=SOFTBOUNCE (mail forwarding loop for groggyhimself@lemis.com)
So the first one got through, and the other two (second not shown here) cause a mail forwarding loop? What's in the headers? Here the one that got through, then one that didn't:
From groggyhimself@w4.lemis.com Mon Nov 4 00:50:33 2019
Return-Path: <groggyhimself@w4.lemis.com>
X-Original-To: groggyhimself@w4.lemis.com
Delivered-To: groggyhimself@w4.lemis.com
Received: from w4.lemis.com (localhost [127.0.0.1])
by w4.lemis.com (Postfix) with ESMTP id 27C4F28129
for <groggyhimself@w4.lemis.com>; Mon, 4 Nov 2019 00:50:33 +0000 (UTC)
Received: (from groggyhimself@localhost)
by w4.lemis.com (8.15.2/8.15.2/Submit) id xA40oXLF030827
for grog; Mon, 4 Nov 2019 00:50:33 GMT
(envelope-from grog)
Date: Mon, 4 Nov 2019 00:50:33 GMT
From: Greg 'groggy' Lehey <groggyhimself@w4.lemis.com>
Message-Id: <201911040050.xA40oXLF030827@w4.lemis.com>
To: groggyhimself@w4.lemis.com
From www@w4.lemis.com Mon Nov 4 01:19:58 2019
Return-Path: <www@w4.lemis.com>
X-Original-To: groggyhimself@lemis.com
Delivered-To: groggyhimself@lemis.com
Received: from w4.lemis.com (localhost [127.0.0.1])
by w4.lemis.com (Postfix) with ESMTP id 5176928114
for <groggyhimself@lemis.com>; Mon, 4 Nov 2019 01:19:58 +0000 (UTC)
Received: (from www@localhost)
by w4.lemis.com (8.15.2/8.15.2/Submit) id xA41JvdO031201;
Mon, 4 Nov 2019 01:19:57 GMT
(envelope-from www)
Date: Mon, 4 Nov 2019 01:19:57 GMT
From: World Wide Web Owner <www@w4.lemis.com>
Message-Id: <201911040119.xA41JvdO031201@w4.lemis.com>
To: groggyhimself@lemis.com
The only obvious difference is that the first message was to groggyhimself@w4.lemis.com, and the second was to groggyhimself@lemis.com. But if one were to bounce, you'd expected it to be the first one.
I still can't work this out. All further messages have got through.
Next, why is all my mail still coming through oldwww.lemis.com (mx1)? It should come in directly to mx0 (the external interface that I call aussie-gw.lemis.com, and Aussie calls 167-179-139-35.a7b38b.mel.nbn.aussiebb.net. Ah, of course, firewalls. Open firewall and see what happens:
Nov 4 17:03:48 eureka postfix/smtpd[33363]: NOQUEUE: reject: RCPT from mta-83-69.sparkpostmail.com[192.174.83.69]: 451 4.3.5 <newyorktimes@lemis.com>: Recipient address rejected: Server configuration problem; from=<msprvs1=18211ZGlSGTCe=bounces-1@bounce.nytimes.com> to=<newyorktimes@lemis.com> proto=ESMTP helo=<mta-83-69.sparkpostmail.com>
Oh. newyorktimes@ is a virtual address for mail from, well, the New York Times. And I the virtual addresses (and a number of other databases) aren't up to date on eureka, to put it mildly:
=== root@eureka (/dev/pts/11) /usr/local/etc/postfix 292 -> l *.db
-rw-r--r-- 1 root wheel 65,536 2 Jan 2013 accept-to.db
-rw-r--r-- 1 root wheel 163,840 30 Dec 2012 access.db
-rw-r--r-- 1 root wheel 344,064 3 Jun 2015 client-access.db
-rw-r--r-- 1 root wheel 65,536 9 Jul 2012 helo-access.db
-rw-r--r-- 1 root wheel 65,536 30 Jun 2013 transport.db
-rw-r--r-- 1 root wheel 65,536 22 Dec 2014 virtual.db
Checking all that is another can of worms. Mañana.
So what have I established today?
In passing, it's interesting to note some of the messages I got on the way:
=== root@w4 (/dev/pts/6) /usr/local/etc/postfix 79 -> /usr/local/sbin/postfix reload
postfix: Postfix is running with backwards-compatible default settings
postfix: See http://www.postfix.org/COMPATIBILITY_README.html for details
postfix: To disable backwards compatibility use "postconf compatibility_level=2" and "postfix reload"
postfix/postfix-script: refreshing the Postfix mail system
That's on a brand new install of the port. But the backwards compatibility message is the same as what I get on eureka, where the port is over 3 years old. Why are there still old-fashioned configuration files?
|
Leonid runs free
|
Topic: animals | Link here |
Off walking down Grassy Gully Road today. Recently we've been letting Leonid run free, and it has worked well. But today he shot off as fast as he could to the first corner, not listening to our calls, and we were afraid that he would disappear round the first corner.
Well, at first he didn't listen. After about 150 m he turned around and ran back just as fast. And ended up on a leash for his pains.
On the way back, he found a bone. He really loves playing with bones, jumping all over the place. Under the circumstances I thought that he would probably not run away, so we let him loose again. More running up and down:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
And sure enough, he didn't run away, though he dropped and broke the bone, and in the end Nikolai got the largest part.
|
Wildflowers everywhere
|
Topic: gardening, opinion | Link here |
This is the time of year for wildflowers, but this year we seem to have more than usual, like here down Grassy Gully Road:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
It doesn't look like much, but normally you'd only see a couple of flowers.
|
Your Academia web site
|
Topic: technology, opinion | Link here |
I've been subscribed to Academia for some time. A number of interesting papers appear there, and the ones I see are free, though clearly they're a commercial venture. From time to time I get a mail saying “The name ‘Greg Lehey’ appears in a document”. That's interesting, of course, but every time I follow up I find that I have to pay to find out what the document is. And somehow there are too many references. Yes, I show up in various places, but (not surprisingly) mainly in computer literature, not quite Academia's focus. And I'm pretty sure that there are no false positives for my name. So so far I have resisted, in the hope maybe of finding a subscriber who can follow the references for me.
Today, though, I got a new offer: “We created a Personal Website for you, powered by your Academia Profile”. OK, what's that? In fact, they hadn't: I had to go through a few steps to choose appearance and a couple of other odds and ends, and then I was presented with:
|
|||||||||||||||||||||||
Not very interesting. The most interesting part was at top right:
|
|
|||||||||
I get the feeling that Academia is a bona fide web site, but this kind of money-grabbing attempt seems to detract from their value.
|
Echium plantagineum!
|
Topic: gardening, opinion | Link here |
Ten days ago I noted a plant in the garden that I couldn't identify:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Yvonne posted it on Facebook, and Margaret Swan recognized it immediately: Echium plantagineum, better known as “Salvation Jane” or “Patterson's Curse”, one of the most noxious weeds I know. I think it's the first I've seen since moving to Dereel. And of course, it had to come out immediately before it seeded:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
A pity it's a weed. It's quite pretty, and only recently Yvonne was mourning the death of our last Echium.
| Tuesday, 5 November 2019 | Dereel | Images for 5 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dog visit
|
Topic: animals, opinion | Link here |
Kim Stanley along today along with her old greyhound mix dog Davy (or is that Davey or Dave?).
It proved to be Davey.
Since her other dog died, Davy has been lonely when she was away, and she's looking for company for him, and when Yvonne offered our vicious hunting dogs, she jumped at the offer. All went well until they wanted to play together. In the 8 seconds between these two photos Davy felt threatened, and we had to pull them apart:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At least we have the photos. It seems that Davy started things, and ours went along. But they calmed down again, and we brought him inside, where he behaved remarkably well.
|
|||||||||||||||||||||||
So he'll be back on Saturday for an hour without Kim.
|
Another grid power failure
|
Topic: Stones Road house, general | Link here |
Another grid power failure today at 12:00:45, this time 4 seconds and thus long enough to be genuine.
|
Military equipment named after Trump
|
Topic: politics, language, opinion | Link here |
The Register reports a more than usually interesting US Air Force contract award today:
US Air Force inks deal with Raytheon on Windows 10 (and other) support for ARSE
This reminds me of the Rolls-Royce projected car “Silver Mist”, renamed “Silver Shadow” after the marketing team discovered what „Mist“ means in German. Clearly the US military are not linguists.
Other sites had a field day, like US Air Force inks deal with Raytheon to insert Windows 10 into ARSE. But maybe it's fake news? No. It took me a while to find a plausible web site, but those in the .gov. domain seem to fit the bill. Advanced Medium Range Air-to-Air Missile Remote Support Equipment (ARSE) states:
Advanced Medium Range Air-to-Air Missile Remote Support Equipment (ARSE)Solicitation Number: FA8675-16-C-0067_P00036Agency: Department of the Air ForceOffice: Air Force Materiel CommandLocation: AFLCMC/EBAK - Eglin
Now wasn't that a nice way of recognizing the Greatest President who Ever Has Been?
|
More house construction
|
Topic: general | Link here |
The house that we have been watching in Harrisons Road is now being erected:
|
||||||||||||||||||||||||
Looks like it's being built by JG King.
| Wednesday, 6 November 2019 | Dereel | Images for 6 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
FreeBSD: Time flies when you're having fun
|
Topic: history, technology, general, opinion | Link here |
I have a daily cron job that sends me email showing the passage of my life, fraction by fraction, along with significant events that match the dates. From today's message:
5/8 life: Monday, 8 March 1993
2/3 life: Saturday, 24 February 1996
Feb 24 "Installing and running FreeBSD" submitted for publication, 1996
70% life: Wednesday, 8 July 1998
A third of my life already! How time flies.
|
Macro photos: the pain
|
Topic: photography, opinion | Link here |
My Disa bracteata (South African Weed Orchid) has survived the exhumation and being put in a pot, and is now flowering more than on Saturday. Time for some more photos, in particular further from below. This photo (from Saturday) suggests that there's more to be seen from that angle:
|
|||||||||||||||||||||||
OK, do it properly. Prepare for focus stacking, attach the external viewfinder, position. Dammit, I can't even recognize or locate the individual flowers! And the controls for the camera are in a position that I can't see, and which I certainly can't reach. In addition, the size of the camera and rail are such that I can't even position correctly. What do I do? Cut the flower stalk and place it differently? That would work, but I'm reluctant to destroy the stalk. So I did what I usually do: procrastinate.
|
xkcd: Software update
|
Topic: technology, opinion | Link here |
I'm still dragging my feet about my software update, both on eureka and on the external servers. Today xkcd brought comfort:
In particular, “An update finally breaks a feature I'm unwilling to lose” reminds me of the breakage that firefox introduced a few releases back, meaning that I can no longer use Emacs bindings.
| Thursday, 7 November 2019 | Dereel → Ballarat → Dereel | Images for 7 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Quora responds?
|
Topic: technology, general, opinion | Link here |
It's been several days since Quora “collapsed” my replies, and I decided that I couldn't be bothered with them any more. I “appealed” anyway, but didn't get any response.
My “couldn't be bothered” certainly seems to be appropriate. I still get messages asking for a personal response to questions, and I'd still be prepared to answer ones that are appropriate to my knowledge. So far, none of them appear to have been worth reading. Here are some, including the one that I think the best question:
Is it safe to unplug a gas stove?How did Hitler gain access to color video recording?Is it expensive to travel around the world?When preparing a sourdough starter, can I use a different flour such as kamut? What should I use for making bread from the starter?Are you interested in buying an original DSLR or a digital camera?Do sunflowers turn to face each other on cloudy days?What is the wealthiest city in Nauru?What is the best lens for an iPhone 11?Why do I look really bad on camera?
The good one? The one about sourdough. But I can resist that.
But one thing that Quora does is pay for questions. Not much, but in the course of many months I accrued the princely sum of $1.82, certainly helped by the fact that I never got answers to the few questions I asked.
But today I suddenly discover:
Partner Program: You've earned $0.34 this week and $2.16 all time.
Why that? The previous week I earnt nothing at all. You'd almost think that they're trying to lure me back with their $0.34.
|
Video downloads: how the other half do it
|
Topic: multimedia, technology, opinion | Link here |
Message today from Lena G. (she's too polite to mention her surname). It's strange for a number of reasons: it's in German, and it refers to a personal page /grog/notes.php?page=Notes/radio, basically a list of download links. But she calls it an article and asked me to link to her page on YouTube download alternatives. There's another issue: I don't mention YouTube downloads at all on that page.
OK, let's follow her link and find what it's all about. Two issues: how to download, and “is it legal?”. She mentions a surprising number of download programs that I have never heard of, all with some limitation. She does mention youtube-dl (which I didn't find at first because she calls it YouTube-DL), but it, too, has disadvantages: spartan “layout” and command-line interface.
Those are disadvantages? It does exactly what it claims to do, and I've never had much in the way of issues with it. The ones that I have had are because it's continually chasing web interfaces, and they're almost always fixed by the time I find them, requiring only an update. And even the port always installs cleanly.
Never mind, if you want lots of eye candy and mouse-pushing you can install YouTube-DLG. But only, it seems, if you run Microsoft.
Somehow we're living in different worlds. All the things that I find good about youtube-dl, she finds bad. And she's prepared to go through intense pain to avoid a command line.
The other part of her article is about the legality of downloading from YouTube. It seems that there are laws that govern that, though she doesn't mention the countries for which her conditions apply (I'm sure that the laws in Switzerland are different from those in Germany, and potentially the laws in Austria differ from both), and her primary reference to legality refers to a torrent site kinox.to.
She concludes that:
The green part, in English: “It is legal if you are not logged in and if the videos are only for your personal use”.
“Not logged in”? What does that mean? Again, I think we're living in different worlds.
And yes, not only do I have permission to quote that image, she specifically asked me to show it.
Of course, that was in 2019. In 2024 it was no longer to be found. The image above is a copy I made at the time.
|
Blood test results and haircut
|
Topic: health, general | Link here |
Into town today for the results of last week's blood test. How long does it take to get there? It's in the middle of Ballarat, and I've found in the past that it takes (a little) more then 30 minutes, and that only 15 of them are outside Ballarat. Today I timed myself again: 25 minutes in relatively slow traffic to the corner of Doveton St and Dana St, about 150 m from the clinic.
But first I needed to park the car in the park house across the road and walk to the clinic. That took a total of 5 minutes! And it wasn't an exception: it took just as long in the other direction. That certainly confirms my decision to have blood tests in Victoria St, where, if I'm lucky, I can park directly in front of the clinic.
Blood test results weren't alarming, but not as good as they have been. Random fluctuation? Cholesterol and blood pressure higher (137/81 IIRC). I'll have to keep an eye on that, and he wants another test in 3 months.
Also compared ear thermometers. Both of mine were surprisingly in agreement at 35.8°. Paul's professional device measured a far more plausible 36.9°. As we decided almost exactly 30 years ago, Triple Modular Redundancy determines that Paul's thermometer was wrong. As I said at the time,
This shows a) the follies of TMR with multiple failures...
On the way home, also finally had my hair cut. After over 4 months it was certainly high time.
|
More macro pain
|
Topic: photography, gardening, opinion | Link here |
More attempts at macro photos of the Disa bracteata today. In the end I gave up trying to focus and took shot-gun photos, up to 220 at a time. No time to process them, and it's clear that I took far too many. But it will have to wait until after the processing to decide whether they're any good.
| Friday, 8 November 2019 | Dereel | Images for 8 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More RCD problems
|
Topic: Stones Road house, general, opinion | Link here |
Another power issue today: the main house RCD tripped, not once, but twice.
Why? The first time seemed related to the use of a microwave oven, but the second seemed unrelated to anything. It was a wet day; could it be that that resulted in enough leakage current to trip the RCD? In any case, it wouldn't have been as bad if each circuit had had its own RCD, instead of one for everything in the house except the UPS. High time to finally fix the house wiring, but I've been trying to get Steve White for months now, without success. Where do I find another good sparkie?
|
Still more macro pain
|
Topic: photography, gardening, opinion | Link here |
Finally got round to processing the extreme macros of the past couple of days. In the end I generated 57 GB of images, mainly TIFF. The first two didn't seem too bad:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The last two images had 195 and 175 frames. How many did I need for the final images? None of them! Somehow I had set my start position incorrectly (reminder: coming closer does not make closer things sharper), and the series were useless. What do I do? I hate throwing away any image, no matter how bad, but something like this makes no sense at all:
|
|||||||||||||||||||||||
OK, start again. In the meantime it has occurred to me to use a small mirror to look up the stalk. But once again the mechanics of the matter got the better of me, and in the end I gave up.
There must be an easier way.
|
Processing subtitles
|
Topic: multimedia, technology, opinion | Link here |
We've been trying to watch an Austrian series, „Vier Frauen und ein Todesfall“. Problem: it's in Austrian, and we really need subtitles to understand the finer points. OK, it comes with a subtitle file (in fact, depending on where I get them from, two, one ending in .ttml and the other in .de-AT.ttml). But mpv wants subtitles in .srt format.
No problems, there's a program ttml2srt, which I have used a couple of years ago, not without pain. I got over that particular problem, but for a couple of weeks now I have been wrestling with this problem:
=== grog@teevee (/dev/pts/5) /spool/Series/Vier-Frauen-und-ein-Todesfall/02 6 -> python3 ttml2srt.py 02-03-Schlachtfest-14562238.de-AT.ttml > 02-03-Schlachtfest-14562238.srt
Traceback (most recent call last):
File "/home/local/bin/ttml2srt.py", line 169, in <module>
print(content)
UnicodeEncodeError: 'latin-1' codec can't encode character '\u2013' in position 31: ordinal not in range(256)
And it aborted at that point (about 7 minutes into the file). Try as I might, I couldn't get rid of that message. Until today, when I read the message more carefully: for some reason the program was converting UTF-8 to ISO 8859. Why? There are no flags for the program. Was it looking at the locale? That's arguably incorrect, but possible. Try again:
=== grog@teevee (/dev/pts/5) /spool/Series/Vier-Frauen-und-ein-Todesfall/02 7 -> LC_ALL=de_DE.UTF-8 python3 ttml2srt.py 02-03-Schlachtfest-14562238.de-AT.ttml > 02-03-Schlachtfest-14562238.srt=== grog@teevee (/dev/pts/5) /spool/Series/Vier-Frauen-und-ein-Todesfall/02 8 ->
So simple! Time to wrap it in a script. Yes, maybe I should have chosen de_AT.UTF-8, but this should work for other German-language stuff too.
| Saturday, 9 November 2019 | Dereel → Ballarat → Dereel | Images for 9 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Three hours for a cup of coffee
|
Topic: food and drink, general, opinion | Link here |
After breakfast this morning, as usual, I pressed the button on the coffee machine to make a cup of coffee. No:
|
|||||||||||||||||||||||
What does that mean? Certainly it's not an encouraging sign. Get a spanner? No, it's a way of telling illiterate coffee drinkers to look up the symbol in the instruction manual, if you can find it. About the most interesting thing in the image is the 1 at the bottom right. Off to drag out the PDF of the “instruction manual” about which I grumbled so long and loud four years ago. There I read:
|
|
|||||||||
No explanation what it might mean, just how to try to work around it. Doesn't that fill you with confidence? Try several times, it might recover. And no mention of the number at bottom right, arguably the most important part of the “message”. About the only thing that I did find was this display, immediately on powering on:
|
|||||||||||||||||||||||
It went away by itself, and the “instruction manual” has information on that too:
|
|
|||||||||
That's a reasonable thing to expect on power on. But it wasn't there before, and it seems reasonable that it is related to the red spanner.
OK, what do I do? Today's Saturday. The earliest I can contact the service people would be on Monday, and it will probably require repair. Round here, that's at least another week, probably two. I had bought a second-hand coffee machine for exactly this eventuality, but it needs ground coffee, and I don't know where it is. On the other hand, I had not expected the machine to last long. It cost me $599, because Philips refused to honour their cashback offer. Compared to the capsule machine I had before, it saved me about $0.80 a day, or about $1,140 over the life of the thing. So the coffee machine has done its work, the coffee machine can go.
And what do I buy? And where? Off to an incredibly slow Good Guys web site to discover only three comparable machines, all from De'Longhi: the “Magnifica”, the “Magnifica” S, and the “Magnifica” S, conveniently accompanied by photos taken from different angles:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
It looks as if the second and third models are the same except for the colour. What does their compare function say? No, completely different feature set. Do I believe it? No, I'm pretty sure that the comparison page is broken.
OK, who else has coffee machines? Harvey Norman? Harris Scarfe? Target? Only Harvey Norman had similar (in fact, the same) models, for $50 more.
OK, off to the Good Guys to ask them the differences. Only the first two of the models there, the second one marked $699 instead of the promised $649. Had some difficulty getting picked on by a salesperson, who checked the price on her phone, but didn't tell me what it was ($649) until I asked some time later. The differences? She called another salesperson and they explained that the more expensive one was newer and more compact: they moved access to the internals from the front to the side, a clear disadvantage. Oh no, just a matter of taste, they said: they've clearly never had to access it in real-life situations. Still, the size made a difference, so I took it.
This description doesn't make clear which model I took. It was the Magnifica S (middle photo), for the promised $649.
And the difference between the silver and black ones? They couldn't say: they had never seen a black one. I had to check when I got back home: the model numbers on the web site are ESAM04110B, ECAM22110SB and ECAM22110B. That suggests to me that the last two are really identical except for the colour.
The difference between the silver one and the one next to it for $1,699, $1,050 more? The expensive one had Bluetooth, so you can control your coffee machine with your mobile phone! Just what I need! It also froths milk automatically: it seems that my last one and the one I was looking at require you to stick the nozzle in the milk to heat it up, but the expensive one comes with a container for the milk. Otherwise they're the same, in particular the “brewing” mechanism.
Somehow there are many appliances made for people with more money than sense. For $150 more I can buy a capsule coffee machine. And there are real coffee machines for up to $3,000 (OK, they call it $2,999).
Interestingly, there were no Philips coffee machines at all. One of the salespeople told me that she hadn't seen them round for years, and a look on eBay confirmed, once I got round their current web site breakage. I wonder how easy it's going to be to get the old machine repaired.
Back home, the usual fun with the instruction manual, clearly written from the perspective of a developer rather than an end-user. But the thing works, and it makes less noise than the old one.
Three hours to make my after-breakfast coffee!
|
Smart fridges and other essentials
|
Topic: general, opinion | Link here |
While at the Good Guys I took a look round the shop for other things of interest. Not much to be seen, though I see there are “air fryers” from Philips that look very much like the el-cheapo versions from ALDI:
|
||||||||||||||||||||||||
Are they bigger? Yes, maybe a little, but they're like our small one (the “coffee machine”). And the prices are exciting, especially from a maker who doesn't seem to be in it for the long run.
Then on to fridges, where once again they have a Samsung “smart fridge” on display, this time with the display working:
|
||||||||||||||||||||||||
That shows clearly that the cameras are in the front corners. While taking this photo, a salesperson came along and we discussed the issues: the cameras just plain Do Not Work in real-life situations. Put my handbag in the fridge to make the point, and he put his pen and keys behind them:
|
||||||||||||||||||||||||
Of course, the keys and pen were not to be seen. Ah, says salesperson, you don't normally put anything that small into the fridge. Be that as it may, you do put bigger things than my bag into the fridge, and only this morning I discovered a jar of sateh sauce that had worked its way to the back and long since gone bad.
But then there's a list of what's in the fridge:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Yes, that would work, but it needs to be a little more specific than “in fridge” or “in freezer”. And how do you get that data? You type it in on the horrible glass display!
On the other hand, a wand for scanning things would help a lot. Do they have them? Of course not. Smart fridges must smart!
|
PIXIO: first impressions
|
Topic: photography, opinion | Link here |
Yvonne and Chris Bahlo back from town shortly after me, with
a package for Yvonne: the PIXIO “Robot Cameraman” that we had ordered on 24 October, with promise that the
excessive € 72 would result in delivery in 5 working days. But they reckoned
without DHL
UPS
Startrak Australia
Post. I received a notification from UPS on Wednesday that the item would be delivered
the following day, and based on my experience with UPS I
wasn't surprised when they didn't. But it's still surprising that the item should have
ended up at Napoleons post
office without any further notification; they only went there at all because Chris had
something to pick up.
So, what's it like? Even more explosion in the dining room:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
That's the sum of the PIXIO, the new coffee machine and some “smart home” stuff that I bought at ALDI in a fit of madness a couple of weeks ago, and which I still haven't tried out.
The PIXIO came with a 5-way USB charger that might be useful elsewhere, which also had the distinction of having the only documentation of any kind, a cut-to-size 7x9 cm laser printer printout. The complicated PIXIO itself arrived with no documentation whatsoever, unless you count the wrapper round the box, which barely documents anything. I've found cause to complain about poor documentation on frequent occasions, but this subpasses anything that I have seen.
OK, I had downloaded the documentation, such as it is, before buying the device, so I could confirm that it also arrived with the wrong cable. That's not a big deal: the cables only work with specified video cameras, and we don't have any. But the standard cable is also the most expensive and the most likely one for us to need, so it's annoying.
| Sunday, 10 November 2019 | Dereel | Images for 10 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PIXIO in practice
|
Topic: photography, multimedia, opinion | Link here |
Reading the instructions for the PIXIO “Robot Cameraman” was not easy. They're a little unstructured, and at one point I thought that important things were missing. But in fact it's not as difficult as the instructions make it look. There are five items: three “beacons”, the tag carried by the person to be followed, and the control unit that turns the camera.
About the most complicated thing is the position of the beacons. They're used for triangulation so that the control unit can determine the relative position of the tag, so clearly they should be as far apart as possible. The instructions include:
|
|||||||||||||||||||||||
That's clear, right? Well, left: put beacon 2 to the left of beacon 1. But the thing turns (it can, indeed, rotate freely). At some point, when it's pointing at beacon 3, beacon 2 is to the right of beacon 1. And the illustration says to place beacon 1 as far as possible from the camera, but that's not where it's shown. In this scenario, very similar to mine, it would be at the right rear corner.
Then there's this illustration:
|
|
|||||||||
What is it really saying? About the only obvious thing is the last scenario: don't place the camera exactly between two beacons. The other “no-no” suggests that you should put them further apart.
In the end I chose none of those layouts. Instead put beacon 1 on rear right (furthest from camera), beacon 2 on rear left, and beacon 3 on a fencepost halfway up the right side.
Why? I was puzzled myself until I discovered that the manual included another choice, not offered on this page. On the previous page there was:
|
|||||||||||||||||||||||
Problem: If I place them like that, I can't put the camera in the middle of the arena without being in a direct line between beacons 2 and 3. Thus my choice of putting it half way down.
OK, how do I do it?
“Start recording” is a kludge, of course. The PIXIO has the ability to remote control specific video cameras from the tag, but I haven't got my head round it yet. And other cameras that I have may be controllable with an infrared remote control. The E-PM2 understands the remote control too, but it always takes a still photo, even when set to video. That's probably not an issue: the E-PM2 doesn't have a microphone input, and we're going to need one of them, so my guess at this stage is that she'll use the Olympus OM-D E-M1 (Mark I).
Here are today's results:
The interesting thing is that the control unit does very well even when one or even two of the beacons are powered down (towards the end of the clip).
Apart from that, there are some strangenesses about the device. It's not easy to place the camera on the device. There's a very cheap connector there, something that (like the toy tripods) they could really improve on:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
YouTube login restrictions?
|
Topic: technology, opinion | Link here |
Mail from Lena G. again today, explaining that youtube-dl is far too complicated for beginners. As I said, parallel universes, but I suppose it's more complicated if you've never been allowed to be in charge of a computer rather than having it in charge of you.
And the question of being “logged in”? She agrees, it doesn't make any sense: it comes from this video clip:
Or does it? That clip (now 3 years old) doesn't mention “logged in” at all, just the question of breaking copy protection. That's clearly illegal if it can be proven, but it's a different kettle of fish.
By chance I also received mail from YouTube today, announcing new terms and conditions. Being logged in? Not mentioned. They did link to this explanation page (which I suspect will rot quickly), where there's no mention of being logged in. The closest I can come is this paragraph:
Google Accounts and YouTube Channels. We’ve provided details about which features of the Service can be accessed without a Google account or YouTube channel, and which features require one.
But that's not a question of legality: it's a question of whether you can access the content at all.
And then there's this:
You are not allowed to: ... access, reproduce, download, distribute, transmit, broadcast, display, sell, license, alter, modify or otherwise use any part of the Service or any Content except: (a) as expressly authorized by the Service; or (b) with prior written permission from YouTube and, if applicable, the respective rights holders;
But that's circular. Still no mention of “login”, but also no mention of where the express authorization is described. Given that this is a US company, my understanding is that any perceived abuse must be clearly prohibited (“due diligence”), and I don't see any such prohibition.
| Monday, 11 November 2019 | Dereel → Napoleons → Dereel | Images for 11 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PIXIO in earnest
|
Topic: multimedia, photography, animals, opinion | Link here |
OK, after yesterday's success with the PIXIO “Robot Cameraman”, Yvonne wanted to do it in earnest. Out to help, but in fact things went pretty well:
Things that need improvement are:
Still, we were expecting some learning issues. On the whole I'm quite happy.
|
Voigtländer Nokton 25 mm f/0.95
|
Topic: photography, opinion | Link here |
Today I received a used Voigtländer Nokton 25 mm f/0.95 lens, serial number 8140610. Why? GAS is one thing, and maybe the satisfaction of being able to show the Nikon users that they're not by any means the first with a standard lens with an f/0.95 aperture. And of course, for more practical reasons, the reduced depth of field, for example here:
|
|||||||||||||||||||||||
The problem is that it's a completely manual focus lens, which doesn't make things easier. The Olympus OM-D E-M1s have focus peaking, but it doesn't work very well at close distances, and in fact this image is slightly out of focus. That could be due to movement: the depth of field is round 8 mm at this distance.
Apart from that, the lens looks OK. I don't see myself using it very much: I don't even need to stop down. For that I can use the Leica Summilux 25 mm f/1.4, which has autofocus and probably better optical qualities.
Apart from that, this is the first lens that I have ever seen that writes the maximum aperture incorrectly: instead of “f/0.95” it is inscribed with “F0.95”. I fear that this dumbing down will continue across the industry.
| Tuesday, 12 November 2019 | Dereel → Ballarat → Dereel | Images for 12 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cutting corners with telephotos
|
Topic: photography, animals, opinion | Link here |
While having breakfast this morning, saw this outside the dining room window:
|
|||||||||||||||||||||||
What terrible image quality! I took them through the window of the dining room. I hadn't expected it to be excellent, but I wasn't expecting such poor quality either.
|
Dentist again
|
Topic: health, opinion | Link here |
Into Ballarat for my six-monthly dental checkup, which translated to $120 for cleaning my teeth. Nothing wrong with my teeth, but I forgot to clean them before coming, so Mario went through them with dental floss. He needed four lengths and considerable effort to complete the job, and left a strand hanging between two teeth. And that's somebody who knows what he's doing and can see where he puts the stuff. Another confirmation that interstitial brushes are a better choice for me.
|
Updating photo workflow
|
Topic: photography, technology, opinion | Link here |
I've been developing (not a pun) my photo workflow for nearly 20 years now, and I'm currently at revision 1.138 of the Makefile. A lot has clearly happened in that time.
One of the central parts is a translation table in a file called Makejpeg. Its format has changed over the years: initially it was makejpeg and contained source file name, destination file name, rotation value and gamma correction value. Since I've been using DxO PhotoLab the last two are redundant, and I've been dropping gamma for some time. And makejpeg contained only photo information; I had a second single line file description that contained the description of the photos. That made no sense, so I merged them into Makejpeg. In principle:
cat description makejpeg > Makejpeg
In the meantime, however, came the issue of scans: three years ago I added a whole lot of data to Makejpeg, so now a typical entry might be:
img225_DxO Taj-Mahal-3-cropped 0 A c 2 F 5 f 11 s 125 l 11
The additional information are: A (set me as author), c 2 (camera 2, the Asahi Pentax “Spotmatic”), F 5 (Film 5, Medium Speed Ektachrome), f 11 (f/11), s 125 (1/125 s shutter speed), l 11 (lens 11, 28 mm f/3.5 Super-Takumar).
Some of this can be useful for other photos as well. The new Voigtländer Nokton 25 mm f/0.95 doesn't supply any lens information, so I was able to add that by the same means. In the process I discovered that my kludged exposure details popup doesn't display UTF-8 characters correctly, so I had to call it Voigtlaender instead of Voigtländer.
The decision to be made: how? Though it uses the Makejpeg file, I hadn't integrated it into my workflow. On the other hand, things like setting the author need to be done for every photo, and so far I've cheated by setting my name in every image I read in, thus also modifying the original file—a no-no.
Currently the workflow is:
So: at what point should I add the additional Exif data? Step 10 is already untidy. Can I somehow combine it with creating the JPEG images? That would fall flat if I were to remove step 8, Ashampoo optimizer, and convert the files directly to JPEG. Can I at least put step 10 into the Makefile?
The good side of all this is that I have the choice. If I were used Microsoft-based off-the-shelf software, the choices would have been made for me, and I probably wouldn't like them.
| Wednesday, 13 November 2019 | Dereel | Images for 13 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
NBN: Goodwill exhausted
|
Topic: technology, general, opinion | Link here |
We've had an Internet link from the National Broadband Network for nearly 6 years now. When it was installed we were overjoyed. As Yvonne put it:
|
|
|||||||||
But that was six years ago. During that time, you'd think that the NBN had done everything to erode this positive sentiment. Today, once again, we had an extended “planned” outage, nearly 6½ hours. I suppose I should be grateful: they had promised three outages of 10 hours each.
But what on Earth are they doing? In those 6 years we've had a total outage duration of over 11 days, or about 3.8 hours per month. It seems that things are getting worse. Looking at the last 7 weeks, we have had:
Start time End time Duration Badness
1568686359 1568686374 15 # 17 September 2019 12:12:39 17 September 2019 12:12:54
1568687257 1568688388 1131 4.077 # 17 September 2019 12:27:37 17 September 2019 12:46:28
1568688877 1568692055 3178 7.362 # 17 September 2019 12:54:37 17 September 2019 13:47:35
1568692480 1568694933 2453 8.471 # 17 September 2019 13:54:40 17 September 2019 14:35:33
1568698770 1568698826 56 0.938 # 17 September 2019 15:39:30 17 September 2019 15:40:26
1568699617 1568699842 225 4.551 # 17 September 2019 15:53:37 17 September 2019 15:57:22
1569445750 1569472304 26554 0.005 # 26 September 2019 07:09:10 26 September 2019 14:31:44
1569610631 1569610712 81 0.026 # 28 September 2019 04:57:11 28 September 2019 04:58:32
1571261914 1571290166 28252 0.002 # 17 October 2019 08:38:34 17 October 2019 16:29:26
1571634521 1571639777 5256 0.010 # 21 October 2019 16:08:41 21 October 2019 17:36:17
1572696085 1572696131 46 0.003 # 2 November 2019 23:01:25 2 November 2019 23:02:11
1573574585 1573575721 1136 0.004 # 13 November 2019 03:03:05 13 November 2019 03:22:01
1573595181 1573618452 23271 0.185 # 13 November 2019 08:46:21 13 November 2019 15:14:12
Summary from 17 September 2019 to 13 November 2019
Total 13 outages, total time 91654 seconds (1 days, 01:27:34)
Longest outage: 28252 seconds (07:50:52)
Start: 17 October 2019 08:38:34
End: 17 October 2019 16:29:26
Average time between outages: 385203 seconds (4 days, 11:00:03)
Average duration: 7050 seconds (01:57:30)
Availability: 98.17%
That's round 10% of all outage time, and the three longest outages (all over 6 hours) have been in the daytime on workdays (26 September, 17 October and today, 13 November), effectively making Internet access impossible for 3 out of 34 work days. The overall availability was only 98.17%, already by far the worst I have experienced in 30 years on the Internet, but looking at work hours it's closer to 92%.
High time for that TIO complaint.
|
HDR images revisited
|
Topic: general, photography, opinion | Link here |
What do you do when you've been forcibly disconnected from the Net for a whole day? There are plenty of things; the garden needs attention, or I could finally try to tidy up my office.
What did I do? While looking for something else, found this entry in my diary: problems getting good HDR images. But now I have Photomatix PRO, and in principle I'm happy with it. How does it fare with those images? Here the images I made then, first enblend, then HDR Projects 4. Run the cursor over either image to switch to the other image:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly the HDR Projects 4 results were useless. I was able to improve on them, but I don't want to have to hand tune every image.
And with Photomatix? Much better. Omitting the useless HDR Projects result, I ended up with (first enblend, then Photomatix):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
There's not really much obvious difference at first. As usual, the devil is in the detail. I had already noted that the contrast of the enblend image was too low, and certainly the details show that. Once again run the cursor over an image to compare it with its neighbour.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But the gradation of the panorama is actually darker than the enblend version. Maybe this is really a kind of image that needs special treatment.
|
Another sick kangaroo?
|
Topic: animals | Link here |
Walking the dogs today, saw a kangaroo sitting in a paddock, about 20 m from the road:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Normally they'll get up and hop off when approached that closely. But this one just looked. He was still there when we came back 20 minutes later. Another dying kangaroo?
|
Broom: invasive?
|
Topic: gardening, opinion | Link here |
There's a broom bush on the corner of Stones Road and Progress Road, and once again it's in flower:
|
||||||||||||||||||||||||
The colour seems different this year. Previously there was much more red in the flowers:
|
|||||||||||||||||||||||
While researching, though, it occurred to me that it's exotic and thus potentially invasive. Is it? The Wikipedia link suggests yes, but that bush has been there for several years now, and I haven't seen any others pop up. My guess is that our landscape doesn't offer the conditions for invasion.
|
Refining steak and kidney pie
|
Topic: food and drink, opinion | Link here |
I've made several batches of steak and kidney pie, most recently started on Saturday, but we discovered that we had so many kidneys left over that we ate them first, so it wasn't until today that we got round to making it. That required a number of adjustments to the recipe: more water, more salt.
And this time we considered and dismissed making it in the “air fryer”, and did it in the oven instead.
The eternal question: how long? In the end, we did them at 210° for 25 minutes, after which they looked like this:
|
|||||||||||||||||||||||
In particular, the gravy started coming out of the vent hole, so clearly it was boiling. But on taking them out of the oven, it seems that they could have had a little longer:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Probably they need 30 minutes, followed by at least 10 minutes cooling down.
| Thursday, 14 November 2019 | Dereel | Images for 14 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Davey stays all day
|
Topic: animals, opinion | Link here |
Kim Stanley brought Davey along today early, and he stayed the whole day. I was a little apprehensive, but after initial excitement he calmed down nicely, and things went well. About the only issue we had was when walking the dogs:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
He's old and has injuries, which have the interesting effect that he walks in a lateral gait, somewhat ambiguously called pace in English. In horses that's a relatively fast gait, but he does it even in a walk. And we were concerned that he wouldn't make the whole 2 km that we normally go, so we returned after about half way. Kim later told me that he should be able to make it; we'll see next time.
|
Piccola vents frustration
|
Topic: animals | Link here |
Piccola wasn't happy about Davey being there, of course, though she was obviously not overly afraid. Clearly he just disturbed her comfort. And at one point, she vented her frustration with a swipe—at Nikolai, her favourite, and an innocent bylyer. Why? She had to hit something, so she hit something that wouldn't hit back. Niko whimpered and looked sad.
|
Xanthorrhoea in flower
|
Topic: gardening | Link here |
The Xanthorrhoea in Grassy Gully Road are flowering:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Why are they always so twisted?
|
Planting tomatoes
|
Topic: gardening | Link here |
Somehow the weather isn't very spring-like at the moment, quite unlike further north, where they've been having some of the most devastating bushfires ever experienced, especially for so early in the season. I've been holding off on planting the remaining tomatoes. But it's still mid-November, almost summer, so they had to go in. A good thing, to: I had put them in a container when they were in the lounge, and outside it had filled up with water. Planted a „Riesentraube“ (which proves to be a cherry tomato variety) by the wind break at the north-east of the house, and an “Oregon Spring” round one of the arches:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder how the “Oregon Spring” will do on the arch.
While I was at it, also spread nearly 5 kg of fertilizer in the garden. I've had another case of a plant nearly dying, like this Coleonema pulchrum:
|
|||||||||||||||||||||||
There have been a number of those, including with relatively hardy plants, and I've only ever seen it here in Stones Road. My current hypothesis is that it's lack of nutrition.
|
Preparing for NBN complaint
|
Topic: technology | Link here |
As if yesterday's outage wasn't enough, today I received a new announcement from Aussie Broadband:
NBNCo has let us know that they are planning network maintenance in your area, and that your service at <strong>29 STONES RD, DEREEL VIC</strong> will be affected.<br>
The details are:
- Start date and time: Wed 4th December 2019 07:00 AEDT
- End date and time: Thu 12th December 2019 20:00 AEDT
- Window: 205.0 hours
You may experience the following interruptions during the maintenance
- 480 min
- 480 min
- 480 min
- 480 min
- 480 min
Another 40 hours! When will it ever end? But clearly it's finally time for a serious complaint. That would be with the Telecom Industry Ombudsman (TIO), and to do that I need to confirm that I have exhausted other attempts. What happened to that complaint I put in months ago? Back to my saved mail and found:
Date: Wed, 3 Jul 2019 14:28:59 +1000
Subject: Fault #6275739 created
Was that the one? Off to look at my ticketsfaults on Aussie. Only one, entered
on 12 April 2019 and now closed. Where did the fault go?
Checked in my diary for 3 July 2019. Yes, that was the occasion. Why is there no record of it on the Aussie support site?
Wrote an email to support about the matter, with the payload:
I've tried to access this fault number on your web site, but it's not there. It appears to be related to the complaint that I refer to at /grog/diary-jul2019.php?subtitle=NBN:%20three%20days%20planned%20downtime!#D-20190704-050320. At the time I was complaining about the perpetual NBN outages, and I was told that you would forward my complaint to the NBN, and that I would hear from them.
I have not heard from the NBN, but I have experienced even more outages, and they appear to be on the increase both in frequency and in duration. Since 17 September I have had 13 outages, mainly in the daytime, with a total duration of 25 hours service interruption. I have just been informed of yet more outages, a total of up to 40 hours between 4 and 12 December.
This is absolutely unacceptable. I have been using the Internet for 30 years now, and I have never had a service that has been nearly as bad. The fact that it's deliberate just makes the matter worse. I appreciate that Aussie's influence is limited, but it exists, and I would appreciate that you do something about it. At the very least I would like to know:
Please ensure that I have a response by close of business, 29 November 2019. If not, I will file a formal complaint with the TIO.
The response was not encouraging:
It usually takes us between 24 and 48 hours to respond to an email ticket.
If it's urgent, we suggest either:* giving our friendly Australian team a call on 1300 880 905 (8am to midnight Australian Eastern Time), OR* logging a fault via the MyAussie app or portal
OK, why does it take them up to 2 days to respond to email, when they can handle things on the phone much more quickly? I've been very happy with Aussie in the past, but I'm beginning to wonder whether that's now literally in the past.
| Friday, 15 November 2019 | Dereel | Images for 15 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Powercor loses to NBN
|
Topic: Stones Road house, general, opinion | Link here |
Wednesday's planned net outage wasn't the only one for the week. Today Powercor tried to match the NBN, with a planned 6 hour outage. That loses automatically, since the NBN outage was nearly 6½ hours. But they were only able to keep it up for 3 hours, 42 minutes.
More to the point, though, was the question: “How do I weather the outage?”. Yesterday evening I changed the inverter settings to charge the battery from grid up to 95% charge instead of the normal 40%, but I couldn't find the setting for minimum. There are two that come into consideration: “SOCmin: Minimum limit for switching to standby” and “SOCrecx: Value to generate voltage in the loads”. The settings (20%) match, but somehow neither seems to be related to “start charging from the grid when SOC drops to this value”. So I left it as it was, and it was only by chance that the battery had 98% charge (grey line) when the power went out at 9:7:29:
|
|||||||||||||||||||||||
And that's the way it stayed. The power came back at 12:49:16, but it's only visible on that graph by the grid feed-in that started then. We didn't need any significant grid power until evening, though the inverter did its usual low-level to-and-fro between grid and inverter.
Could it have been different? Yes, of course. Firstly, we knew about the outage, and I was waiting for it. Otherwise we could have been running the air conditioner and not notice the problem until it was too late. Surprisingly, I did receive a warning message from Ingecon:
Date: 15 Nov 2019 01:00:45 +0100
From: ingeras.noreply@ingeteam.com
Serial number Start End Description Type
0AM172B16A22 (E10017220395) 11/15/19 9:07:38 AM Grid voltage out of range Active for more than 1h
Apart from the incorrect date format, it's interesting that it decided that the power failed 9 seconds after it really did, and that the message wasn't sent until 11:00, nearly 2 hours later. And this was the only message. Nothing further when it continued, or when it finished. Basically, not much use.
|
More garden work
|
Topic: gardening | Link here |
Mick Solly around this morning to finally finish the first lawn mowing of spring. And he didn't bring his ride-on mower with him.
But he managed to tidy up relatively well with the push mower and whipper-snipper. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The down side is that that's the last of the green grass for this year.
|
Boobialla after all
|
Topic: gardening, opinion | Link here |
For some time I have been puzzled by this bush in front of the shade area:
|
||||||||||||||||||||||||
What is it? Lorraine Carranza took a cutting to Formosa Gardens and came back with the information that it was Boobialla, poisonous, and that we should get rid of it.
Was it? Is it? I did some research at the time, but for some reason didn't mention it in this diary. But finally it's flowering:
|
|||||||||||||||||||||||
Yes, that's clearly Myoporum insulare, better known as Boobialla. Should we get rid of it? Maybe. Not because it's poisonous, but because it's far too big, and the flowers are barely visible. But it's too late this year; it'll come out in the autumn.
|
Still more flash pain
|
Topic: photography, gardening, opinion | Link here |
The photos of the Myoporum insulare were an absolute pain. I took the first two with available light, but it was clear that that wasn't enough. OK, that's what I have the mecablitz 15 MS-1 ring flash for. Put it on the M.Zuiko Digital ED 60 mm f/2.8 Macro and the Olympus OM-D E-M1, along with the FL-LM3 to trigger the flash, and off to take photos.
After 6 attempts, I gave up. The flash didn't fire. Dammit, I've been having issues with (electronic) flash almost since I first got one. Why is it such a pain?
Afterwards, a more rigorous examination showed:
The Olympus FL-LM3 that came with the Olympus OM-D E-M1 Mark II doesn't work on the OM-D E-M1 Mark I. Why? It relies on power from the camera. But so does the FL-LM2 that came with the Mark I. Other flashes work on the Mark I, but not that one. And of course that was the combination I had tried.
But that's not all. Unlike the FL-LM3, the FL-LM2 doesn't have an on-off switch. You need to lift the head to turn it on. I had guessed that the FL-LM3 might have been part of the problem, and tried the FL-LM2 without “turning it on”.
The mecablitz doesn't fire just because the trigger flash fires, at least not when set to understand the Olympus flash protocol The trigger needs to be set in remote flash mode. That wasn't the case in the image that I finally used, but fortunately the exposure was adequate.
There is a difference between “On” and “off” in the remote flash menu for the E-M1 models. Here an example, without any other flash, showing what you get if you press the shutter while looking at the LCD display:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly “off” is a relative term.
Why is this all such a pain? 55 years and still not there.
|
A new home for Wotan
|
Topic: animals, opinion | Link here |
Wotan (and despite his Icelandic breeding not Oðinn), once Yvonne's favourite horse, injured himself years ago and is no longer rideable. Yvonne has found a number of homes for him, but somehow they all didn't work.
She didn't give up, and today Danielle Teo and two daughters came along to look at him. They're not overly experienced with horses, but Yvonne is happy with them, and there's a good chance that he will go to them in the near future. But nomen est omen: Oðinn was the wanderer.
| Saturday, 16 November 2019 | Dereel → Ballarat → Dereel | Images for 16 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More storage space
|
Topic: general, language, opinion | Link here |
ALDI had some shelving on offer today, just what I needed for the garage, so off early to pick some up.
Started trying to assemble them in the evening. The assembly was supposed to take 30 minutes and not require any tools except for a hammer that was suddenly mentioned towards the end. After 45 minutes I had the first one about 80% assembled:
|
|||||||||||||||||||||||
After that I had to prepare dinner, so it had to wait for tomorrow.
At first the “instructions” looked a cut above the average for this kind of equipment, but that was just the first impression. What language! It refers to two different kinds of cross members (“cross beams”), horizontal and vertical. What's a vertical cross member? It's one along the long side. Why “vertical”? Either the person writing the documentation was lying drunk on the floor, or it's an artefact of translation (the shelves come from Poland, so presumably the documentation was originally in Polish). If that's the case, I can only assume that the translator was working without the illustrations.
There are more issues, but I'll leave that until it's complete.
|
Goodbye, LEMIS
|
Topic: technology, general, opinion | Link here |
LEMIS ceased trading in 2007, but I've dragged my heels for years to get it deregistered. Finally it's done, as I discovered after asking my accountant what was going on—on 26 September. That's something like 40 years since it came to be, back in Friedrichsfeld. How time flies!
|
Rumours of my demise...
|
Topic: photography, politics, opinion | Link here |
Interesting, but slightly disturbing, news today: Olympus closure is near, can happen in less than 8 months. Huh? That's a blog entry of a certain Vitaly Kiselev, whose home page seems to consist mainly of ban notices. The content implies, rather than clarifies, that he's talking about the imaging division.
OK, some random detractor. What does Olympus say? 奥林巴斯公司 公布新经营战略称不会关停相机业务, surprisingly in Chinese. Google translate offers Olympus announced that its new business strategy will not shut down the camera business .
OK, who's behind this? We see lots of use of web sites to influence politics. This one looks to me like a (possibly paid) attempt to influence business. My first thoughts were: “Sony” and “Tony Northrup”.
Why? I don't trust Sony's business integrity—we saw that years ago with their abuse of the Gnu public license, though it seems that the web has forgotten this matter, maybe the same way that it forgot the elk test (which Wikipedia insists on calling a “moose test”).
And I've already noted that Tony Northrup appears to be paid to put down the Micro Four Thirds system.
And sure enough, off he goes:
There's a lot of stuff in there, but some of the highlights:
“Don't ask the company”. No, of course not, that would be against the content that he's trying to get across. Yes, he's right that the answer you get might not be as accurate as you like, but companies have obligations to their shareholders, so they can't produce outright lies. And Olympus has said that they are not
“Smart phones produce better image quality than Micro Four Thirds”. Yes, of course. Pull the other one, it's got bells on.
At 15:36 into the clip: “smart phones produce more background blur”. Well, they might simulate it, but that's about as far as it goes.
18 seconds later, 15:54: “Limited FF-equivalent fast lenses”. NO! That's a downright lie. There are more fast lenses available for Micro Four Thirds than for any other system. Nikon has just brought out its first f/0.95 lens in 50 years or so; Canon still doesn't have one. And there are at least 6 for Micro Four Thirds.
Clearly Tony is talking about this stupid “aperture equivalence”: an f/0.85 lens (the fastest available, and only for Micro Four Thirds) is “equivalent” to an f/1.7 lens on a “full frame” camera. But it's not. It's equivalent to a (non-existent) f/0.85 lens.
The stupidity is that this “equivalence” is only related to depth of field. As I discussed four months ago, it's not clear that less depth of field is an advantage; I have far too many photos from my 35 mm days that have insufficient depth of field. And if he thinks that this is an issue, he should explain the “background blur” technique on mobile phones. But of course, that's not the point: he's trying to find arguments against Micro Four Thirds.
At 17:00 he notes that Olympus is not doing much R&D. Yes, that's an issue. He considers the autofocus on the OM-D E-M1X to be inferior. I can't argue, but it has object recognition that promises to be better than current techniques. And the cameras have things like focus stacking that no other camera has. Then there are things like high speed image capture (up to 60 frames per second), including “Pro capture”, a silly name for a method to save images taken before the shutter release was pressed. In many ways, other manufacturers are trying to catch up.
17:30: Olympus is losing money. Yes, correct, and they've explained how they intend to fix the problem, and have demonstrated that they're on track. No mention of that from Tony, only “how could they turn it around?”. Well, read what they said, and look at the figures. “Analyze it like an investor”. Yes, read what they said, and look at the figures. In other words, he's making claims that don't match the data, not for the first time.
Olympus has done this in the past, discontinuing the Four Thirds system. Well, sort of. It replaced it with the upwards compatible Micro Four Thirds system. I don't see that as an argument; other makers are doing the same thing.
Somewhere: Olympus doesn't have a lot of lenses available. Tony knows that this claim is wrong. I looked at the relative availability 6 months ago and came to the conclusion that they're way ahead of the competition, certainly helped by Panasonic.
So: all just nonsense? No, this could be a self-fulfilling prophecy. And Olympus has left its users in the lurch in the past, with the sudden end of the OM system. It's concerning at least for those reasons.
| Sunday, 17 November 2019 | Dereel | Images for 17 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Finishing the first shelf unit
|
Topic: general, opinion | Link here |
Continued with the shelf unit this morning, and finished it pretty quickly. But it wasn't without surprises. Despite all care, I had managed to mismatch the height of no fewer than 3 cross-members, like this one (from yesterday's photo):
|
|||||||||||||||||||||||
The units are held together by tongues on the cross members sticking into slits in the columns. They're supposed to go in and out again on the same side of the column, but it's very easy to get them coming out on the other side or jam on the edge of the slit:
|
|||||||||||||||||||||||
I spent much of the time trying to get them to fit in correctly. I found that I could push in the end of the tongue with a small pipe wrench did the job.
|
Tool terminology
|
Topic: language, opinion | Link here |
Are these metal strips that fit into the columns on the shelving units really called tongues? The instructions carefully avoid giving them any name at all. But after much consideration, I can't think of an alternative term.
And then: I used a pipe wrench to push in the tongues on the shelving units. Or did I? Pipe wrench? What's that? I thought I knew my terms, but clearly others don't. What I'm talking about is this kind of tool, and I used the nose to push in the tongue:
It's amazing how confused the terminology is. I found this particular image at tongue-and-groove pliers, a term I have never heard of. But the article gives alternate names: water pump pliers, adjustable pliers, groove-joint pliers, arc-joint pliers, Multi-Grips, tap or pipe spanners, gland pliers, and the image above is titled “multiple pliers”. And there are other wrenches also called pipe wrenches, but which look completely different:
My guess is that every city has its own terminology. No wonder they never finished the Tower of Babel.
|
Carpobrotus in full flower
|
Topic: gardening | Link here |
The smaller Carpobrotus in the island in front of the house are in full flower:
|
|||||||||||||||||||||||
And down Harrisons Road the one Solanum aviculare plant is flowering more than I have seen before:
|
|||||||||||||||||||||||
Is this due to the moist spring?
|
Can opener pain
|
Topic: food and drink, opinion | Link here |
I've been using can openers for at least 60 years. They're not exactly high-tech, and for the past 50 years or so the standard opener looks pretty much like this:
|
|||||||||||||||||||||||
But I've had difficulty with them in the past: cans are giving way to single-use plastic, and it's not clear whether this is a good or a bad thing. But the cans themselves are changing: they're thinner and more easily deformable, and I suspect that the lid is indented further from the top, making it difficult to open properly. So three years ago we bought a new opener, which, though strange, seems to work.
But today I couldn't find it, so I had to use the old one. What a mess! It wasn't enough by itself; I needed other tools too:
|
|||||||||||||||||||||||
And that's without my bleeding hand, caused by the screwdriver slipping. By the time I got round to looking at it again, there wasn't much to see.
Why is this so difficult?
|
Blue spot fish
|
Topic: food and drink, opinion | Link here |
ALDI had “Blue spot emperor fillets” on special this week. What's that? A kind of fish. It doesn't seem to be well known. According to an Australian government web site, Bluespotted Emperor is Lethrinus punctulatus, but that (Wikipedia) link redirects to Lethrinus lentjan, a different species with the names pig-face bream, pink ear emperor, pink-eared emperor, purple-eared emperor, purple-headed emperor, red spot emperor and redspot emperor.
Still, it looked good, so I bought a pack, and we made sweet and sour fish out of it. Verdict? Boring. Maybe we should have followed ALDI's recommendation and just pan-fried it.
| Monday, 18 November 2019 | Dereel | Images for 18 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Coffee machine problems
|
Topic: food and drink, opinion | Link here |
One of the big differences between the old coffee machine and the new one is that the old coffee machine had a toy display that produced (mainly) images rather than text:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's a bit-mapped display, so it could show text too, and just occasionally it does.
And the new machine? It's the newer of the two that I looked at, and one of the advantages that the salespeople mentioned was that it had a display; the older model had none at all. But what a display! And what reliability! This morning I tried to make a cup of coffee, but all I got was:
|
|||||||||||||||||||||||
What does that mean? Broken already, after only 9 days? Where's that instruction manual? At least the old machine had symbols that vaguely meant something, and they were colour-coded to show the seriousness. But the new display only has fixed images, all in red. To set “water hardness 1”, the softest water (they don't cater for people like me with no water hardness at all), you go through various incantations that light up a varying number of these symbols:
|
|
|||||||||
In passing, admire the missing symbol at top, probably related to a misassumption on the part of the author.
So what does one symbol mean? My guess was “grounds container full”. Well, almost. From the instruction manual (not the one delivered with the machine, which is formatted differently, though the content appears the same):
|
|
|||||||||
It was not flashing, though a warning triangle, undocumented here, was:
|
|||||||||||||||||||||||
Nope, container empty, fits nicely.
FOOL! You've been looking in the wrong place. That was on page 9, “description of lights”. Why are you looking there? On page 18 there is:
|
|
|||||||||
It also includes the helpful solution: “Fill the beans container”.
And sure enough, that was the (obvious) solution, made 10 times worse by appalling ergonomics and documentation. De'Longhi has now well and truly made it into the 20th century.
|
New power pole
|
Topic: general | Link here |
While walking the dogs, found this:
|
|||||||||||||||||||||||
Clearly it's a power pole, and it appears to have been the reason for last week's power outage.
But why is it there? There's another pole only 20 m away on the other side of the road, and it's in the middle of a paddock. Yes, there's at least one transformer on it, along with other equipment and what looks like a junction box with a coiled-up flexible conduit underneath on the right-hand side. Are they planning to connect something there? What? A house? This used to be a sheep paddock, and I didn't see anything for sale there.
|
More Internet pain
|
Topic: technology, opinion | Link here |
Yvonne processed some photos today, by no means for the first time, but she ran into trouble syncing them to our external server. No response!
Dammit, has it gone down? No, I could access it fine. But only from eureka. It proved that teevee couldn't access it either. After some examination, it proved to be the firewall (with ipfw and natd). The rules of interest are:
=== root@eureka (/dev/pts/11) ~ 363 -> ipfw show
00021 97446 86838832 allow ip from 45.32.70.18 to 167.179.139.35
...
00050 608016833 700320792987 divert 8668 ip from any to any via xl0
So how does this work? lagoon (192.109.197.134) sends a datagram to www (45.32.70.18). natd on eureka translates the sender IP address and sends it off to www with eureka's external address 167.179.139.35 as sender address. www receives the datagram and sends a reply. Explicitly accepted by rule 21.
And then? It should end up at natd, but it seems that it gets delivered (and discarded).
So what should I be doing? Clearly I can't establish connection with machines behind natd from outside. But I can deliver mail, which was the purpose: 167.179.139.35 is also the address of mx0.lemis.com, my lowest-order MX, and that's the reason for the rule in the first place. The workaround is obvious: limit the rule to SMTP:
00021 0 0 allow ip from 45.32.70.18 to 167.179.139.35 dst-port 25
That only works because none of the systems behind natd exchange SMTP traffic with the outside world. But what about HTTP? I had thought of some limited HTTP access from outside, and that would block web browsers on the other machines. I'm still puzzled.
| Tuesday, 19 November 2019 | Dereel | Images for 19 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PIXIO fail
|
Topic: photography, multimedia, opinion | Link here |
Anke Hawke is coming for another clinic this weekend, and clearly we want something better in the way of recordings than last time, where I had a relatively painful time (mainly) behind the camera.
In principle we now have the sound side of things sorted out. The only other issue was getting the device level.
OK, put the OM-D E-M1 Mark II and microphone receiver on the PIXIO, remove the panorama head from my tripod, leaving the leveling base in place, mount the whole thing on the tripod, and off.
What a fiddly camera mount this thing has! And it's quite flimsy. I went to check that the weight of the camera and lens wasn't too much. But no, according to the specs the maximum load is “2.5 kg (5 lbs)”, and my combination only weighs 1.14 kg. Still, there's plenty of room for improvement. That could be as simple as a small ball head.
Out to the arena and set the thing up. Turn on beacons. Turn on badge. Turn on PIXIO. Four flashes of the green LED, and then it turned off again. Dammit, what does that mean? Should I try it longer? It continued to blink as long as I pressed, but then turned off again. Next time round it alternated with a red LED.
Back to RTFM. Ah, I shouldn't have turned the badge on yet. OK, try again with the badge turned off. Same thing.
RTFM again. Green/red alternate means that the device is in slave mode. How could that happen? Hold down the on/off button long enough.
Finally it occurred to me what I had read the very first time:
5. To avoid self-discharge of the robot, unplug the robot after the end of the robot charge, turn the robot ON for a few seconds and then turn it OFF before storing it.
What does that mean? Clearly a bug in the charging circuitry, and a workaround to match. And I hadn't done so the last time. Battery discharged? How do I know? There's no display, and no mention of it in the “instructions”.
OK, abort, take the whole thing apart, in the process noting that there's no real way to unscrew the PIXIO from its mount: the only part you can take hold of rotates on the head. Connect all five components to the battery charger. The beacons and the badge took about 5 minutes to charge. The PIXIO itself took 2 hours, clearly completely discharged. After charging, turned it on as instructed, and discovered that the LED was much brighter, and that it continued to flash until I turned it off again.
What a disaster! The triangulation works well, but the rough edges, both in the charging circuitry and the documentation, are unacceptable.
|
Dianella revoluta
|
Topic: photography, gardening, opinion | Link here |
One of the prettier flowers that I have seen is Dianella revoluta, which is relatively rare, but this year there are more than most. Brought back a twig from Westons Road while walking the dog, and tried some focus stacked images. They still need to be processed, but the first non-stacked attempts were less than encouraging:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I would never have thought that the (natural) light was so uneven at the back of the office.
| Wednesday, 20 November 2019 | Dereel | Images for 20 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Nasi goreng revisited
|
Topic: food and drink, language, opinion | Link here |
Nasi goreng is a Malay and Indonesian term meaning literally “fried rice”. I've known it for about 65 years: early in our time in Kota Bharu we were walking through the market, and I saw some rice. Clever 5-year-old says “goreng!”.
Wrong, doubly. Malay adjectives follow the noun, so it's “nasi” that means rice, and “goreng” means fried. But “nasi” would also be wrong in this context. There are many words for rice: “padi” is the rice in the field, “beras” is the word for husked rice, and “nasi” means cooked rice.
And of course we ate lots of nasi goreng. It was one of the first dishes that I tried to cook myself, and I've done it many times since. But it wasn't until four years ago that I wrote down a recipe, mainly for other people, and which I promptly forgot.
Nasi goreng is best for leftover rice, and we had plenty of it this week. OK, make breakfast. Rice, bacon (my personal favourite), spring onions, peas, prawns, soya sauce (dark and light), eggs, cucumber—pretty much my ”breakfast” slant in the recipe. I don't need to look it up.
But it tastes boring! What's missing? Off to look online and offline (recipe books), and found significant differences. Indonesian(-language) recipes include trasi (Indonesian for belacan), and one has sausages (though they can be Chinese sausages). My only Malay-language recipe includes celery, chicken livers and tomato sauce. Wendy Hutton's Singapore Food includes raisins! And Rosemary Brissenden's “South East Asian Food”, which I have had for nearly 50 years, includes raw rump steak (½ lb for “one plate” of rice)! To think I once considered the book authentic.
But they had one thing in common that was missing from my recipe: garlic. Why was it missing? It wasn't in the recipe. I just forgot.
One thing I did learn: use very little dark soya sauce. It's almost overpowering.
|
Dan Murphys: No customer service
|
Topic: technology, food and drink, opinion | Link here |
Everybody knows Dan Murphy, the author of TECO, the first real computer editor. But in Australia there's another Dan Murphy, a large retailer of alcoholic drinks. And they're still trying to establish an online presence, and I have some kind of “account” with them.
Today they came up with an interesting offer: Adelaide Hills Sauvignon Blanc for $25 a bottle, but for members only $10. Is it worth it? They're selling in units of 6 bottles, so it was worth a try.
Oh, you have to log in. No worries, details are saved, press “log in”. After 5 minutes of spinning, gave up and tried in a different browser instance (also firefox), this time logging in first. Through the usual steps, pressed “pay now”. Brief hang, then:
|
|
|||||||||
That's interesting, particularly since I chose PayPal. Retry? No, though it suggested retrying, that part of the page had disappeared.
OK, try again with Chromium. Same thing. But I got mail from PayPal:
Date: Tue, 19 Nov 2019 19:49:48 -0800
From: "service@paypal.com.au" <service@paypal.com.au>
To: Greg Lehey <groggyhimself@lemis.com>
Subject: You've received a refund from Dan Murphy's Online
Dan Murphy's Online has sent you a $60.00 AUD refund for your purchase on 19 November 2019.
Now wouldn't that be fun if they had refunded me money without booking it in the first place? But no, it seems that PayPal has preferential notifications. The corresponding booking notification came 5 minutes later at 19:55:05 -0800.
OK, follow the instructions that say I can send mail to, an appropriately named customerservice@danmurphys.com.au. Cut, paste and send a message to them.
Date: Tue, 19 Nov 2019 20:11:23 -0800 (PST)
From: Mail Delivery Subsystem <mailer-daemon@googlemail.com>
Subject: Delivery Status Notification (Failure)
** Address not found **
Your message wasn't delivered to customerservice@danmurphys.com.au because the address couldn't be found, or is unable to receive mail.
The response was:
The email account that you tried to reach does not exist. Please try double-checking the recipient's email address for typos or unnecessary spaces. Learn more at https://support.google.com/mail/answer/6596
Marvellous! In passing it's interesting to note that they use Gmail for their mail.
OK, how about the Contact Us link? Not found. Clearly they don't want contact. Oh, yes, you can click on the “Contact Us” link on the “not found” page, and that links to a web form that thinks it's an email interface. But it doesn't like my web browser:
Please upgrade to a supported browser to get a reCAPTCHA challenge.
Damn you, Dan Murphy! Do you want to sell me things or make life difficult? OK, tried with a different browser, and finally sent a complaint. I don't know why I bother. There are other vendors who don't give me that much trouble.
|
Garden progess
|
Topic: gardening | Link here |
It's been over two years since we bought two Clematis plants. One is doing acceptably, but the other, a General Sikorski, suffered one problem after another, being almost swallowed by a Tropaeolum. But it didn't give up, and today, after over two years, we finally have a flower:
|
|||||||||||||||||||||||
Yes, it doesn't look like much, but there are more buds on the way.
Other things aren't doing as well. This chili plant has suffered, I think, from being sprayed with fertilizer:
|
|||||||||||||||||||||||
And this Lilac is suffering from not enough water:
|
|||||||||||||||||||||||
Watered it well. It'll get a dripper Real Soon Now.
| Topic: Stones Road house, general, opinion | Link here |
Found this in the garden below the PV arrays:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It looks very much as if it could be part of the mounting brackets. There's nothing obvious, but it's worth checking on, so sent a message off to Tomas Kucera of Effective Electrical about the matter.
Looking at the clamp, there's a good chance that it was dropped and never used: there's no evidence of any pressure on the bolt. And I got a quick reply saying the same thing:
It will be one that was dropped during the installation process.Not serious the boys have boxes full of them and would have put the correct number on the panels.
At least it confirms the nature of the clamp, and that I had reported it.
| Thursday, 21 November 2019 | Dereel | Images for 21 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Bloody SMEG range hood!
|
Topic: general, opinion | Link here |
I've complained about our SMEG range hood almost since we got it, but in the end I gave up against the force of the Establishment. And we can just about get by with it.
But one thing that has annoyed me in the past were these silly halogen lamps: they're unreliable (we must have been through 6 of them so far) and extremely difficult to change.
|
||||||||||||||||||||||||
Last week the left-hand one burnt out—I think it's the one I replaced only 4 months ago—and it took me forever to insert the replacement in the socket. I think that the socket must have sustained some minor damage that makes it difficult to insert, not made any easier by the fact that it's located on the side of the reflector, and there's no way to apply pressure evenly.
And today the other one failed! Once again, I couldn't get the bulb into the socket. When I did find a way to apply more force, I managed to snap off one of the legs before I could insert it.
Nothing for it. The reflector is attached to the body with a rivet. Bored it out, dismantled the assembly, and I was able to access the socket at a normal angle. Then it was relatively easy to insert the globe, though it's not clear how long it will stay that way if I have to replace the globes every few months.
And reassembly? Rivets are clearly not the situation. How about screws? Do I have one? I don't know. But it turned out that I didn't need one: the reflector fitted pretty firmly in place without one.
Photos? No time. There will, sadly, be a next time.
Now why couldn't they have used conventional illumination?
|
Davey stays the night
|
Topic: animals | Link here |
Kim Stanley along early this morning with Davey, who was to stay the night. She's off to hospital next week, and we'll have to look after him—if things work out. Otherwise it's off to a boarding kennel for him.
Things worked out well, and this time we were able to walk him the length of Grassy Gully Road:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Another temperature plummet
|
Topic: general, opinion | Link here |
One of the strange things about Victorian weather is how temperatures can plummet in a matter of minutes. I first noted this on Black Saturday, where the temperature dropped from about 45° to 31.7° in 10 minutes, a drop of 1.3° per minute:
These weren't my own measurements: I hadn't written my weather station software yet, so I couldn't measure the steepest drop. But on 6 January 2018 I could: up to 2.2° per minute, though the overall drop wasn't as much, only 8.2° in 12 minutes.
But then today we had an even more extreme case. In 10 minutes the temperature dropped by “only” 12°, but between 13:18:57 and 13:19:59 (62 seconds) it dropped by 2.9°, an average of 1° in 21 seconds:
+----------+--------------+
| time | outside_temp |
+----------+--------------+
| 13:16:54 | 37.3 |
| 13:17:56 | 36.8 |
| 13:18:57 | 35.3 |
| 13:19:59 | 32.4 |
| 13:21:00 | 31.1 |
| 13:22:02 | 30.4 |
| 13:23:03 | 28.9 |
| 13:24:04 | 27.6 |
| 13:25:06 | 26.9 |
| 13:26:07 | 25.9 |
| 13:27:09 | 25.4 |
+----------+--------------+
As the other data show, this wasn't just a glitch in a reading, and as on Black Saturday, there was a second drop later on. By 15:44 the temperature was 15°, a drop of 23.2° since midday.
| Friday, 22 November 2019 | Dereel | Images for 22 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PIXIO: Close, no cigar
|
Topic: multimedia, photography, animals, opinion | Link here |
Anke Hawke arrived this afternoon, and she promptly gave Yvonne a riding lesson. Time for me to set up the PIXIO “Robot Cameraman” and take some real video.
How did it work? Parts of it were excellent. One issue was that the camera (the Olympus OM-D E-M1 Mark II) could only record 29 minutes at a time. I thought those restrictions were dead and gone, and it's a nuisance. After we finished, I discovered further issues:
In passing, what a mess this stuff is! There were no fewer than 9 battery-powered devices: the three PIXIO beacons, the PIXO device itself, the PIXIO badge, the camera, the microphone receiver and the two microphone transmitters. Just setting up the PIXIO devices for charging is a pain. The camera itself did better than I expected. I started with a freshly charged battery, and after about 75 minutes of recording it still showed 55% charge.
|
Cailles aux raisins secs: insights
|
Topic: food and drink, opinion | Link here |
Cailles aux raisins secs for dinner tonight. They were alright, but I think I've identified two issues for improvement. Firstly, cook them longer. I had been doing them for 30 minutes, but I think that could be extended considerably. 60 minutes, maybe? And secondly, I served the lardons and raisins separately rather than putting them on the birds. Bad idea: much got left over. Next time they go back on the birds.
| Saturday, 23 November 2019 | Dereel | Images for 23 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More PIXIO fun
|
Topic: multimedia, technology, animals, opinion | Link here |
Today was the first day of Anke's clinic, but somehow she and Yvonne found time to do some more instruction in our arena, and I tried again to get the PIXIO “Robot Cameraman” to track the rider. Last time I had tried an only slightly modified version of this arrangement from the manual:
|
|||||||||||||||||||||||
The only difference was that I had placed the camera equidistant from the sides. It also needed to point in the other direction (“out of the screen”), but that has nothing to do with the orientation. With this layout the PIXIO tended to lose track of the rider when coming from beacon 1 and passing beacon 2.
Why? There was no obvious reason. Problems with beacon 2? I found that it turned on when attached to the charger, and I had to turn it off again. Or then the manual says
Check that there are direct lines of sight between the beacons, and between the robot and each beacon. Or at least check that any obstacle is small.
Could it be that the rider came too close to the beacons? I could mount them further away (there's another fence round the two sides of the arena that come into question), but that's not what the images show.
OK, today tried putting beacon 3 on the other side of the arena. Since the camera is in the middle, this is almost completely symmetric. The only difference is that beacon 1 is special.
The results? The thing still loses sync. But not in the same place. They're in many places in this clip:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I'm going to have to reconsider the whole thing. First, though, tomorrow I'll use the other layout that they suggest in the manual:
|
|||||||||||||||||||||||
On the positive side, the batteries all held out. I had used NiMH batteries in the microphones this time. It looks like becoming a habit to recharge 11 batteries every time we use the thing. Interestingly, the one battery that doesn't seem to require such frequent recharging is the camera battery. So much for people who say that mirrorless cameras have poor battery life.
Today's videos are part 1, part 2 and part 3 (Anke riding Carlotta).
|
Joining video clips
|
Topic: multimedia, technology, opinion | Link here |
So I have a number of videos from my Olympus OM-D E-M1 Mark II, all less than 4 GB in size. How do I join them?
Off looking on the web, and of course there's lots of stuff in the Microsoft space. But surely something like mencoder can do it? Off looking there and found surprisingly little basic information. But with a bit of experimentation discovered that I could join them with:
=== grog@eureka (/dev/pts/14) ~/Photos/20191123/orig 245 -> mencoder -ovc x264 -oac mp3lame -o 1.avi 4B24413[678]*
It wasn't all plain sailing, though. It set off with its usual plethora of marginally useful information, partially masking some of the problems:
MEncoder SVN-r37899-snapshot-3.4.1 (C) 2000-2016 MPlayer Team
success: format: 0 data: 0x0 - 0xece0c7b3
libavformat version 57.48.103 (internal)
libavformat file format detected.
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x21cf050]Protocol name not provided, cannot determine if input is local or a network protocol, buffers and access patterns cannot be configured optimally without knowing the protocol
[lavf] stream 0: video (h264), -vid 0
[lavf] stream 1: audio (pcm_s16le), -aid 0, -alang jpn
VIDEO: [H264] 1920x1080 24bpp 50.000 fps 50202.8 kbps (6128.3 kbyte/s)
[V] filefmt:44 fourcc:0x34363248 size:1920x1080 fps:50.000 ftime:=0.0200
==========================================================================
Opening audio decoder: [pcm] Uncompressed PCM audio decoder
AUDIO: 48000 Hz, 2 ch, s16le, 1536.0 kbit/100.00% (ratio: 192000->192000)
Selected audio codec: [pcm] afm: pcm (Uncompressed PCM)
==========================================================================
Opening video filter: [expand osd=1]
Expand: -1 x -1, -1 ; -1, osd: 1, aspect: 0.000000, round: 1
==========================================================================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
libavcodec version 57.55.101 (internal)
Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
==========================================================================
MP3 audio selected.
Could not find matching colorspace - retrying with -vf scale...
Opening video filter: [scale]
Movie-Aspect is 1.78:1 - prescaling to correct movie aspect.
[swscaler @ 0x21fdff0]bicubic scaler, from yuv420p to yuv420p using MMXEXT
[swscaler @ 0x21fdff0]using unscaled yuv420p -> yuv420p special converter
x264 [info]: using SAR=1/1
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 AVX2 LZCNT BMI2
x264 [info]: profile High, level 4.2
Pos: 0.0s 2f ( 0%) 0.00fps Trem: 2min 0mb A-V:0.002 [0:0]
1 duplicate frame(s)!
Pos: 0.0s 3f ( 0%) 0.00fps Trem: 2min 0mb A-V:0.004 [0:0]
1 duplicate frame(s)!
Pos: 0.0s 4f ( 0%) 0.00fps Trem: 3min 0mb A-V:0.004 [0:0]
1 duplicate frame(s)!
Pos: 0.6s 30f ( 0%) 0.00fps Trem: 6min 0mb A-V:0.042 [0:233]
Skipping frame!
Pos: 1.1s 57f ( 0%) 33.06fps Trem: 9min 0mb A-V:0.040 [0:246]
Skipping frame!
Pos: 1.2s 60f ( 0%) 30.63fps Trem: 10min 0mb A-V:0.022 [0:247]
Writing header...
ODML: vprp aspect is 16:9.
Setting audio delay to 0.048s.
Writing header...
ODML: vprp aspect is 16:9.
Setting audio delay to 0.048s.
Writing header...
ODML: vprp aspect is 16:9.
Setting audio delay to 0.048s.
Writing header...
ODML: vprp aspect is 16:9.
Setting audio delay to 0.048s.
Writing header...
ODML: vprp aspect is 16:9.
Setting audio delay to 0.048s.
Pos: 5.6s 283f ( 1%) 3.03fps Trem: 150min 708mb A-V:0.036 [10500:262]
What are those duplicate frames? They show up as video noise at the beginning of the output video. They're gone after a few seconds, but why are they there at all?
The other issue is the glacial speed. It maxes out over 6 processors on a relatively fast machine, converting at roughly 40% of the recording rate (taking over an hour of elapsed time and 5½ hours of CPU time to process a 30 minute clip):
USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND
grog 54141 664.4 2.0 1000452 679868 14 SN+ 11:44am 327:25.24 mencoder -ovc x264 -oac
Can ffmpeg do better? To be investigated.
|
Garden flowers in late spring
|
Topic: gardening, opinion | Link here |
Today was a month before the December solstice, time for the monthly garden photos.
Probably the biggest success this month are the roses. This time last year I was also happy with the results:
|
|||||||||||||||||||||||
But clearly they're better this year:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
One thing that has changed is the Tropaeolum that was growing against the fence. That proved to be responsible for the near-death of the Clematis General Sikorski, which has now recovered relatively well. The rather timid flower of a few days ago has now opened, and more are on their way:
|
|||||||||||||||||||||||
But things are only relative. The other clematis, which flowered last year, is hardly overly active:
|
|||||||||||||||||||||||
Still, the idea of more fertilizer is gradually making itself felt.
The Banksia integrifolia continues to flourish, and to my surprise there are what appear to be three different kinds of flower:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why is that?
The Box elder that we thought dead has recovered:
|
|||||||||||||||||||||||
The hanging baskets in the entrance are still alive, but there's no obvious difference in their appearance:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Honeysuckle at the west side of the dog run is in full flower:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The last water Iris of the season is still flowering, and the first water lily has just appeared:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And the first Strelitzia reginae for nearly 3 years is promising to flower:
|
|||||||||||||||||||||||
It's still a bit behind the state of 3 years ago, though. This was on 30 November 2016:
|
|||||||||||||||||||||||
And these small blue wildflowers are doing well:
|
|||||||||||||||||||||||
As the ant indicates, they're tiny, about 1 cm across. I must finally get round to taking some good photos of them.
The Grevillea robusta that we planted 5 months ago has not done as well as I had hoped:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
My guess is that horses are to blame; the mesh on top wasn't there when I planted it.
|
Spam beats NBN
|
Topic: technology, opinion | Link here |
Seen in my inbox today:
207 + 22-11-2019 Aussie Broadband To grog@lemis.c ( 584) + nbn Unscheduled Outage
208 N + 21-11-2019 UltraWifiPro Holiday To grog@lemis.c ( 94) N + Give the Gift of High Speed Internet This Holiday
Now wouldn't high speed Internet be a good idea? Reliability would be even better.
To be fair to the National Broadband Network, this is one of the very few unscheduled outages. Normally they schedule them for just about the entire time. Was I hit by this one? I don't know. The description was so vague that I couldn't tell.
|
Turducken again
|
Topic: food and drink, opinion | Link here |
Big dinner for the Anke Hawke clinic today, and since I wasn't involved in the clinic, I had to do the cooking. We had, with some reservations, bought another turducken, mainly because we couldn't find anything else that met our criteria, and after all, they might have improved it in the meantime
How long does it take to thaw a 1 kg turducken? According to the package, 2 days in the fridge! I took it out in the morning, and by 14:00 it was still almost completely frozen.
Panic! Started gently thawing in the microwave oven, and an hour later about 1 cm from the surface had thawed. Yvonne came by and we discussed it, but by that time, with little outside help, it was almost completely thawed, so we could eat it after all.
But look at the roast! It's in a net, of course, because of its composition. But what's that stuff on top? It smells of fenugreek! What a way to spoil a “Christmas” bird! And it wasn't shown on the packaging.
|
|||||||||||||||||||||||
Scraped it off and hoped that the flavour wouldn't remain:
|
|||||||||||||||||||||||
Then there's the question of cooking times, of course. ALDI stated that it should be cooked at 190° (without any distinction for fan-forced or normal ovens) for 95 minutes covered in aluminium foil, and then for 25 minutes without. I set 180° fan-forced, and it reached 80° after 86 minutes, and after 100 minutes it was nicely browned:
|
|||||||||||||||||||||||
All well and good, but that was 20 minutes earlier than planned, and it wasn't enough. But somehow we managed, though for the first time we had nothing left over after such a meal. On the positive side, the taste of fenugreek didn't remain.
|
Nikolai finds a friend
|
Topic: animals | Link here |
Once again Nele Kömle came and brought Katherine with her. Katherine discovered a like for Nikolai, and spent a lot of time attending to him, to his obvious pleasure:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Sunday, 24 November 2019 | Dereel | Images for 24 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PIXIO success
|
Topic: photography, multimedia, opinion | Link here |
The clinic continued today, and I was up and in the arena by 8:00 for another lesson for Yvonne. As planned, this time I positioned the beacons as planned:
|
|||||||||||||||||||||||
And how about that, it worked! No loss of position. Gradually we're getting there. Why did it not work before? There's a warning in the instructions against placing the camera exactly between two beacons:
|
|
|
|||||||||
But that wasn't the layout I had. The diagonal line between beacon 3 and the opposite corner doesn't come within 4 m of the camera. If anything, that diagonal is closer the way I did it today. So that is probably not the reason. Maybe the overall distance between beacon 1 and 2?
The microphones are also no longer an issue: they just work.
|
ffmpeg instead of mencoder?
|
Topic: technology, multimedia, opinion | Link here |
mencoder was able to join my videos yesterday, but at what cost! One minute of video requires about 15 minutes of CPU time to convert! Can ffmpeg do better? Tried today and got similar results. The only difference was that it was much more complicated, and the first time I only managed to copy the first input file in about 90 minutes of CPU time.
But looking back at the output from mencoder, that was to be expected:
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
libavcodec version 57.55.101 (internal)
Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
When I have time, I'll look for other alternatives.
|
Harrisons Road progress
|
Topic: Stones Road house, general, opinion | Link here |
Five years ago today our house was standing and had a roof:
|
|||||||||||||||||||||||
Today the JG King house in Harrisons Road was a step further on. It had been “wrapped”:
|
||||||||||||||||||||||||
In general it seems that it's about a week to 10 days ahead of our house at the same time of year. Can I predict that they'll move in round 25 April?
|
Wild plants
|
Topic: gardening | Link here |
I take lots of photos of wildflowers, but of course there are many other plants as well. Yvonne identified some down Harrison's Road today:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I should spend some time identifying them.
|
Air-fried spare ribs
|
Topic: food and drink, opinion | Link here |
Yvonne bought some pork spare ribs recently, something that we normally don't eat. How do you cook them? Took a look in my cookbooks and online and came to the surprising recognition that they're all almost the same. Managed to quickly reject this one, not because of the recipe itself (again, very similar to the others), but because of:
Equipment
There will be no steenking measuring cups and measuring spoons in my kitchen! And how many knives do you need? A strong meat knife to cut the ribs, a small knife to cut the ginger and garlic, and that's it. But also an unmentioned blender to mix the marinade. And the mention of the brand name for the air fryer suggests paid advertising. Understandable, but not for me.
This recipe also had a couple of interesting instructions:
Set the air fryer at 375F for 8 minutes.Check to ensure the ribs have an internal temperature of 165F before serving.
Presumably the first means to cook them in the air fryer at “375F” (190°) for 8 minutes, not spending 8 minutes doing the setting. And how do you measure the “internal” temperature of a spare rib? In any case, 74° seems a little on the low side.
In the end printed out this recipe, which required a screen capture: it seems to come with its own printing routine which produces nonsense. It doesn't directly address air fryers, but that was the experiment anyway: instead, it wants the ribs cooked in an oven at 165° for 40 minutes. With some guesswork, modified the marinade to this:
| quantity | ingredient | step | ||
| 64 g | hoi sin sauce | 1 | ||
| 20 g | tomato sauce (“ketchup”) | 1 | ||
| 30 g | soya sauce (“kicap asin”) | 1 | ||
| 10 g | dark soya sauce (“kicap manis”) | 1 | ||
| 30 g | honey | 1 | ||
| 30 g | rice wine | 1 | ||
| 20 g | lemon juice | 1 | ||
| 16 g | grated ginger | 1 | ||
| 2.4 g | five spice powder | 1 | ||
| 24 g | garlic | 1 | ||
| 800 g | pork spare ribs | 2 |
Mix the marinade, cut the ribs into individual pieces, and marinate at room temperature for 2 hours. I placed them in an air fryer (the “hair dryer”) and grilled at 200° for 20 minutes:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The result? The taste was good, though probably too much garlic, but the meat didn't fall from the bones like it should have done. I shudder to think what results the first recipe above would have produced after only 8 minutes at 190°. I tried giving some another 10 minutes at 160°, but it didn't make much difference. Next time I'll try cooler and longer. Here's my recipe, which will evolve.
| Monday, 25 November 2019 | Dereel | Images for 25 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Davey comes for a week
|
Topic: animals | Link here |
Kim Stanley along with Davey this morning. She's off to hospital for an operation, so we'll have him until she gets out, which is planned for Saturday. Gradually he's making friends with Piccola.
|
Recovery day
|
Topic: general | Link here |
Anke's gone, but we had put lots of things to one side, and today we tried to catch up to normality. Four dishwasher loads since yesterday. I didn't finish my diary for Saturday until this evening.
|
Microsoft problems?
|
Topic: photography, technology, opinion | Link here |
Yvonne had 860 photos to process, made more difficult by her propensity to zoom in postprocessing. And she ran into trouble: the system seemed to hang, page refreshes didn't work. OK, stop DxO PhotoLab and try again. More problems. And then I noticed something strange:
=== grog@eureka (/dev/pts/29) ~/Photos/20191122 1358 -> ruptime
dischord down 0:42
eureka up 75+05:42, 22 users, load 1.87, 1.79, 1.70
euroa up 1+01:26, 0 users, load 0.00, 0.00, 0.00
lagoon up 34+04:11, 1 user, load 0.31, 0.25, 0.18
teevee up 34+04:10, 0 users, load 0.07, 0.14, 0.10
Why was dischord showing as “down”? It was running. OK, try rebooting. It took forever. And starting took even longer. Press the power button: it stopped immediately, rather than requiring the normal 5 second press.
Power on again. Menu: “Start Windows normally” or “Perform Startup Repair (recommended)”. OK, let's see what “startup repair” can do. After a while it came back and suggested recovering to an earlier point in time, presumably some kind of snapshot, but it was too polite to use that terminology. No, I can always try that as a second course of action.
That carried on with the same display as before. After about 5 minutes nothing had happened, so I pressed the thoughtfully provided “Cancel” button. Ha ha, only joking:
|
|||||||||||||||||||||||
OK, leave it run a bit longer. But after 20 minutes I asked on IRC, and Daniel O'Connor told me “format and reinstall, you're fucked”.
Damn! How I hate Microsoft. But I won't give up that easily. After another 20 minutes I powered it down (this time really needing a long press) and started normally. This time it came up completely normally.
So what happened? This is Microsoft, and it's not telling. My guess is:
The only part of the problem that doesn't fit is that ruptime showed it being down. Why? Performance problems, maybe? You wouldn't expect them to hinder the Microsoft equivalent of rwhod for 45 minutes. Maybe it just crashed.
In any case, it brought a couple of things home to me: firstly, I should do more frequent backups of my Microsoft boxen (the last was over 2 months ago, mainly because of configuration difficulties), and secondly, whether I like it or not (I don't), I'll need to upgrade to Microsoft “Windows” 10 soon. And that requires bringing eureka up to date.
|
Nokton experience
|
Topic: photography, opinion | Link here |
For the first time today I took a significant number of photos with my new Voigtländer Nokton 25 mm f/0.95 lens. Almost the only purpose is for shallow depth of field, though clearly it allows for relatively fast shutter speeds. I tried a photo on Saturday night, using the Olympus E-PM1. Not a good idea: there's no adequate focus help, and the best I got was:
|
|||||||||||||||||||||||
|
|
|||||||||
Still, with the right camera the bokeh should be worth it, right? Somehow it wasn't as pronounced as I had hoped for. Today I took a couple of comparison photos with the Nokton at f/0.95, the Leica Summilux 25 mm f/1.4 at f/1.4, and for the sake of comparison the Nokton at f/1.4. The results are hardly any different:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Which is which? It's really hard to say. In fact, the first two are with the Nokton, first at f/1.4 and then at f/0.95. And though it's clear that the trees in the background are out of focus, it's not overly pronounced.
On the other hand, one thing is obvious: it's not clear that the f/0.95 is genuine. The two images at f/1.4 had a shutter speed of 1/5000 s, so the f/0.95 should have been 1/10000 s. In fact, it's 1/8000 s. That suggests only ⅔ stop difference, or f/1.1. I noticed similar things with the 50 mm f/1.4 Super-Takumar decades ago, but at the time I guessed (possibly correctly) that it might have been an artefact of the mirror mechanism. That's clearly not the case here, especially as the same lens creates the same result as the Summilux at f/1.4. In addition, the Nokton is the first lens I have ever seen that has ⅓ stop detents in the aperture ring instead of the ½ stop ones I was used to. But between “f/0.95” and f/1.4 there is only one detent. If that's correct, it's not clear what advantage the Nokton has over the M.Zuiko Digital ED 25mm f/1.2 PRO apart from price.
In fact, the lens does have ½ stop detents. I don't know how I got the impression that they were ⅓ stops.
The other issue is the sharpness of the lens. Here comparison crops of the Nokton and the Summilux:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And that's not even at full aperture: they're both at f/1.4. Here the Nokton at f/0.95, then at f/1.4:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
All in all, rather disappointing. I had planned never to sell a lens again, but I might make an exception in this case.
|
Videos: done!
|
Topic: multimedia, animals | Link here |
Finally finished the transcoding of the video clips we took over the weekend. We now have:
Friday, part 1: Yvonne on Carlotta (25:13).Friday, part 2: Yvonne on Carlotta (16:41).Friday, part 3: Anke on Carlotta (24:34).Saturday: Yvonne on Carlotta (29:00).Sunday: Yvonne on Carlotta (37:35).
It would be nice to put image links here, but they take so long to render that it's impracticable. And together the five videos take up 133 minutes of video, the length of a good feature film.
|
More Google pain
|
Topic: technology, general, opinion | Link here |
While uploading YouTube videos today, I found it convenient to do it from the browser instance running on eureka:0.3. Normally I do it from eureka:0.1 or eureka:0.2.
It's the same IP address and the same browser (ancient version of firefox), but Google doesn't care:
|
|
|||||||||
Read the message sent to your mobile phone, enter the digits shown on the browser into the phone. Done!
Well, not even done. Not one, but two messages:
A new device just signed in to your Google Account. You're getting this
email to make sure it was you.
Check activity
<https://accounts.google.com/AccountChooser?Email=groggyhimself@gmail.com&continue=https://myaccount.google.com/alert/nt/1574648753000?rfn%3D31%26rfnc%3D1%26eid%3D-8117328810543791175%26et%3D1%26anexp%3Dhsc-control_b>
You
And then
Your account groggyhimself@lemis.com is listed as the recovery email for
groggyhimself@gmail.com. Don't recognize this account? Click here
<https://accounts.google.com/AccountDisavow?adt=AOX8kir-QMncSKpOvTyrfZW3hL_M0miqHTjwOjwuxexdxBviKzYjLrR0PD5e&rfn=31&anexp=hsc-control_b>
How about some details? Oh, for that you need to go near a browser:
|
|
|||||||||
What's wrong with this? There are so many things that don't make sense:
Protection is good. But this is mainly irritation.
| Tuesday, 26 November 2019 | Dereel | Images for 26 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Viewing the video clips
|
Topic: multimedia, technology, opinion | Link here |
So now we have a total of five big video clips, nearly 10 GB in size (the values are MB):
=== grog@eureka (/dev/pts/17) ~/Photos/20191125 116 -> du -sc ../201911*/orig/*avi
1733 ../20191122/orig/1.avi
1052 ../20191122/orig/2.avi
1981 ../20191122/orig/3.avi
2578 ../20191123/orig/1.avi
2544 ../20191124/orig/1.avi
9886 total
That's still only a fraction of the size of the originals:
=== grog@eureka (/dev/pts/17) ~/Photos/20191125 117 -> du -sc ../201911*/orig/*MOV
66 ../20191111/orig/4B113692.MOV
35 ../20191119/orig/4B193845.MOV
3800 ../20191122/orig/4B223954.MOV
3799 ../20191122/orig/4B223955.MOV
1771 ../20191122/orig/4B223956.MOV
3803 ../20191122/orig/4B223957.MOV
2409 ../20191122/orig/4B223958.MOV
3799 ../20191122/orig/4B223959.MOV
3808 ../20191122/orig/4B223960.MOV
1541 ../20191122/orig/4B223961.MOV
87 ../20191122/orig/4B223962.MOV
3803 ../20191123/orig/4B233963.MOV
3808 ../20191123/orig/4B233964.MOV
3177 ../20191123/orig/4B233965.MOV
2 ../20191123/orig/4B234108.MOV
3792 ../20191124/orig/4B244136.MOV
3801 ../20191124/orig/4B244137.MOV
2681 ../20191124/orig/4B244138.MOV
3683 ../20191124/orig/4B244139.MOV
15 ../20191124/orig/4B244140.MOV
49670 total
Came into Yvonne's office today to find her displaying the version uploaded to YouTube. Why?
There turned out to be multiple reasons: there's no way to access the local copies from a web browser, we don't know the names (clearly I could have been more imaginative), and... mpv wasn't installed on lagoon, Yvonne's computer.
OK, use mplayer instead. But that displayed at approximately ¼ the normal speed. Why? I didn't check, since I've moved on from mplayer anyway, but I wonder if it's related to the X driver configuration.
OK, install mpv:
=== root@lagoon (/dev/pts/7) /eureka/home/grog 6 -> pkg install mpv
...
New packages to be INSTALLED:
mpv: 0.29.1_12,1
...
[1/1] Installing mpv-0.29.1_12,1...
[1/1] Extracting mpv-0.29.1_12,1: 100%=== root@lagoon (/dev/pts/7) /eureka/home/grog 2 -> mpv /Photos/grog/20191124/orig/1.avi
Shared object "libva.so.2" not found, required by "mpv"
Huh? What happened there? I recall seeing this before, but that doesn't make it any more acceptable.
=== root@lagoon (/dev/pts/7) /eureka/home/grog 9 -> locate libva.so
...
/usr/local/lib/libva.so
/usr/local/lib/libva.so.2
/usr/local/lib/libva.so.2.100.0
But /usr/local/lib/libva.so.2 was no longer there. The locate database is rebuilt every night, so it seems reasonable to believe that pkg removed it when installing mpv. But no mention of that, and no obvious reason. /usr/local/lib/libva.so.2.100.0 was still there, and comparison with other systems showed that /usr/local/lib/libva.so.2 was just a symlink to /usr/local/lib/libva.so.2.100.0, so I could just relink it.
But that wasn't all. There were many other missing libraries:
=== root@lagoon (/dev/pts/7) /eureka/home/grog 15 -> ldd /usr/local/bin/mpv|grep "not found"
libva-x11.so.2 => not found (0)
libva-drm.so.2 => not found (0)
libva-drm.so.2 => not found (0)
libva-x11.so.2 => not found (0)
libdl.so.1 => not found (0)
libdl.so.1 => not found (0)
Remove and reinstall libva? I didn't even consider it. There were clearly other dependencies, and that way madness lies. Instead I did the symlinks... until libdl.so.1, for which I didn't find one. Never mind, there's one on teevee:
=== root@lagoon (/dev/pts/7) /usr/local/lib 22 -> mount teevee:/ /mnt=== root@lagoon (/dev/pts/7) /usr/local/lib 23 -> cp -p /mnt/usr/lib/libdl.so* /usr/lib/=== root@lagoon (/dev/pts/7) /usr/local/lib 24 -> ldd /usr/local/bin/mpv|grep "not found"=== root@lagoon (/dev/pts/7) /usr/local/lib 25 -> mpv
ffmpeg library versions:
libavutil 56.31.100 (runtime 56.14.100)
libavcodec 58.54.100 (runtime 58.18.100)
libavformat 58.29.100 (runtime 58.12.100)
libswscale 5.5.100 (runtime 5.1.100)
libavfilter 7.57.100 (runtime 7.16.100)
libswresample 3.5.100 (runtime 3.1.100)
ffmpeg version: 4.0
mpv was compiled against a different version of FFmpeg/Libav than the shared
library it is linked against. This is most likely a broken build and could
result in misbehavior and crashes.
mpv does not support this configuration and will not run - rebuild mpv instead.
Exiting... (Fatal error)
That's enough. I need to do this properly, which will require a complete pkg upgrade. I shudder to think what other problems that will uncover.
|
Faster mencoder transfer?
|
Topic: multimedia, technology, opinion | Link here |
On IRC, Peter Jeremy suggested alternative processing methods for joining my video clips:
<peterj> groggyhimself: Transcoding video using a CPU is slow. Is there a reason you need to
transcode and can't just use ovc=copy oac=copy ?
That makes sense. But then look at the difference in size: the transcoded files are 80% smaller. I think that makes up for the CPU usage.
|
Another power failure
|
Topic: general | Link here |
Another very short grid power failure this evening at 21:51:20, possibly only a brownout.
| Wednesday, 27 November 2019 | Dereel | Images for 27 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Upgrading lagoon
|
Topic: technology, opinion | Link here |
Yvonne off shopping today, time to upgrade the ports on lagoon, her system. She was gone for 4 hours, and I needed the whole time.
Once again I got:
Installed packages to be REMOVED:
OpenEXR-2.2.1
emacs25-25.3_1,3
ImageMagick-6.9.9.28,1
...
Why does pkg always want to remove Emacs? It happens so regularly that I come to expect it.
And once again it downloaded hundreds of binaries—716 of them, taking about 2 GB of downloads. And once again, when it was done, it reconsidered, and came up with:
Number of packages to be removed: 45
Number of packages to be installed: 85
Number of packages to be upgraded: 576
Number of packages to be reinstalled: 44
The only serious package to be removed was still Emacs. And after performing the upgrade and reinstalling Emacs, things seem to work. I suppose that's something, but the whole process scares me so much that I keep putting it off.
|
More Carpobrotus
|
Topic: gardening | Link here |
While walking the dogs, saw this on the corner of Bliss Road and Harrisons Road:
|
|||||||||||||||||||||||
It proved to be the small kind of Carpobrotus, of which we already have a couple of varieties. Yvonne liked the colour, so now we have a couple more cuttings:
|
|||||||||||||||||||||||
|
More web site overload
|
Topic: technology, opinion | Link here |
Last weekend was interesting for another reason. Syncing my diary to the external web server (www.lemis.com, alias w4.lemis.com) took forever, and then failed:
Connection closed by 45.32.70.18
rsync: connection unexpectedly closed (0 bytes received so far) [sender]
rsync error: error in rsync protocol data stream (code 12) at io.c(226) [sender=3.1.2]
Wed 27 Nov 2019 14:15:07 AEDT
real 25m2.776s
user 0m0.000s
sys 0m0.034s
Took a look at the top display, and once again the load average was up to 140 or so. What caused that? When I could finally access the log files, I found many entries like:
ecs-159-138-158-142.compute.hwclouds-dns.com - - [27/Nov/2019:03:00:06 +0000] "GET /yvonne/photos/Photos.php?dirdate=20190322&size=2&imagesizes=22222222222222222222232222222220222222222222223220222222222222222222222222222220222220 HTTP/1.1" 200 45693 "-" "Mozilla/5.0 (Linux; Android 7.0; FRD-AL00 Build/HUAWEIFRD-AL00; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/53.0.2785.49 Mobile MQQBrowser/6.2 TBS/043602 Safari/537.36 MicroMessenger/6.5.16.1120 NetType/WIFI Language/zh_CN"
ecs-159-138-155-7.compute.hwclouds-dns.com - - [27/Nov/2019:03:00:00 +0000] "GET /grog/diary-mar2018.php?dirdate=20180314&imagesizes=1111111111111111111111111111111111112111111111111111111111111111111112111121111111111111111111111111111211121111212111111111111111111211111111211111111111111111111112 HTTP/1.1" 200 721294 "-" "Mozilla/5.0(Linux;U;Android 5.1.1;zh-CN;OPPO A33 Build/LMY47V) AppleWebKit/537.36(KHTML,like Gecko) Version/4.0 Chrome/40.0.2214.89 UCBrowser/11.7.0.953 Mobile Safari/537.36"
ecs-159-138-152-225.compute.hwclouds-dns.com - - [27/Nov/2019:03:00:06 +0000] "GET /yvonne/photos/Photos.php?dirdate=20190322&size=2&imagesizes=22122221221222222222222222322222222223222222322222222222222222232222222222202222222220 HTTP/1.1" 200 25552 "-" "Mozilla/5.0(Linux;U;Android 5.1.1;zh-CN;OPPO A33 Build/LMY47V) AppleWebKit/537.36(KHTML,like Gecko) Version/4.0 Chrome/40.0.2214.89 UCBrowser/11.7.0.953 Mobile Safari/537.36"
ecs-159-138-153-59.compute.hwclouds-dns.com - - [27/Nov/2019:02:59:49 +0000] "GET /yvonne/photos/Photos.php?dirdate=20190322&size=2&imagesizes=22222222222222222222222222222222222222222222322122222222220022222232222222222222222222221 HTTP/1.1" 200 464229 "-" "Mozilla/5.0(Linux;Android 5.1.1;OPPO A33 Build/LMY47V;wv) AppleWebKit/537.36(KHTML,link Gecko) Version/4.0 Chrome/43.0.2357.121 Mobile Safari/537.36 LieBaoFast/4.51.3"
There are a number of things of interest here:
OK, it looks like it will pass. What do I do in the meantime? A second server, of course! How hard can it be? Using Vultr,
And the reality? It Just Worked! In fact, I didn't even need to change the host name, though it was clearly the right thing to do. By the time I got round to looking at the log files, it was already serving web pages, though it's not clear how I can balance them.
The only puzzling thing was:
=== grog@tyo (/dev/pts/1) ~ 1 -> host www
Host www not found: 3(NXDOMAIN)=== grog@tyo (/dev/pts/1) ~ 2 -> host www.lemis.com
www.lemis.com has address 45.32.13.124
www.lemis.com has address 45.32.70.18
Huh? That looks like something wrong with /etc/resolv.conf, and sure enough:
=== grog@tyo (/dev/pts/1) ~ 3 -> cat /etc/resolv.conf
# Generated by resolvconf
nameserver 108.61.10.10
But why didn't this happen on w4? Because I had never tried it. When I did, it behaved in the same way. So using DHCP has its advantages, but this doesn't seem to be one of them.
The other thing that was surprising was the length of time it took to restore the snapshot. It was 80 GB in size and taken in Los Angeles, but the data was restored in Tokyo. It only took about an hour, which would suggest a speed of about 150 Mb/s, though I suppose they used some aggressive correction.
In passing, it's interesting to note how long an apachectl graceful can take. At 2:05 (UTC) I renamed the log file and ran apachectl graceful. It took 12 minutes for the first new server instances to write to the new log file.
|
Chicken Kiev? Or Kyiv?
|
Topic: food and drink, history, opinion | Link here |
37 years ago today (or tomorrow?), Yvonne moved in with me in Rosbach vor der Höhe. How time flies! In celebration of that she bought some pre-prepared Chicken [à la] Kiev. Or is that Kyiv?
How to cook them? Last time we deep fried them, but now we have the “air fryer”, and it seemed reasonable to use that. Gave them 15 minutes at 180°, but it didn't seem to be enough. Maybe 18 minutes would be a better attempt next time.
That is, if there's a next time. My comment from last time still seems appropriate: cook first, add butter and garlic later.
| Thursday, 28 November 2019 | Dereel | Images for 28 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Web server load balancing
|
Topic: technology | Link here |
So how do you balance web servers? It seems that, for Apache, the answer is mod_proxy_balancer. Problem: my overload had gone away, so I could destroy Tokyo (OK, tyo.lemis.com). At least I know where to look next time. And the obvious thing to do is to install a web server on ffm.lemis.com, which isn't doing very much of use.
|
Lens sharpness again
|
Topic: photography, opinion | Link here |
A few days ago I discovered that the Voigtländer Nokton 25 mm f/0.95 was disappointingly fuzzy at full aperture. Discussion on the M43 Tech Talk Facebook group suggests that this is normal for this lens, and that it should sharpen up nicely by f/2.8 or f/4.
That doesn't make much sense, though, because I have the Leica Summilux 25 mm f/1.4, which is sharper at full aperture. But I've already noted that the Summilux is particularly sharp. How does the Nokton compare with other lenses, such as the Asahi Optical Super-Takumar 50 mm f/1.4? Out to take some comparisons. Here's the Super-Takumar at f/4 down to f/1.4 in 1 EV steps:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And here's the Nokton at f/4 down to f/0.95:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The first thing of interest is that both lenses don't show a full stop difference in exposure between the widest and second widest aperture. The Super-Takumar has a shutter speed of 1/2000 s at f/1.4 and 1/1600 s at f/2—exactly the difference that I noticed over 50 years ago. At the time I guessed (possibly correctly) that it might have been an artefact of the mirror mechanism and the exposure measurements, but clearly that's not the case here. And the Nokton did the same thing: 1/4000 s at f/1.4 and 1/5000 s at f/0.95. Why the difference in speed? Partially, maybe, because of yellowing of the Super-Takumar over the last 50 years, and partially because of the different background due to the different focal length.
And the sharpness? Here are 450x600 excerpts (click once to get original size). Again, first the Super-Takumar:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
And then the Nokton:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
So the Nokton isn't as bad as I had feared; it's certainly better than the Super-Takumar. But it's still difficult to compare. The Super-Takumar has a lot of flare wide open, which may hide the sharpness. Here another attempt. First the Super-Takumar at f/1.4, then the Nokton at f/1.4 (run the cursor over an image to compare it with its neighbour):
|
|
|
|||||||||
|
|
|
|||||||||
And here the Nokton, first at f/1.4, then at f/0.95:
|
|
|
|||||||||
|
|
|
|||||||||
So the Nokton isn't that bad after all. But is it worth it? The difference in Bokeh between f/1.4 and f/0.95 is not overly evident, and for the former I can use the Summilux.
| Friday, 29 November 2019 | Dereel | Images for 29 November 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More gardening
|
Topic: gardening, opinion | Link here |
Mick Solly along this morning again, mainly to finish mowing the lawn for the first time this spring (which ends tomorrow). He was severely hampered by the lack of his ride-on lawn mower, which had failed on him, but somehow got it done anyway with a smaller push mower.
He also transplanted a couple of plants: the Schinus molle (Californian pepper tree) that we planted in July 2018, and which isn't quite dead yet. Here then and last month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
To do so he had to take out the birch that died last year. That wasn't easy. Removing the trunk was: it just broke off. But digging out the roots made a real mess of his garden fork:
|
||||||||||||||||||||||||
Why? It had been planted without removing the plastic pot:
|
||||||||||||||||||||||||
Who did that? . It's probably the reason that it died, though it's not clear why the roots couldn't have fought their way out of the pot.
The Schinus looks pretty sickly, but it does have a number of shoots coming out:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Hopefully it will grow well now.
Also transplanted the Grevillea thelemanniana that we had planted in the middle garden bed five months ago. They don't seem to have been at all happy there, unlike the Grevillea preissii that we planted near the entrance at the same time, so we moved them there:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
Ports upgrade: sting in the tail
|
Topic: technology, opinion | Link here |
The upgrade to lagoon on Wednesday went relatively smoothly—I thought. Then Yvonne told me that she was having difficulties with her firefox. Oh. I had forgotten to stop firefox before upgrading it. In to take a look. Stop firefox. Start firefox. Hangs. That's very similar to problems I had under Microsoft on dischord a couple of weeks ago, and which I couldn't solve. It's been almost almost exactly 15 years since I started using firefox, and I still hate it with a passion.
OK, start chrom* again. Ah, isn't that nice?
|
|||||||||||||||||||||||
It's too polite to state its name, going to the trouble of disabling the window manager in the process. And also too polite to display the obligatory www in the URL! It even changes http: into Not secure. After all the trouble I've gone to make it clear that lemis.com is not a valid URL! Who does it think it is? Ugh!
To be fair, yes, I recall that you can teach it manners; there's some configuration knob to make it use the system decorations. But why should you need to? At the very least it's a significant violation of POLA.
Yes, some people will consider it a matter of taste. But it's not. lemis.com is not www.lemis.com, and there are significant differences. And why does it want to go against what the installation has chosen for window decorations? I know this happens in Microsoft, and so far I've thought that this was just a bug in their implementation. But why would people want to make it difficult to establish what a window is?
It didn't work much better than firefox. Various sites just hung, and I found lots of low-level error messages on the system console, again too polite to say who they were:
|
|||||||||||||||||||||||
The reference in the last line to chrome-search suggests that once again it's chrome. Google has many clever people. How can they stomach this kind of nonsense?
So we were left with Opera, which worked. For a while:
|
|||||||||||||||||||||||
OK, something has gone seriously wrong. Opinions on IRC varied: my fault for not updating packages sooner, my fault for using packages rather than compiling ports, my fault for mixing ports and packages (something that is considered a feature, not a bug, but which I didn't do), my fault for messing around trying to fix library breakage that happened when I tried to install on Tuesday. Arguably there is some merit in this last one.
So: what do I do? By this time it was evening. I bit the bullet and gave her web access on dischord, the Microsoft box, while I ponder how to upgrade the system to something functional. It hurts.
| Saturday, 30 November 2019 | Dereel | Images for 30 November 2019 |
| Top of page | ||
| previous day |
|
Air frying ikan bilis
|
Topic: food and drink, opinion | Link here |
Nasi lemak for breakfast today. And I was out of cooked peanuts and ikan bilis. OK, I've been able to cook them in the “air fryer” in the past, so do so again, in the process reducing the cooking time from 20 to 12 minutes.
It wasn't enough. And the quantities that you can make in that time are pretty minimal. I think I'll give up and use conventional methods.
|
Recovering lagoon
|
Topic: technology, opinion | Link here |
What's the easiest and least disruptive way to recover the mess I made of lagoon? By the time I woke up this morning, the answer was clear: I need a Vulture. Off to create a new, small and temporary virtual machine, this time in Sydney, for no particular reason.
OK,
Create a new vulture, using FreeBSD 12.1, a relatively new release. The IP address brought back memories: 139.180.165.47. That reminds me of my Telstra addresses decades ago, the ones that showed up in The Complete FreeBSD.
Check out the 12-STABLE source tree. Oh. No svn. OK, install it:
# pkg install svn
Updating FreeBSD repository catalogue...
FreeBSD repository is up to date.
All repositories are up to date.
pkg: No packages available to install matching 'svn' have been found in the repositories
Fool! It's called Subversion, not svn! Installed that and started checking out the source tree, for some reason at a snail's pace: it must have taken 2 hours.
Install lots of ports. Since I really needed bash and Emacs, I installed them first (dragging in 124 dependencies in the process). Then the other ones that I recalled. I'll do a full list when I'm relatively sure.
Build a new world. Oh. Dropped a clanger:
c++ -target x86_64-unknown-freebsd12.1 --sysroot=/usr/obj/usr/src/amd64.amd64/tmp -B/usr/obj/usr/src/amd64.amd64/tmp/usr/bin -O2 -pipe -I/usr/obj/usr/src/amd64.amd64/lib/clang/libclang -I/usr/obj/usr/src/amd64.amd64/lib/clang/libllvm -I/usr/src/contrib/llvm/tools/clang/lib/Basic -I/usr/src/contrib/llvm/tools/clang/lib/Driver -I/usr/src/contrib/llvm/tools/clang/include -DCLANG_ENABLE_ARCMT -DCLANG_ENABLE_STATIC_ANALYZER -I/usr/src/lib/clang/include -I/usr/src/contrib/llvm/include -DLLVM_BUILD_GLOBAL_ISEL -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -DNDEBUG -DLLVM_DEFAULT_TARGET_TRIPLE=\"x86_64-unknown-freebsd12.1\" -DLLVM_HOST_TRIPLE=\"x86_64-unknown-freebsd12.1\" -DDEFAULT_SYSROOT=\"\" -DLLVM_TARGET_ENABLE_AARCH64 -DLLVM_TARGET_ENABLE_ARM -DLLVM_TARGET_ENABLE_MIPS -DLLVM_TARGET_ENABLE_POWERPC -DLLVM_TARGET_ENABLE_SPARC -DLLVM_TARGET_ENABLE_X86 -DLLVM_NATIVE_ASMPARSER=LLVMInitializeX86AsmParser -DLLVM_NATIVE_ASMPRINTER=LLVMInitializeX86AsmPrinter -DLLVM_NATIVE_DISASSEMBLER=LLVMInitializeX86Disassembler -DLLVM_NATIVE_TARGET=LLVMInitializeX86Target -DLLVM_NATIVE_TARGETINFO=LLVMInitializeX86TargetInfo -DLLVM_NATIVE_TARGETMC=LLVMInitializeX86TargetMC -ffunction-sections -fdata-sections -gline-tables-only -MD -MF.depend.AST_ASTImporter.o -MTAST/ASTImporter.o -fstack-protector-strong -Qunused-arguments -fno-exceptions -fno-rtti -gline-tables-only -std=c++11 -stdlib=libc++ -Wno-c++11-extensions -c /usr/src/contrib/llvm/tools/clang/lib/AST/ASTImporter.cpp -o AST/ASTImporter.o
*** Signal 9
Stop.
make[5]: stopped in /usr/src/lib/clang
How did that happen? Who sends a SIGKILL during a build? It was repeatable, and it wasn't until some time later that the reason occurred to me:
Nov 30 02:15:54 vultr kernel: pid 74703 (c++), jid 0, uid 0, was killed: out of swap space
Nov 30 02:15:55 vultr kernel: pid 74564 (make), jid 0, uid 0, was killed: out of swap space
Yes, this system didn't have any swap, and it also didn't have much memory. Next time I do this I'll create a swap file.
After that partial success, packed the entire system into a compressed tarball, slightly under 10 GB in size, and moved it to lagoon at a quite respectable speed:
=== grog@lagoon (/dev/pts/6) /var/tmp 8 -> scp 139.180.165.47:/var/tmp/rootfs.tar.bz .
rootfs.tar.bz 100% 9085MB 2.2MB/s 1:08:25
That's an average speed of about 17.6 Mb/s, not bad for a 25 Mb/s downlink. I had speeds of up to 2.9 MB/s (23.2 Mb/s). Certainly the fact that it was in the same network helped: in fact, over half the latency between the systems was across the NBN link.
Then unpack, restart the build, destroy the vulture (which cost me $0.09), and give up for the day.
|
Garage sale. Where?
|
Topic: general, technology, opinion | Link here |
Yvonne back from Ballarat with the news that there was a 5 town garage sale again, like they had last year. OK, details please. It's on the DEREEL OUTPOST Facebook group. OK, take a look:
|
|
|||||||||
That's not exactly what I expected. Oh, says Yvonne, the addresses are on the “rokewood, dereel, cape clear, cressy, skipton district buy swap and sell” list. OK, take a look. Members only. Apply for membership, but of course nothing happened. And Yvonne was out somewhere, so she couldn't help.
More looking around. There's a map on the DEREEL OUTPOST list:
|
|
|||||||||
But it's not legible. Where is it? They don't say. Could that be the Dereel Hall? Checked with my photo yes, it is. But it doesn't help me much with the search for the garage sales.
Finally Yvonne came back, and we found the entry on the “rokewood, dereel, cape clear, cressy, skipton district buy swap and sell” list. A fuzzy, cropped mobile phone photo of most of a list specifying only addresses. Ugh!
Anyway, Yvonne wanted to look in Rokewood, where there were two sales, including a “car boot sale” on the RSL grounds.
OK, where's my GPS navigator? Looked at the window of the car. Not there. Ah, there it is:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Somehow that completed today's frustration—I thought. To Rokewood. Nothing. The RSL grounds were deserted, and the other address also seemed inactive. Back to Dereel, checking at Ruth Uren's place. She had a sign out the front, nicely covered. No more garage sale today: it was 15:45.
On to the Dereel Hall, where yes, indeed, I found the map:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I wonder who uses maps like this. Every man and his dog has a mobile phone, and that's easier. But yes, by comparison with other such maps that I've seen, it doesn't look bad.
So, in the end: an afternoon of extreme frustration, with nothing else to show for it.
|
Garden progress?
|
Topic: gardening, opinion | Link here |
Yesterday Mick Solly, the gardener, commented on the number of partially dead plants that we have. He's right. It doesn't make sense, especially in this case (a Hebe):
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
They're all in a row (which is longer), and only the middle one is suffering. And there are others:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
On the other hand, other plants are doing well. The Clematis General Sikorski is actually flourishing and flowering better than the other one, the Edo murasaki on the trellis:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
And the „Gruß an Aachen“ rose is flowering better than I have seen it since we have been here:
|
||||||||||||||||||||||||
But it also looks as if we'll get a flower from our Strelitzia nicolai:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I've never seen one before.
|
Stones Road changes
|
Topic: Stones Road house, photography, opinion | Link here |
It's been just over 6 years since we first saw the plot of land in Stones Road. At the time, I took a panorama photo from the north-east end of the plot:
|
|||||||||||||||||||||||
Today was house photo day again, so I thought it would be fun to try to reproduce it. It was more difficult than I thought: the best indicators of position are the trees, but they have changed considerably in that time. Here's my best guess:
|
||||||||||||||||||||||||
It's surprising how much has changed in that time. Looking west, we have:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
