|
| Greg's diary |
| April 2020 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Wednesday, 1 April 2020 | Dereel | Images for 1 April 2020 |
| Top of page | ||
| next day | ||
| last day |
|
PIXIO or Pixem?
|
Topic: photography, multimedia, opinion | Link here |
Chris Bahlo is interested in the PIXIO “Robot Cameraman”, and she asked me a number of questions yesterday, including whether she could use it with her Olympus OM-D E-M10 Mark II. That's a simple answer: Yvonne used it in January, and it worked well. The only issue is the sound, as I noted at the time: it doesn't have an input for external microphones. That's the only reason I bought the (much more expensive) E-M5 Mark III for Yvonne.
But Chris had considered using a mobile phone if the E-M10 wasn't compatible. That would require a PIXEM, so I took a look at that. It appears to be identical to the PIXIO except for the mount:
The PIXIO has firmware that controls the zoom of some Sony video cameras, clearly something that influences a significant part of the price. Like cameras derived from still cameras, mobile phones don't have an external zoom control—in fact, most if not all don't have zoom at all. But the PIXIO and PIXEM cost the same, and the only adaptation I see for mobile phones is the phone mount. And of course, getting back to the original issue, mobile phones don't have inputs for external microphones either, so there are only down sides.
But why vertical? It's clear that the normal way to hold a phone is vertical, to match the shape of a human face. But why do people take photos and videos that way? Particularly videos don't look good like that. Can't phones take landscape oriented images?
Off to try things out with my three mobile phones. I couldn't work out how to take videos with the Samsung GT-I9100T: the app seems similar to the other two, but I couldn't select the symbol for video. Potentially I'm missing some setting elsewhere, but the phone is 10 years old or so, and I couldn't be bothered. More to the point are the other two:
=== grog@eureka (/dev/pts/65) ~/Photos/20200401/orig 903 -> mpid *.mp4 *.3gp
VID_20200401_174331.mp4 VIDEO: [H264] 1920x1080 24bpp 29.583 fps 16816.4 kbps (2052.8 kbyte/s)
VID_20200401_174400.mp4 VIDEO: [H264] 1920x1080 24bpp 29.583 fps 16846.7 kbps (2056.5 kbyte/s)
VID_20200401_174519.3gp VIDEO: [MP4V] 1280x720 24bpp 29.537 fps 11357.0 kbps (1386.3 kbyte/s)
VID_20200401_174540.3gp VIDEO: [MP4V] 1280x720 24bpp 29.537 fps 12399.7 kbps (1513.6 kbyte/s)
The Nokia 3 has a video resolution of 1280×720, while the Nokia 5.1 Pro has a resolution of 1920×1080—the first potential advantage of the 5.1, though of course I wouldn't use it except in an emergency. But why not 1080×720 in landscape mode? With mplayer the landscape versions displayed correctly, while the portrait clips were on their side. I don't even know how to correct that. So why do people predominantly take portrait clips?
And while I was at it, also took some photos. The Nokia 3 cheated by making a noise but not taking a photo, so I only got one. Attempts later were confusing: I seemed to get a thumbnail, but not a real photo. I'll need to check on that one later.
What I did get was:
Samsung I9100T:
|
|||||||||||||||||||||||
Nokia 3:
|
||||||||||||||||||||||||
Nokia 5.1 Pro:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
It's interesting that the only image that wasn't rotated correctly was a portrait image from the I9100T. But its Exif data included:
Orientation : Rotate 90 CW
So arguably this was an issue with my workflow. But the others rotated the image automatically. The Nokia 5.1 Pro had no Orientation information in the Exif data, and the Nokia 3 had:
Orientation : Horizontal (normal)
A couple of other things of interest: despite the format of the phone itself, the images have a 4:3 aspect ratio, though the same sensor produces a 16:9 aspect ratio for the videos in the two that I was able to take. The other thing that is very obvious is the extreme difference in quality between the three phones. The Nokia 5.1 Pro is almost acceptable.
It wasn't until later that I noticed that I hadn't looked at the PIXEM images very carefully: yes, the predominant feature is vertical, but that's just to mount the phone horizontally. I also discovered that some mobile phones do indeed support zoom, and the PIXEM makers claim to support it, though they're very cagey about which phones: I couldn't find out which. And on closer examination the PIXEM does indeed mount the phone in landscape orientation:
But those are just details, and I discovered them too late to make any difference.
|
More teevee pain
|
Topic: technology, opinion | Link here |
While playing around with the photos on teevee this afternoon, tried:
=== root@teevee (/dev/pts/8) /usr/src 23 -> emacs
(emacs:80274): dbind-WARNING **: 15:17:03.053: Error retrieving accessibility bus address: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
What's that? The DISPLAY was set to eureka:0.1, and it only happened with Emacs under root. And it only happened since the reboot yesterday.
It was repeatable, sort of. While scratching my head, this message appeared:
(emacs:80314): dbind-WARNING **: 15:21:08.304: Error retrieving accessibility bus address: org.freedesktop.DBus.Error.TimedOut: Failed to activate service 'org.a11y.Bus': timed out (service_start_timeout=120000ms)
After that, it worked. But what's the issue here? At which end? On the face of it, it appears to be an issue with D-bus, which I'd rather avoid, like all programs that only half think out colour schemes.
| Thursday, 2 April 2020 | Dereel | Images for 2 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Relativizing Android file access ease
|
Topic: technology, photography, opinion | Link here |
Downloading yesterday's photos from the mobile phones was interesting. Mounting the file systems on teevee is clunky. I haven't found a way to integrate it into /etc/fstab, so I need to start things manually, not helped by at least flachmann continually disabling its FTP server.
But even when things were mounted correctly, I had difficulties:
=== root@teevee (/dev/pts/11) /flachmann/DCIM/Camera 27 -> l
-rw-r--r-- 1 root wheel 4,657,086 1 Apr 18:43 IMG_20200401_174318.jpg
-rw-r--r-- 1 root wheel 5,463,532 1 Apr 18:43 IMG_20200401_174323.jpg
-rw-r--r-- 1 root wheel 42,016,296 1 Apr 18:43 VID_20200401_174331.mp4
-rw-r--r-- 1 root wheel 34,467,850 1 Apr 18:44 VID_20200401_174400.mp4=== root@teevee (/dev/pts/11) /flachmann/DCIM/Camera 28 -> mv * /Photos/grog/20200401/orig/
mv: fastcopy: read() failed: VID_20200401_174331.mp4: Input/output error
mv: fastcopy: read() failed: VID_20200401_174400.mp4: Input/output error=== root@teevee (/dev/pts/11) /flachmann/DCIM/Camera 29 -> l
-rw-r--r-- 1 root wheel 42,016,296 1 Apr 18:43 VID_20200401_174331.mp4
-rw-r--r-- 1 root wheel 34,467,850 1 Apr 18:44 VID_20200401_174400.mp4
The problem was repeatable, but I was able to download the files with ftp. A problem with the FreeBSD-side software? Hard to say, especially with the unspecific error message. But I downloaded files of similar size from taskumatti with no difficulties:
=== root@teevee (/dev/pts/11) /taskumatti/DCIM/Camera 39 -> l
...
-rw------- 1 root wheel 39,781,868 31 Jan 17:49 VID_20200131_164903.3gp
-rw------- 1 root wheel 18,844,144 1 Apr 18:45 VID_20200401_174519.3gp
-rw------- 1 root wheel 22,796,546 1 Apr 18:45 VID_20200401_174540.3gp
That was running WiFi FTP Server, while flachmann was running WiFi File Transfer in FTP mode. Maybe that points towards the problem, but since it's becoming clear that flachmann will be returned, I'm not too worried.
The other issue was with the photos on taskumatti. As I noted yesterday, I only got one:
-rw------- 1 root wheel 1,214,305 1 Apr 18:45 IMG_20200401_174512.jpg
That's the only one in /DCIM/Camera. But in /DCIM/.thumbnails, tastefully encoded with a different method, there's:
-rw------- 1 root wheel 43,125 1 Apr 18:45 1585723535361.jpg
-rw------- 1 root wheel 56,947 1 Apr 18:45 1585723558025.jpg
-rw------- 1 root wheel 53,111 2 Apr 13:02 1585789336665.jpg
-rw------- 1 root wheel 49,137 2 Apr 13:02 1585789343318.jpg
What are those numbers? They look like time_t timestamps, and they are, sort of. They appear to be milliseconds since the Epoch. Removing the last 3 digits, the first one has the time:
=== root@teevee (/dev/pts/11) /usr/src 60 -> date -r 1585723535
Wed 1 Apr 2020 17:45:35 AEDT
And sure enough, it's a thumbnail of the one image I got, /DCIM/CameraIMG_20200401_174512.jpg. But where did the other three go? They're the photos I tried to take, but they never got stored in /DCIM/Camera. Space constraints? No, I have 11 GB free. Did the server somehow set the permissions on /DCIM/Camera incorrectly? How do I find out. The good news is that I'd have to be in a real fix to want to take real photos with the phone.
|
More clematis
|
Topic: gardening | Link here |
My long-suffering clematis “General Sikorski” continues to recover:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Hopefully it will continue to recover.
|
Pregnant goldfish?
|
Topic: animals | Link here |
For some time, subjectively over a month, we've had our eye on this goldfish:
|
|||||||||||||||||||||||
Why is it so fat? Clearly it can't be pregnant, but do they fill up with eggs before releasing them? If so, this one must be good for hundreds of offspring. Or is it some disease?
|
Steak and kidney pie again
|
Topic: food and drink, opinion | Link here |
Individual steak and kidney pies for dinner this evening. I've been refining the baking time for some time: it needs to be well browned on the top, but the pastry in the form should also be cooked through. Last time I decided to cook at a lower temperature (190°) and with heat from below, and expected them to be ready in 20 to 30 minutes. After 16 minutes they looked like this:
|
|||||||||||||||||||||||
That's a long way from being cooked: the pastry is still translucent. So I changed to heat from above, still at 190°, and after 30 minutes had:
|
|||||||||||||||||||||||
That's barely OK, and the pastry underneath, though better, was still not completely cooked. What do I do next time? 210° with circulation only?
| Friday, 3 April 2020 | Dereel | Images for 3 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
SARS-CoV-2: A life changer?
|
Topic: health, general, opinion | Link here |
The progress (if that's an appropriate word) of SARS-CoV-2 and the associated COVID-19 has been amazing. Only a little over a month ago the WHO was saying that it could still be possible to avert a pandemic. Now half the world is locked at home. How different is that?
Well, for me, not at all. I've been voluntarily “locked at home” for nearly 30 years now, with the exception of the extensive travel that I did in the first half of that time. When I started working from home (15 January 1991), I thought that it would only be a matter of time before “everybody” did it. Well, clearly not people who have to interact with objects, which makes up a large proportion of the population. But most technical people can work from home. Think of call centres: wouldn't it be so much easier for them to work from home? But no, companies don't like the lack of supervision that that implies.
What I didn't expect, though, was that many people suffer when working from home. They miss the company. And that in modern times, when communications are so much easier!
But what about people who have to work on site? Next week Mick Solly is due to come and do the garden. Is he allowed to? Yvonne tells me that we're under “Stage 3” restrictions, whatever that means. This afternoon I spent something like 15 minutes trying to find out what that meant, in the process tripping over things like an ABC News article discussing what they could look like. But elsewhere I found references that suggest that they're in place, including on the ABC site.
But there must be some official site that states clearly what is allowed and what isn't. Went looking for a gov.au site, with low signal to noise ratio, like this one. With some further searching, found this page, which appears to be official, and is very specific in some cases—like the $1,652 on-the-spot-fines, which leave me wondering how they will be enforced if the person in question doesn't have $1,652 on the spot, and how they came to that strange sum (there's also $9,913 for businesses, which, by $1, isn't even exactly $1,652 × 6). On the other hand, it's vague in places where it's important. When are you allowed to leave your house? All they say is (layout and (lack of) punctuation original):
you should only be outside for one of the following 4 reasons:
What does that mean? Much of it is straightforward enough, but what are compassionate needs? How do I parse “exercise in compliance with the public gathering requirements”? I've never done exercise in any kind of compliance, and I can exercise my compliance at home if I wish to. What's “work”? I'd guess that Mick's work applies, especially since he's not exactly in contact with other people or places that could be infected.
And why did it take me 15 minutes to find even this minimal information? There should be a link to it (or, preferably, something better) on every page related to the outbreak.
And then I had heard of restrictions on interstate travel. Do they exist? If so, why aren't they mentioned? Potentially they're a federal restriction, but do I need to go looking for separate federal restrictions? What about in other states? Doubtless the restrictions are different. People, can't you even present the requirements clearly, completely and unambiguously? Instead we get silly terms like “stage 3”, which, as I have seen, can mean “You are not permitted to use water outside on the 31st day of the month”.
|
DPI debunked
|
Topic: photography, opinion | Link here |
One of the silliest pseudo-measurements that I have seen for photos is Dots per inch. Apart from enshrining archaic units of measure, it has no meaning for a digital image. But some people swear by it; I've seen claims that setting DPI correctly is the most important thing in photo processing.
Today I found a page that agrees with me wholeheartedly. It must be right!
| Saturday, 4 April 2020 | Dereel | Images for 4 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Miserable weather
|
Topic: general, Stones Road house, opinion | Link here |
The mild weather has finally given up. Got up this morning to find that 36 mm of rain had fallen overnight, making a total of 56 mm so far this month. And the door to the verandah had misted over, something that I had never seen in Germany, where it's much colder:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Thank you, JG King, for the antediluvian single-glazed panes.
|
Wide angle distortion
|
Topic: photography, opinion | Link here |
How do you take a really wide angle view of an area? The traditional approach is with a rectilinear lens like the M.Zuiko DIGITAL ED 7-14 mm f/2.8 PRO:
|
|||||||||||||||||||||||
Yes, there's some distortion visible on the left, where the tops of the doors are higher than the bottom. But that only has a horizontal field of view of 72°. We can do better:
|
|||||||||||||||||||||||
Yes, the field of view is much wider, at 95°. But oh, horror, that's a fisheye projection, full of curved lines!
OK, let's “defish”, converting most of it to rectilinear:
|
|||||||||||||||||||||||
That's not distorted, right? Those three cabinets (left, middle, right) all really have the same width. There's a good reason why there are no really wide angle rectilinear lenses.
Of course, the correct approach is an equirectangular or cylindrical projection (taken later):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Which was taken with which lens? The width shows it, not the perspective, which is the same. The first image was stitched from 3 images taken with the Olympus M.Zuiko Digital ED 8mm f/1.8 Fisheye PRO, and the second from 4 images taken with the M.Zuiko DIGITAL ED 7-14 mm f/2.8 PRO. But the native projection of the lens is no longer important.
|
Understanding COVID-19
|
Topic: health, opinion | Link here |
Coursera is offering a couple of free courses on COVID-19, both from Johns Hopkins University, the epicentre (note the modern word) of COVID-19 research. I signed up for COVID-19 epidemiology, which was due to start on 31 March, but I had difficulty accessing it. Finally yesterday evening I got a mail message:
From bounce@t.mail.coursera.org Fri Apr 3 20:33:44 2020
Date: Fri, 03 Apr 2020 09:33:34 +0000
Welcome to Fighting COVID-19 with Epidemiology: A Johns Hopkins Teach-Out.
I'm so happy you will be joining us! The Teach-Out begins on March 31st at
10:00 AM EDT, but you can also join at any time after that.
So clearly they had some delays. Wouldn't it be nice so say so? Off to watch the course. I can forgive them for having a few rough edges, but somehow that went too far. One example, relating to monkey pox, specifically the 2003 Midwest monkeypox outbreak:
|
|||||||||||||||||||||||
The material had already clarified that the categories are listed in order of seriousness: suspected, probable, confirmed. That makes perfect sense. But look at the criteria: a suspected case requires a fever and unexplained rash, and then at least two other symptoms. A probable case only requires a fever or a vesiculopustular rash! That doesn't make any sense—in fact, so little that I still can't decide what the correct criteria should be. Probably you need fever and a vesiculopustular rash for a probable case.
One bad mark. Later comes what appears to be an exercise:
|
|||||||||||||||||||||||
What does ID mean? What relevance does the date have? What is today's date? Why so many patients with no symptoms of monkey pox, and none with the symptoms? Does this example even relate to monkey pox? He doesn't say. But that doesn't matter, because if it's an exercise, it's broken.
Should I continue? I'm very disappointed. Yes, things are hectic at the moment, but I would have expected more accuracy from Johns Hopkins.
|
Backing up videos
|
Topic: multimedia, technology, opinion | Link here |
It's been just over 3 years since we gave up using broadcast TV and switched to the World-wide web. We quickly acquired a large collection of various series, mainly from German TV broadcasters. And since they have a tendency to broadcast repeats out of sequence, we kept the entire series even after we watched them.
That's a recipe for bloat. I started off with a 3 TB disk and then replaced it with (yet another) 8 TB disk. And it's filling up too.
Looking at the older series, it's clear that the resolution leaves something to be desired. 200 GB of films with a maximum resolution of 720x480, significantly below the resolution of standard TV broadcasts. Do we even want to watch something that fuzzy? Somehow 960×720 is the lowest acceptable resolution nowadays. We could probably delete the lot, but I hate doing that.
OK, dragged out a 2 TB external disk, one of my older photo backup disks. Surprisingly, there were 164 GB of files on the disk that aren't on my current disks. Hopefully they're just images that I have reprocessed, but how can I be sure? Went through them with
=== root@eureka (/dev/pts/11) ~ 569 -> mklinks merge /Photos/grog /photobackup/Photos/grog
That replaced all identical files on the disk with symlinks to the version on disk, and it's what showed me the size of the differing files:
=== grog@eureka (/dev/pts/65) ~ 1009 -> du -s /photobackup/Photos/
163850 /photobackup/Photos/
But how do I keep track of what I have backed up? I did this once before, and I don't know what happened to the disk.
| Sunday, 5 April 2020 | Dereel | Images for 5 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The pain of time change
|
Topic: general, photography, technology, opinion | Link here |
Daylight Saving Time ended today, time to put the clocks back. Now that even Microsoft understands DST, it's not as difficult as it used to be. But there were still 12 clocks to reset: the (conventional) oven, two microwave ovens, three analogue clocks and six cameras.
OK, three of the cameras have an 802.11 (“Wi-Fi”) link, and the Olympus OM-D E-M5 Mark III even has a Bluetooth link. With these links, at least in theory, you can set the time to a 1 second granularity simply by connecting. The other three can only set time to a granularity of 1 minute.
But how do you connect? Not for the first time I had problems. In the past I had established that it requires you to disconnect from any other 802.11 network, so you had to first go into the Settings menu, which conveniently hid all the saved connection information under “Saved networks”, though for some reason it was happy to display connection details for the photovoltaic inverter:
|
|
|||||||||
|
|
|||||||||
I had hoped that the new Nokia 5.1 Plus might do better, but there was no difference. And this time I needed to type in the 8 digit “password” for each camera, something that I hadn't had to do before. There used to be some configuration parameter for that, but I get the impression that that was too convenient, and that they have removed it.
OK, select the Saved network. Cancel or Forget? People, are you trying to annoy me? I still haven't understood how to tell it to connect. Maybe it does it automatically in its own slow time. At some later point I was connected, but I still don't know how.
So: two cameras (OM-D E-M1 Mark I and E-M1 Mark II) updated. And the Olympus OM-D E-M5 Mark III? I've done it before, but this time I couldn't find out how to start the “Wi-Fi” connection from the camera. OK, RTFM:
Select [Device Connection] in the > Playback Menu and press the Q button.
· You can also connect by tapping [loudspeaker symbol] D in the monitor.
Now isn't that obvious? Playback menu. Why didn't I think of that? More to the point, why is there no [loudspeaker symbol] D on the monitor? That corresponds to the WiFi on the other cameras. But it's not there. The manual doesn't explain how to display it again, and in the end I gave up and climbed down into the menus. After all, OI.Share is as good as useless for anything else: apart from the pain of setting it up, as well as losing connectivity to the rest of the world, it can't even take normal photos. It insists on JPEG in a different format from what it normally saves, even if I select raw images.
After that, things worked about as well as I could have hoped for. A quarter of an hour! Am I the only person who finds the functionality unacceptable?
|
Understanding Olympus functionality
|
Topic: photography, opinion | Link here |
As if I hadn't had enough pain with Olympus already today, found a message on the M43 Tech Talk Facebook group. A user had bought an Olympus OM-D E-M1 Mark III and was complaining because he couldn't do focus confirmation the way he used to do on his Canon. It seems that on (at least some) Canons, if you push the AEL/AFL button equivalent while in manual mode, it enables something that will beep when some part (which?) of the image is in focus. Without focus peaking, that's not a bad function. But user wanted his E-M1 to do the same thing, and it seems that the E-M1 Mark III (and also the E-M1X, but not any other model) has a default setting to “mode 3”, which assigns autofocus to this button.
That's interesting in itself. A couple of months ago I noted that this mode 3 was very useful, and I have been using it forever. But user thinks it's wrong. That's not manual focus!
A lively discussion ensued, in which it became clear that I had to modify my focus assist page. But the most interesting thing, apart from the fact that user couldn't accept the fact that the AF mode is helpful, and that he can ignore it if he doesn't like it, is that Olympus seems to have decided to make mode 3 the default for newer cameras. Clearly user is in the minority, and that's understandable: if you don't want MF to have this added functionality, what do you allocate to that button?
| Monday, 6 April 2020 | Dereel | Images for 6 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Red Energy responds
|
Topic: Stones Road house, general, opinion | Link here |
Call on the phone this morning from <mumble>, whose name, spelt out, proved to be Pratim, the complaints resolution man at Red Energy. He wanted to talk about the complaint I made to the ombudsman on Friday.
So why was he calling me? You would expect that he would have first read in the voluminous mail exchange that I wanted to communicate by email (that was one of the points I raised in my complaint), and the confirmation from the ombudsman included the text (no idea where the * come from):
*Preferred method of contact:* Email
But no, it seems that the mail he received from the ombudsman asked him to call me on the phone or send email. Sometimes I wonder what the purpose of these fields is.
Agreed that he would check the mail and then get back to me. At least he sounded a little more sensible than most of the people I've dealt with.
Only a little later he replied with an enormous PDF of my last bill—11 MB for 3 pages—and some indication that he had read my mail:
Why phone? What is the purpose of doing everything by phone? Later I got another message telling me that they have applied the rates agreed (in other words, accepting my complaint), and mentioning a revised bill, to be received shortly, along with another indication that I would have to call on the phone to reinstate my direct debit authorization. No bill arrived today.I have also requested your direct debit to be cancelled. Please note that if you wish to reinstate it we will require your consent which means a we will need to speak to you over the phone to obtain it.
|
Repotting hibiscus
|
Topic: gardening | Link here |
I put a small pot with a cutting of my Hibiscus rosa-sinensis “Uncle Max” in Yvonne's bedroom a few months back. It liked the position, and has grown to over 2 metres, the tallest we've ever had:
|
||||||||||||||||||||||||
But it has never flowered, and clearly it needed repotting:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Surprisingly, Yvonne didn't want it back, so now we have two hibiscus in the lounge room:
|
|||||||||||||||||||||||
The old one on the right is gradually recovering. It's nearly 12 years old now. How old do Hibiscus get?
|
|||||||||||||||||||||||
It's the original that Uncle Max gave to me on my 60th birthday, so I don't want to just throw it out.
| Tuesday, 7 April 2020 | Dereel | Images for 7 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Updating eureka, attempt 15
|
Topic: technology, opinion | Link here |
I last updated eureka, my main machine, well over 4 years ago:
FreeBSD eureka.lemis.com 10.2-STABLE FreeBSD 10.2-STABLE #2 r290972: Wed Nov 25 11:38:38 AEDT 2015 grog@stable.lemis.com:/usr/obj/eureka/home/src/FreeBSD/svn/10/sys/GENERIC amd64
I've had a number of attempts to update it both before and after then, but there are a number of issues. Firstly, I hate to reboot. Now that we have the photovoltaic electricity installed, we no longer have power outages, and eureka has been up for 7 months, with promise of staying up much longer.
Then my X configuration is bizarre, and after 30 years of experience, X is still not as reliable as I would like. Upgrading my configuration to a new system has almost always given me issues, including last time.
Given the rarity of the event, should I perform hardware upgrades? eureka's motherboard is over 6 years old. If you believe Moore's law, modern processors should be 8 times as fast as then. Or is that 16 times? Depends on the law. But in fact there isn't that much difference. Yes, the AMD Ryzen Threadripper 3990X is the current top runner with 78,319 CPU marks (but only 2,607 per thread), compared to 6,881 for my Intel i7-4771 (which, I'm sure, was once rated only slightly below 10,000), but it wasn't the fastest in those days, just one that I considered affordable. The Threadripper costs $4,000, far more than what I would want to pay. And surprisingly the single thread rating of the Intel is 2,206, barely any different from the Threadripper. The fastest CPU under $400 US is currently another Threadripper, the AMD Ryzen Threadripper 1920X, rated at 24,489 CPU marks, not really worth the trouble. So I'll stick with the current processor and motherboard.
And what about disks? When I bought the current eureka, I included a 120 GB SSD for the root file system. There was no obvious difference in performance, and it died far too early. But clearly SSDs are the way of the future, and the prices have dropped significantly. It looks as if a 1 TB SSD would be a good idea: that would allow me to put just about everything I access frequently on the disk.
But looking at what's available, it seems that SATA SSDs (what I had last time) are a bad idea: they're limited to about 500 MB/s by the interface. The alternatives of PCIe and M.2, both of which can offer up to 4 GB/s.
M.2 requires dedicated slots on the motherboard, so that would require a different motherboard. But PCIe seems OK, and I have spare slots on the motherboard. But it's not clear that there's enough space between them and the graphics cards:
|
|||||||||||||||||||||||
That would require trying things out and being prepared to fail.
The real issue, though, is: how do I manage the upgrade so that I can continue to work even if something goes wrong? That's the real question. In the past, from 1992 until 2008, I ran two separate machines, joined together with x2x. That had the advantage that I never needed to take both systems down.
And maybe that's what I should be doing now. Many steps: split into two systems (in the process reviving dereel), upgrade dereel, upgrade eureka when dereel is working well, and at some later time put a 1 TB fast SSD in eureka. Motherboard upgrades postponed to some future date.
So today I found an old system (probably the old eureka) with an AMD Phenom 9550 and 6 GB of memory, and looked around for some spare disks with enough capacity. I have plenty of old disks, but most of them are small:
|
|||||||||||||||||||||||
To my surprise, I found 2 1 TB disks (one marked “flaky”), a 2 TB disk, a 3 TB disk and even a 6 TB disk, the last out of teevee only 10 months ago. It'll make a good second disk, which I can mount on eureka or dereel.
For today I took the good 1 TB disk, which proved to have come from lagoon, and booted it as eureso:
=== root@eureso (/dev/pts/1) ~ 55 -> df
Filesystem 1048576-blocks Used Avail Capacity Mounted on
/dev/ada0p2 99,176 35,625 55,616 39% /
/dev/ada0p4 99,176 0 91,242 0% /destdir
/dev/ada0p5 705,709 180,997 468,255 28% /home
OK, first copy the root file system from eureka to eureso. While that was running, though, decided that it was high time to start again from scratch. I've been updating the system on eureka for at least 13 years, and some parts possibly as long as 30 years (starting with allegro in May 1990). I've already established that there are some strangenesses in eureka's mail configuration that I can't work out: after removing all traces of some aliases from the configuration, the server still accepted mail for them. I can't even guess where the information is stored. So starting from scratch would fix this and potentially a number of other issues.
| Wednesday, 8 April 2020 | Dereel | Images for 8 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dual server install, next step
|
Topic: technology, opinion | Link here |
As planned yesterday, continued with my installation of a second system. The machine was already running an old instance of lagoon, masquerading as eureso, and it had a complete copy of eureka's root file system on another partition. But that wasn't what I wanted: I wanted a clean new install.
How do you do that with FreeBSD? I used to give classes in this sort of thing, but they started with a boot floppy disk, something mercifully obsolete. Nowadays we have bootable DVD images and bootable memory sticks. But that's not what I want; I just wanted a bootable root file system, and there seems to be no way to download one.
So I downloaded a DVD image and created a new virtual machine, then copied the file system to my new machine.
Various things happened. My first attempt to reboot failed: instead of rebooting, it hung in the “beep” that precedes rebooting. Hardware issues? It wasn't made any easier by the fact that I had connected the display to my eureka:0.0 monitor, which switches back to a different input when a display goes away. So I have no idea.
Made some process, in the process coming to the realization that I had migrated to real hardware (especially if it's flaky) far too early: I don't really need to take that step until almost everything is installed. And when I looked later on and discovered that the system had hung again, I decided that it was time to give up. Tomorrow I continue with a VM, where I can make a number of clean snapshots.
|
Where's my electricity bill?
|
Topic: Stones Road house, general | Link here |
Two days ago Red Energy accepted my complaint, told me a new sum and said that the adjusted invoice would be arriving soon. But so far nothing has arrived. A new message to their complaints person, and got a response telling me that it would be coming by snail mail. Also a response to my query about why I need to call up to reinstate the direct debit authority:
You can most certainly do it in writing or even online.
So why the repeated warning that I would have to make a phone call?
|
teevee crash!
|
Topic: technology, multimedia | Link here |
As if that wasn't enough, when I went into the lounge room to watch the news on teevee, it hung. Why? I couldn't get the firefox instance to stop. Into the office, where I could access teevee with no problems. The firefox process was stuck in a STOP state, something that I usually associate with a process dumping core. But I had disabled core dumps on teevee, and it was maxing out a processor core.
Dammit, restart the X server. But that was hanging too, waiting on some obscure kernel lock whose name I forgot to write down. A look at /var/log/messages showed:
Apr 8 17:04:05 teevee kernel: NVRM: GPU at PCI:0000:01:00: GPU-85983119-e9ef-ac66-6817-fbed5657b871
Apr 8 17:04:05 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000001 00000094 00010049 00000007 00000000
Apr 8 17:04:05 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000001 000000c0 0001004b 00000007 00000000
Apr 8 17:04:18 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 0000045c 0001002c 00000007 00000000
Apr 8 17:04:18 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000088 0001000a 00000007 00000000
Hardware problems? Probably. I wonder what it really means. But I just wanted to watch the news. Reboot the machine.
Oh. I still hadn't fixed the X configuration. OK, take a quick look, and decide that maybe it will work anyway. So I pressed the Big Red Button (black, power on/off) and rebooted.
Once again the automatic start of X failed, simply with the message that the server had stopped. OK, this time save the old log file to compare it with the new one. Start X manually. No problems. Remap the keyboard with the new, modified .xmodmap file. No problems.
So why didn't it start automatically? Off to take a look at the log files:
=== grog@teevee (/dev/pts/6) ~ 22 -> l -rt /var/log/Xorg*0408*
-rw-r--r-- 1 root wheel 43,858 8 Apr 17:03 /var/log/Xorg.0.log.old-20200408
-rw-r--r-- 1 root wheel 21,144 8 Apr 18:51 /var/log/Xorg.0.log-20200408
Oh. The old log file was the file from before the crash. What happened to the log file from the aborted automatic start? It seems that it didn't create one.
What is going on here?
In passing, took a look at the old log file to see whether it could help understand the hang. But no, there was just a disconnect/reconnect sequence, something that I have seen before, 777017 seconds (about 9 days) after the server was started. No idea how that relates to when the hang happened. Why does't X use clock times?
|
Supermoon!
|
Topic: photography, general, opinion | Link here |
This morning Jari Kirma posted a photo on IRC:
Nice photo. It wasn't until some hours later that I discovered that it was a supermoon, apparently the biggest for who knows how long.
That photo looks as if it was taken with a mobile phone; the moon is clear, but it doesn't exactly dominate the image. OK, in the evening, for some reason, the sky was relatively clear, and I was able to get some photos with a longer focus lens:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But why is there chromatic aberration on the first image? It was centered, so normally any chromatic aberration would be the same on both sides, but here it's red on the bottom and blue on top. The moon was rising, but the shutter speed was 1/400 s, surely short enough to avoid any motion artefacts, and in any case, why should motion of that nature cause chromatic aberration?
The third photo shows a different problem: lack of dynamic range. It was exposed 11 EV more than the first image,but the trees in the foreground are still just silhouettes, and the moon is (obviously) overexposed. I should try some exposure bracketing if the weather is still acceptable tomorrow evening.
In passing, it's interesting that I can't detect any camera shake in the last image. The leaves are a particularly difficult subject in that regard. It was taken at 1/6 s with a focal length of 100 mm, which in this case corresponds to 200 mm “full frame”. My old rule of thumb told me that I would get camera shake below 1/200 s. Isn't image stabilization wonderful?
| Thursday, 9 April 2020 | Dereel → Ballarat → Dereel | Images for 9 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
dereel update
|
Topic: technology | Link here |
As planned, today I continued my installation on the virtual machine that I had set up to bootstrap the next dereel. Also created (yet another) page describing what needs to be done. That's a moving target. Today I started building the latest FreeBSD 12.1-STABLE, which ran for a surprising time—in fact, the rest of the day.
In the meantime, how do I reinstall the ports? pkg upgrade has its limitations: first, I've experienced repeatedly that it fails. And then there's the issue of whether I even need the ports. The alternative is to make a list of all installed ports, but install only those ports that I recognize. If something is missing, I can use the old list to guess what it might be.
OK,
=== grog@eureka (/dev/pts/19) ~ 14 -> mkdir public_html/HOWTO/portlists=== grog@eureka (/dev/pts/19) ~ 15 -> pkg info > public_html/HOWTO/portlists/all-on-eureka-20200409
That gives me a list with entries like
trapproto-3.4.3 DEC-XTRAP extension headers
trousers-0.3.13_1 Open-source TCG Software Stack
twm-1.0.9 Tab Window Manager for the X Window System
unzip-6.0_7 List, test, and extract compressed files from a ZIP archive
urlview-0.9.20131021 URL extractor/launcher
Many problems here: I have no idea what trapproto and trousers are. twm is an ancient window manager that I almost certainly don't need. And what's unzip? It doesn't seem to actually create the archives. How does it relate to zip? The only one on that list that I definitely need is urlview, which makes up for a deficiency in mutt to identify URLs in an email message and get a web browser to display them. But do I need urlview-0.9.20131021? No, I want the current modern version. How do I install it? For most such things, pkg install urlview should do. In general, it would be a good idea to just remove the first hyphen and everything behind it, collapse the list to a single line and feed it to pkg install.
More tidying showed all sorts of things that definitely didn't need to be there any more, and maybe never needed to be there:
pilot-link-0.12.5_2,1 Suite of tools to connect and sync your Palm handheld
pt_BR-gimp-help-html-2.8.1_2 The GIMP User Manual in Brazilian Portuguese
ru-gimp-help-html-2.8.1_2 The GIMP User Manual in Russian
zh_CN-gimp-help-html-2.8.1_2 The GIMP User Manual in Simplified Chinese
I recall growling about GIMP in the past, but clearly pilot-link is a good example of why it's a good idea to install from scratch. It has way outlived its usefulness and the hardware; I don't think I've used it in the last 18 years.
On the other hand, there are others that are just plain confusing:
python-2.7_2,2 The "meta-port" for the default version of Python interpreter
python2-2_3 The "meta-port" for version 2 of the Python interpreter
python27-2.7.15 Interpreted object-oriented programming language
python3-3_3 The "meta-port" for version 3 of the Python interpreter
python34-3.4.5 Interpreted object-oriented programming language
Yes, there are at least two different flavours of python. But which two? And these names show another issue: which parts of the port name apply to the port itself, and which are version numbers? The simplistic idea of just truncating the names at the first hyphen is clearly not going to work. It also gives rise to names like:
mysql56
mysql56
That was once mysql56-client and mysql56-server. I'm left with an obsolete version number, but not the complete description.
And then there's stuff like:
tex-basic-engines-20150521 Basic TeX Engines
tex-kpathsea-6.2.1_1 Path searching library for TeX
tex-ptexenc-1.3.3_2 Library for Japanese pTeX and its tools
tex-texmflocal-1.9 Meta-port that creates a site-local $TEXMF directory
tex-web2c-20150521_1 TeX implementation translating WEB to C
Is the TeX port really called tex-basic-engines? You'd think people were trying to hide it.
In the end I created a short list packages.base with those packages that I recognized and wanted to install on the new system, along with a second file packages.maybe that might potentially be of interest.
Finally I reduced packages.base to something approaching port names and tried installing them, with the small optimization of installing X and Emacs first. Of course they didn't all exist, so pkg refused to install any of them:
===== Thu 9 Apr 2020 11:37:41 AEST on dereel: pkg install xorg emacs ImageMagick a2ps aalib apache apache24 audacity bind910 chromium cowsay cpuid curl dcraw dvd+rw espeak exif exifprobe exiv2 ffmpeg firefox fvwm ghostscript9 gimp gindent gv hugin kdenlive klondike lame lensfun lightzone linuxlibertine lsof lynx mencoder mplayer mpv mtools mutt mysql80-server nmap nvidia nvidia p5 patch pbzip2 php php56 php56 php56 php56 pidgin pigz portlint postfix procmail projectx qpopper rawtherapee rdesktop rsync rtorrent samba42 squid subversion sudo tightvnc transcode ufraw urlview vigra virtualbox virtualbox vlc vobcopy webcamd wget wireshark x264 x2x xcpustate xearth xev xfig xfstt xgc xhost xine xinit xinput xkbcomp xkbevd xkbutils xkeyboard xkill xlsatoms xlsclients xmag xmessage xmixer xmlcatmgr xmlcharent xmlprpr xmlrpc xmodmap xpat2 xpdf xpdfopen xpr xprop xproto xrandr xrdb xrefresh xset xsetmode xsetroot xterm xtrans xtset xv xvid xvidtune xvinfo xwd xwininfo xwud yajl youtube_dl zip zsh
Updating FreeBSD repository catalogue...
FreeBSD repository is up to date.
All repositories are up to date.
pkg: No packages available to install matching 'ImageMagick' have been found in the repositories
1.84 real 0.01 user 0.00 sys
OK, find the names the hard way, one by one:
=== root@dereel (/dev/pts/2) ~ 3 -> for i in a2ps aalib apache apache24 audacity bind910 chromium cowsay cpuid curl dcraw dvd+rw espeak exif exifprobe exiv2 ffmpeg firefox fvwm ghostscript9 gimp gindent gv hugin kdenlive klondike lame lensfun lightzone linuxlibertine lsof lynx mencoder mplayer mpv mtools mutt mysql80-server nmap nvidia nvidia p5 patch pbzip2 php php56 php56 php56 php56 pidgin pigz portlint postfix procmail projectx qpopper rawtherapee rdesktop rsync rtorrent samba42 squid subversion sudo tightvnc transcode ufraw urlview vigra virtualbox virtualbox vlc vobcopy webcamd wget wireshark x264 x2x xcpustate xearth xev xfig xfstt xgc xhost xine xinit xinput xkbcomp xkbevd xkbutils xkeyboard xkill xlsatoms xlsclients xmag xmessage xmixer xmlcatmgr xmlcharent xmlprpr xmlrpc xmodmap xpat2 xpdf xpdfopen xpr xprop xproto xrandr xrdb xrefresh xset xsetmode xsetroot xterm xtrans xtset xv xvid xvidtune xvinfo xwd xwininfo xwud yajl youtube_dl zip zsh; do echo === $i; pkg search $i; done | tee -a Log.log
That gave me yet another long list to investigate. In the meantime, at least managed to install X and Emacs. Next in line was ImageMagick. But that failed because of the breakage with Emacs that I had run into last month: Emacs wants version 6, other programs want version 7. Started to rebuild Emacs without ImageMagick support, like I had done last month, but of course the build tools haven't been installed yet.
So where now? Defer ImageMagick and install those of the rest that will install cleanly, and return to ImageMagick after that.
But that's all for another day. I have other things to do.
|
More weed spraying
|
Topic: gardening | Link here |
Yvonne wanted to spray weeds on the only part of the property that she worries about, the arena, and maybe also the dog run. So today I showed her how to use the new spray unit, in the process spraying some tenacious sorrel round the rose garden. She decided against the dog run, which is finally responding to her last spray efforts, and did the arena. Only later did it occur to me that I had planned to spray the north bed too.
|
Ballarat in the time of COVID-19
|
Topic: general, opinion | Link here |
Mick Solly is coming tomorrow, Good Friday or no Good Friday, and I want him to replace the submersible drain pump in front of the house. So it would be a good idea to have a pump to install.
In principle I have one, but it's old, and since the previous one died (short circuit or similar) after not very long, it sounded like a better idea to buy a new one. So off to Bunnings.
How do you get in? Walked up to the entrance, and was turned away. Now it's only an exit:
|
|||||||||||||||||||||||
|
|
|||||||||
Why? In principle Bunnings have symmetrical entrances (the other one is visible in the background). So I went in that way anyway, and was chased out. Because of the virus. What does that mean? She couldn't explain either, but was becoming visibly distressed, so over to the other side, where they had cordoned the place off and made people go three times the distance past guide ropes that would be a perfect site for the virus to attach themselves to:
|
|||||||||||||||||||||||
I suppose that makes sense when they need to limit the number of people who can enter the premises at any one time. But that's not now, as the photo clearly shows. The guard on the right did lower the barrier to allow an employee to roll in some (uncleaned) trolleys, so why not leave it open until the sudden rush of customers happens, the one that has never happened since they opened the place? Somehow people seem to be reacting without thinking.
On then to ALDI, noting on the way that the lack of traffic and people in the streets that I noted three weeks ago has dissipated. Customer levels seemed normal both at Bunnings and at ALDI. But there's still no toilet paper, after well over a month. What are people doing with it? And even the cornflour that I was looking for was rationed:
|
|||||||||||||||||||||||
Not a serious problem. The last time we bought cornflour was a year or two ago, and it's gradually running out. I only needed 500 g.
|
Google gets on the bandwagon
|
Topic: technology, opinion | Link here |
It was bad enough that GIMP installed multiple copies of the documentation in different languages, but that was (probably) years ago, and they may have learnt since then. What's less understanding is why I continually get a ding-ding ding-ding on my mobile phones with silly questions like:
|
|
|||||||||
|
|
|||||||||
Why do they do this? I never asked for it, and how many people need all three? It's painful enough entering English text on a mobile phone “keyboard”.
|
Quora: run by idiots or androids?
|
Topic: general, technology, opinion | Link here |
A few months ago I stopped answering questions on Quora because of their heavy-handed and non-transparent
approach with perceived violations of their policy. I got the impression that the
identification was done by computer programs, but that doesn't exonerate them for not
responding to my complaint “appeal”.
Gradually I started answering a few questions, though nothing like the number I used to answer. Clearly too many. Today I received:
Quora Moderation collapsed your answer for violating a policy on Quora.Your answer may need improvement
What's up now? The question was about visa regulations, and my answer said “don't ask Quora, ask the consulate”. What policy does that violate? But I'm allowed to “appeal”. Did that, and based on my previous experience not expecting any answer.
Submit an Appeal
Quora Moderation has flagged your answer as not complying with a policy on Quora: Your answer may need improvement Show details.
Your answer is currently hidden. If you'd like to make the answer visible again, you can edit your answer to conform to the policy, then submit an appeal below.
OK, I “appealed”, in a manner which they deserve:
People, what kind of stupidity is this? This is the only valid answer. This continual “may need improvement” and lack of response to this kind of “appeal” makes me wonder why I even bother with Quora. You’re certainly not doing anybody a favour with this kind of action. Has a human being even seen this? What policy have I violated? What would a good answer look like in your opinion?
I guessed that it was some automatic procedure, potentially an automatic response to something, which in turn might be a reader who didn't like it. They do the same with downvotes: one downvote and you're gone, no matter how serious or good your response may be.
But in the course of the evening another three “collapses” occurred, all about the same thing: questions about visas. Do they not like me saying “don't ask Quora”?
Another thing is that all of these answers were over 2 years old. Why now? Sometimes I wonder why I bother.
| Friday, 10 April 2020 | Dereel | Images for 10 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More garden spraying
|
Topic: gardening | Link here |
Mick Solly was a no-show today: his wife reminded him that it was Good Friday, and that they should visit graves of the departed. Understandable, though I don't know how that fits the current curfews.
But the weather was much better than forecast. They had talked of rain all day long, but in fact none came until the evening, and I had time in the morning to finish the weed spraying that I had forgotten yesterday.
|
Continued upgrade pain
|
Topic: technology, general, opinion | Link here |
Yesterday Arstechnica published an article on installing FreeBSD. Various FreeBSD lists started discussing it, producing more text than I could handle at the moment. I skimmed through the article and recognized both things that I recognized and things that I didn't. It seems that the article was well-written and, though the author comes from the Linux camp, relatively objective. Things like “mouse doesn't work out of the box” are clearly wrong. Installing GNOME is, in my world, also wrong, but it should work, and the author had trouble with it.
On the other hand, some of it I can definitely relate to:
The apparent lack of a simple xorg package turns out to be a big, fat lie—if I had just typed pkg install xorg, it would have worked. But as I eventually discovered, pkg search doesn't return meta-packages. Yay.
Clearly I'm not a newbie to FreeBSD, and I'm not really installing for the first time, but all this should be easier. Today I spent almost the whole day writing up what I did yesterday. But I do think that it's worth going into this much detail; it may save me pain in the future.
What I did do: the kernel build failed after 11 hours with a spurious NFS write error (to the log file) that I've seen before, along with an error message that doesn't exactly match:
tee: Make.log.dereel: Input/output error
/eureka/home/src/FreeBSD/svn/stable/12/sys/amd64/vmm/amd/svm.c:47:10: fatal errortee: Make.log.dereel: Input/output error
: 'machine/reg.h' file not found
#include <machine/reg.h>
^~~~~~~~~~~~~~~
1 error generated.
*** Error code 1
...
make: stopped in /eureka/home/src/FreeBSD/svn/stable/12
39969.89 real 25871.58 user 3361.23 sys
That worked well enough when I restarted the build, not without being told off for remembering how to restart:
=== root@dereel (/dev/pts/0) /usr/src 16 -> Make -DNOCLEAN buildworld kernel
su: xtset: command not found
===== Fri 10 Apr 2020 09:42:08 AEST on dereel: Make -DNOCLEAN buildworld kernel
make[1]: "/eureka/home/src/FreeBSD/svn/stable/12/Makefile.inc1" line 452: warning: NOCLEAN option is deprecated. Use NO_CLEAN instead.
But then it completed, and I was able to confirm that uname castration has proceeded even further:
=== grog@dereel (/dev/pts/0) ~ 1 -> uname -a
FreeBSD dereel 12.1-STABLE FreeBSD 12.1-STABLE GENERIC amd64
Previously it at least gave the revision number of the tree at the time of the build. Once again create /etc/src.conf (a name that seems to mean nothing):
# Don't drop build date, dammit!
# What a silly name for the knob.
# It's about as "intuitive" as the name of this file.
WITHOUT_REPRODUCIBLE_BUILD=dammit
But that wasn't all. When I create or modify a system file, I check it in with RCS. But there was none! It seems that they've removed this basic utility and made a port of it. Shades of Linux! At least the port installed simply.
And after another kernel build, install and reboot things looked sane:
FreeBSD dereel 12.1-STABLE FreeBSD 12.1-STABLE #2: Fri Apr 10 10:49:21 AEST 2020 grog@dereel:/usr/obj/eureka/home/src/FreeBSD/svn/stable/12/amd64.amd64/sys/GENERIC amd64
OK, on with ports installed. Next was ImageMagick. But the ports are too polite to use such a straightforward name. Instead I had to search for it:
=== root@dereel (/dev/pts/2) ~ 8 -> pkg search ImageMagick
ImageMagick6-6.9.10.90,1 Image processing tools (legacy version)
ImageMagick6-nox11-6.9.10.90,1 Image processing tools (legacy version)
ImageMagick7-7.0.9.20 Image processing tools
ImageMagick7-nox11-7.0.9.20 Image processing tools
...
OK, ImageMagick7 it is. In a moment of madness I tried to install it, well knowing that Emacs didn't like it. But how about that, not the complaint I expected:
<code><font color="blue">=== root@dereel (/dev/pts/2)</font> <font color="red">~</font> <font color="blue">9</font> -> <b><tt>pkg install ImageMagick7</tt></b></code>
Updating FreeBSD repository catalogue...
FreeBSD repository is up to date.
All repositories are up to date.
The following 1 package(s) will be affected (of 0 checked):
New packages to be INSTALLED:
ImageMagick7: 7.0.9.20
...
Proceed with this action? [y/N]: y
[1/1] Fetching ImageMagick7-7.0.9.20.txz: 100% 8 MiB 465.4kB/s 00:18
pkg: cached package ImageMagick7-7.0.9.20: size mismatch, fetching from remote
Fetching ImageMagick7-7.0.9.20.txz: 100% 8 MiB 1.2MB/s 00:07
pkg: cached package ImageMagick7-7.0.9.20: size mismatch, cannot continue
Consider running 'pkg update -f'
Huh? This pkg database was virgin only a couple of days ago. OK, pkg update -f it is, and sure enough,
=== root@dereel (/dev/pts/2) ~ 13 -> pkg install ImageMagick7
The following 2 package(s) will be affected (of 0 checked):
New packages to be INSTALLED:
ImageMagick7: 7.0.9.20
Installed packages to be REINSTALLED:
ImageMagick6-6.9.10.90,1
Number of packages to be installed: 1
Number of packages to be reinstalled: 1
Huh? Where did this ImageMagick6-6.9.10.90,1 come from? OK, let's see what happens:
Proceed with this action? [y/N]: y
[1/1] Fetching ImageMagick6-6.9.10.90,1.txz: 100% 7 MiB 277.8kB/s 00:27
[1/2] Deinstalling ImageMagick6-6.9.10.90,1...
[1/2] Deleting files for ImageMagick6-6.9.10.90,1: 100%
[2/2] Installing ImageMagick7-7.0.9.20...
[2/2] Extracting ImageMagick7-7.0.9.20: 100%
[2/2] Installing ImageMagick6-6.9.10.90,1...
pkg: ImageMagick6-6.9.10.90,1 conflicts with ImageMagick7-7.0.9.20 (installs files into the same place). Problematic file: /usr/local/bin/Magick++-config
Well, isn't that clever? It tried to install two conflicting packages. That's enough for me.
| Saturday, 11 April 2020 | Dereel | Images for 11 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The daily ports grind
|
Topic: technology, opinion | Link here |
OK, where did I get to yesterday? I couldn't look in the diary entry, because I hadn't written it yet. What's in the log file?
=== root@dereel (/dev/pts/3) /var/cache 3 -> emacs Log.log
ld-elf.so.1: Shared object "libMagickWand-6.so.6" not found, required by "emacs"
Damn! That ImageMagick installation failed, but it left me without dependencies for Emacs. The quick answer was to reinstall Emacs, but that wasn't a real option yet.
So: install all the ports that would install easily:
for i in a2ps aalib apache audacity bind916 chromium cowsay cpuid curl dcraw dvd+rw espeak exif exifprobe exiv2 ffmpeg firefox fvwm ghostscript9 gimp gindent gv hugin kdenlive klondike lame lensfun lightzone linuxlibertine lsof lynx mencoder mplayer mpv mtools mutt mysql80-server nmap nvidia nvidia p5 patch pbzip2 php pidgin pigz portlint postfix procmail projectx qpopper rawtherapee rdesktop rsync rtorrent samba42 squid subversion sudo tightvnc transcode ufraw urlview vigra virtualbox virtualbox vlc vobcopy webcamd wget wireshark x264 x2x xcpustate xearth xev xfig xfstt xgc xhost xine xinit xinput xkbcomp xkbevd xkbutils xkeyboard xkill xlsatoms xlsclients xmag xmessage xmixer xmlcatmgr xmlcharent xmlprpr xmlrpc xmodmap xpat2 xpdf xpdfopen xpr xprop xproto xrandr xrdb xrefresh xset xsetmode xsetroot xterm xtrans xtset xv xvid xvidtune xvinfo xwd xwininfo xwud yajl youtube_dl zip zsh; do echo === $i; pkg install -y $i; done | tee -a /var/tmp/portinstall
That seemed to install quite a few ports, so I left it at that. I can check the log file tomorrow. Instead, went back to Emacs. Build from source without ImageMagick. That wasn't completely straightforward: help2man, whatever that is, failed:
===> Configuring for help2man-1.47.13
env: ./configure: No such file or directory
===> Script "configure" failed unexpectedly.
This just Should Not Happen. Installed the help2man package, and had no further problems. And the result worked, as before. Now I can remove ImageMagic6 and install ImageMagick7:
=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 35 -> pkg remove ImageMagick6
...
[1/1] Deinstalling ImageMagick6-6.9.10.90,1...
[1/1] Deleting files for ImageMagick6-6.9.10.90,1: 100%=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 36 -> pkg install ImageMagick7
...
The following 3 package(s) will be affected (of 0 checked):
New packages to be INSTALLED:
ImageMagick6-nox11: 6.9.10.90,1
ImageMagick7: 7.0.9.20
libwmf-nox11: 0.2.8.4_15
Number of packages to be installed: 3
The process will require 56 MiB more space.
8 MiB to be downloaded.
Huh? That's not what it said last time. And what's libwmf anyway? OK, try it and see:
Wonderful. What now? Try to remove the nox11 (pronounced “No X11”) version. Not there, so try to install it:
=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 37 -> pkg delete libwmf-nox11-0.2.8.4_15
No packages matched for pattern 'libwmf-nox11-0.2.8.4_15'=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 38 -> pkg install libwmf-nox11-0.2.8.4_15
pkg: transcode has a missing dependency: lame
pkg: mencoder has a missing dependency: lame
Checking integrity... done (2 conflicting)
- libwmf-nox11-0.2.8.4_15 conflicts with libwmf-0.2.8.4_15 on /usr/local/bin/libwmf-config
- libwmf-nox11-0.2.8.4_15 conflicts with libwmf-0.2.8.4_15 on /usr/local/bin/libwmf-config
Checking integrity... done (1 conflicting)
- ImageMagick6-nox11-6.9.10.90,1 conflicts with ImageMagick6-6.9.10.90,1 on /usr/local/bin/Magick++-config
Checking integrity... done (0 conflicting)
The following 8 package(s) will be affected (of 0 checked):
Installed packages to be REMOVED:
gimp: 2.10.18,2
gimp-app: 2.10.18_1,1
gimp-gutenprint: 5.3.3
libwmf: 0.2.8.4_15
py27-gimp: 2.10.18_1
New packages to be INSTALLED:
ImageMagick6-nox11: 6.9.10.90,1
libwmf-nox11: 0.2.8.4_15
Installed packages to be REINSTALLED:
flac-1.3.3
Proceed with this action? [y/N]:
In passing, it's interesting to note these lame dependencies. lame doesn't mean that they're lame; it's the name of the package. I've seen it for some time, so it seems to be a bug in the lame port.
But at least we now have an indication of why this strange behaviour occurs: it seems that GIMP wants ImageMagick6 too. That's bad. I've already commented that Emacs doesn't really need ImageMagick, but GIMP most certainly does. Surely it should be the latest version.
OK,
=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 39 -> pkg delete gimp
[1/1] Deinstalling gimp-2.10.18,2...=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 40 -> pkg install ImageMagick7
Checking integrity... done (1 conflicting)
- ImageMagick6-6.9.10.90,1 conflicts with ImageMagick7-7.0.9.20 on /usr/local/bin/Magick++-config
Checking integrity... done (4 conflicting)
- libwmf-nox11-0.2.8.4_15 conflicts with libwmf-0.2.8.4_15 on /usr/local/bin/libwmf-config
- libwmf-nox11-0.2.8.4_15 conflicts with libwmf-0.2.8.4_15 on /usr/local/bin/libwmf-config
- ImageMagick6-nox11-6.9.10.90,1 conflicts with ImageMagick7-7.0.9.20 on /usr/local/bin/Magick++-config
- ImageMagick6-nox11-6.9.10.90,1 conflicts with ImageMagick6-6.9.10.90,1 on /usr/local/bin/Magick-config
Checking integrity... done (0 conflicting)
The following 2 package(s) will be affected (of 0 checked):
Installed packages to be REMOVED:
emacs: 26.3_3,3
New packages to be INSTALLED:
ImageMagick7: 7.0.9.20
Dammit, I've built Emacs without ImageMagick dependencies. It seems that the information didn't make it to the package database. That's bad from a consistency viewpoint, but I can live with it. I'm even still in the Emacs ports directory, so after pkg deletes Emacs, all I need is:
=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 41 -> make install=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 42 -> emacs
su: emacs: command not found
Dammit, I've just removed Emacs. But the ports tree doesn't know about it. OK,
=== root@dereel (/dev/pts/3) /usr/ports/editors/emacs 43 -> make deinstall reinstall
===> Deinstalling for emacs
===> emacs not installed, skipping
===> Deinstalling for emacs
===> emacs not installed, skipping why twice?
...
===> Registering installation for emacs-26.3_3,3
(emacs-26.3_3,3) /eureka/home/src/FreeBSD/svn/ports/editors/emacs/work-full/stage//usr/local/bin/emacs-26.3 - required shared library libMagickWand-6.so.6 not found
(emacs-26.3_3,3) /eureka/home/src/FreeBSD/svn/ports/editors/emacs/work-full/stage//usr/local/bin/emacs-26.3 - required shared library libMagickCore-6.so.6 not found
Installing emacs-26.3_3,3...
Somehow there's something seriously wrong with the consistency between packages and ports. But why doesn't Emacs build with ImageMagick7? Time to investigate. Change the Makefile:
--- Makefile (revision 531446)
+++ Makefile (working copy)
@@ -176,7 +176,7 @@
M17N_IMPLIES= XFT
M17N_LIB_DEPENDS= libm17n.so:devel/m17n-lib
M17N_CONFIGURE_WITH= m17n-flt
-MAGICK_LIB_DEPENDS= libMagickCore-6.so:graphics/ImageMagick6
+MAGICK_LIB_DEPENDS= libMagickCore-7.so:graphics/ImageMagick7
Why, doesn't it build with ImageMagick7? It built fine and worked. The real question is why it failed last month. But first, asked on IRC, and Juha Kupiainen built from source with that patch, noting that it installed ImageMagick7 even with make all.
So: next to try it out on teevee again. But that was enough pain for one day.
|
What is ISO, part 36
|
Topic: photography, opinion | Link here |
A year ago I mused about the meaning of “ISO”, really meaning the ISO standards for digital camera sensor sensitivity. Now DPReview has come up with a pair of articles (well, at least the first one). So far no deep insights, but at least no nonsense. They're a little too hands-off with their description of how the standard (they chose ISO 12232:2019) applies to camera hardware. Maybe the second article will be closer to reality. About the only thing that I did learn was two TLAs: SOS and REI. SOS (Standard Output Sensitivity) is more standardized: it describes the exposure needed to get “an sRGB JPEG with lightness values of 118 from a middle grey target”. The lightness value is simply the value of the pixel.
Why 118? I still don't completely understand. JPEGs have pixel values from 0 to 255, so a half-full (or half-empty) pixel would have a value of 128. But I seem to recall that JPEGs consider any value over 250 to be burnt out. That would still leave the mid point at 125, even if you ignore the noisy low values. Possibly the text of the standard (unfortunately not freely available) would clarify it. But that definition is very different from the film definitions, which describe the threshold where any image information occurs.
The other thing of interest is that the standard specifically refers to JPEG images. This implies that raw images don't have an intrinsic ISO value; you can create 5 different JPEG images with different ISO values from the same raw image. Somehow this suggests to me that there's something basically wrong in the definitions. A JPEG is a derived image, not one that the camera produces automatically.
| Sunday, 12 April 2020 | Dereel | Images for 12 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Busy day
|
Topic: general | Link here |
Up early this morning: Mick Solly was coming, possibly as early as 8:00, to do the garden, Easter Sunday or no Easter Sunday. And CJ Ellis was coming to put in a couple of gates.
8:00 came. 9:00 came. All day went by. No Mick. He didn't show up at all. Where is he? I suppose I should give him a call, but normally he's very good with calling if he can't make it. Did he leave a message on my phone? Yes!
|
|
|||||||||
Oh. Look at that date. That must have been the person from Elysian Energy who ignored my wishes and called me on the phone, triggering a search for the cause of the non-ringing phone. Time to tidy things up. And no sign of Mick.
CJ along as planned and put in a couple of posts:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In principle we also wanted to put gates on them, but various things got in the way: the soil is very moist, we didn't have the correct fittings, and we also needed to install the electric gate opener, which requires other tools and a whole amount of guesswork, since the (relatively detailed) instruction manual doesn't match what I received. So we deferred the rest until later in the week.
|
Slow cooker galbi jjim (갈비찜)
|
Topic: food and drink, opinion | Link here |
I'm subscribed to a number of email recipe lists, and recently I received one from Sue (apparently without further names) of My Korean Kitchen: Slow Cooker Korean Short Ribs, a translation of galbi jjim (갈비찜).
It looked interesting, and in particular it seemed that it would be something that wouldn't offend Yvonne's far too European taste. So I asked her to get some short beef ribs, and today I started the task of 8 hours' cooking.
The ribs were interesting. They looked almost nothing like the ones that Sue used (though she's in Brisbane).
|
|||||||||||||||||||||||
They were also amazingly expensive, $17 per kg. I wouldn't have expected them to be over $10.
OK, prepare as the recipe, trying to interpret things like cups and Tbsps and things. It didn't occur to me until much later that this was an “Australian” recipe, so the cups were round 4% more, and the Tbsps were ⅓ more than the correspondingly name US “units”. And how much ¼ cup of brown sugar is is mainly guesswork; I tried measuring and decided that it was about 110 g, but those were based on US cups.
Ended up by choosing nearly double the quantity of ingredients for the sauce, since my ribs had so much more meat on them, and they also weighed a little more. Ultimately I used:
| quantity | ingredient | step | ||
| 1.75 kg | beef short ribs | 1 | ||
| 785 g | carrots | 1 | ||
| quantity | ingredient | step | ||
| 120 g | Apple | 1 | ||
| 100 g | Pear | 1 | ||
| 140 g | Onion | 1 | ||
| 290 g | Water | 1 | ||
| 5 g | Chicken stock powder | 1 | ||
| 220 g | Light soya sauce | 1 | ||
| 110 g | Brown sugar | 1 | ||
| 64 g | Honey | 1 | ||
| 40 g | Garlic | 1 | ||
| 23 g | Ginger | 1 | ||
| 68 g | Rice wine | 1 | ||
| 10 g | Sesame oil | 1 | ||
| 10 | Peppercorns | 1 |
Many of these choices depended on what I had. The original wanted “baby potatoes”, but it proved to be the only recipe I found that used potatoes, and since I had just the right quantity of carrots, I used them instead. The original wanted an “Asian” or Williams pear, but there was also mention of red apples. The water and stock powder were instead of beef stock, which I proved not to have. Ginger wasn't in the original recipe, but I had seen it in others, and it seemed the right thing to do.
OK, put it in my sous-vide machine, which also lays claim to be a slow cooker:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Turn on to “low” (or “LO”, as the cooker calls it). What temperature does that correspond to? How long does it take to warm up? After a good hour, the edges were warm (the sous-vide half claimed 60° or so), but the middle was round 26°. After some examination discovered that the thing heats only from the ends. That seems to be enough for a water bath, but here there's no circulation, and it would presumably have taken all day to warm up. The machine works well as a sous-vide machine, but it's clearly useless as a slow cooker.
OK, take the basin out of the machine and put it on the cooktop with a thermometer:
|
|||||||||||||||||||||||
What temperature am I aiming for? Off to browse for suitable information, all of which was in Fahrenheit. It seems that the “low” setting should be about 80°, considerably more than I expected, and “high” is closer to 95°. OK, we can do that; it proved that the cooktop needed to be set at 2½ or 3, but it took a long time to get that far.
In the process of searching for temperatures, came across this startling statement in the Wikipedia page:
Raw kidney beans, and, to a lesser extent, some other such beans as broad beans or fava beans, contain the highly toxic lectin phytohemagglutinin. Boiling destroys this lectin, but the far lower temperature of a slow cooker does not.
That's the first I have heard of that, but it appears to be plausible. Wouldn't you think that the packaging for this kind of bean would include that warning?
OK, an hour before it's finished. What does it look like?
|
|||||||||||||||||||||||
Not much difference. And the carrots were still raw! It seems that, while the temperature in the middle was round 80°, the temperature on the surface was closer to 45°. Not the way to cook it.
Too late to correct. Put it on the gas flame and brought it to the boil, so that after an hour it was cooked—but not exactly “slow cooked”.
The result? The meat tasted alright, but the fat hadn't gelatinized the way it should have, and parts of it were inedible:
|
|||||||||||||||||||||||
I'm going to have to slow cook the rest of it after all. In the oven?
In passing, I probably didn't choose the best recipe. This one, by Maangchi, is much more elaborate, and states:
Traditionally it’s a dish for special occasions such as Korean festival days like Chuseok, New Year’s Day (Seollal), or family’s birthdays,
What I produced didn't come close to being anything special. Next time I'll look more at Maangchi's recipes.
|
Daemon influence?
|
Topic: photography, general, food and drink, opinion | Link here |
While taking photos of the food this morning, the flash unit went off a number of times at random for no obvious reason. In one case I wasn't even in the room.
What causes that? The flash trigger is relatively new. Is it defective? Is it responding to power surges from the grid? Is there an intermittent contact? Played around a bit with the connections, and it didn't happen again.
But then, while measuring the temperature of the “slow cooker”, the thermometer went mad. It switched from Celsius to Fahrenheit, showed incorrect temperatures at random, and even switched into a mode that I couldn't identify, but which had nothing to do with temperature measurement. It also didn't appear to have anything to do with the timer function that it can also perform.
Defective? OK, years of work with Tandem Computers have had their effect: I have a backup device, so I used that. And the same thing happened again with the new device!
What's causing that? I now have three devices that misbehaved in a way I've never seen before. The thermometers are clearly not affected by the grid voltage (and I later confirmed that the inverter hadn't noticed any irregularities). What is it?
| Monday, 13 April 2020 | Dereel | Images for 13 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Salvaging the galbi jjim
|
Topic: food and drink, opinion | Link here |
Most of yesterday's galbi jjim was left over, and it hadn't been really “slow cooked”. So repeat today.
How? Clearly the “slow cooker” function of the sous-vide cooker was useless. It didn't work well on the stove either. So I put it in the oven. What temperature? When cooking baked beans I set the oven to 120°. I wanted the meat to cook at 80°. How much warmer? Started at 110° oven temperature, but that proved to be far too hot. To my surprise, after some adjustment, I discovered that the correct oven temperature setting was 80°. I had expected evaporation to cool the contents below the ambient temperature, but that doesn't seem to be the case. It's also interesting to note how exactly the surface temperature in the middle corresponded to the oven setting. On the edges of the basin it dropped to 77°, certainly still close enough.
|
SMS to “landlines”?
|
Topic: technology, general, opinion | Link here |
Call from Mick Solly this afternoon. He had been incapacitated yesterday. Did I get his SMS? No, where did he send it? To my “landline” (really a VoIP phone in the “landline” address space).
How can that work? He doesn't know either, but his phone accepted it. That sounds strange, so I tried it. Yes, sure enough, you can send an SMS to a destination that's not equipped to receive or present it.
Some discussion on IRC suggested that some Telstra and Optus phones really can display SMSs. But wouldn't you think that the phone companies would do something to present the others, like offer a diversion to a mobile phone, or convert to voice?
Off to look at MyNetFone, who really do offer a means of sending SMS, but not receiving them. But this page shows how do send SMS messages from a web browser; that could be useful, though first attempts weren't very successful.
|
More gate work
|
Topic: Stones Road house, general, opinion | Link here |
CJ Ellis along somewhat unexpectedly today to do some more work on the gates, still somewhat hampered by the lack of tools and components. At least he got the gate mounted; the rest will be round the weekend.
|
Use for superfluous tripod components
|
Topic: photography | Link here |
When I fitted the leveling head to Yvonne's tripod, I had a problem: the tightening screw of the Arca-Swiss mount had a tightening screw that fouled the leveling head. I had to add a macro rail in between:
|
|||||||||||||||||||||||
I'm still looking for something better, but today I saw something that made me wonder whether I should bother:
|
|||||||||||||||||||||||
Maybe I should just leave it like that.
|
Freezing aspic
|
Topic: food and drink, opinion | Link here |
A year ago Yvonne made some salmon in aspic. We ate some and froze the rest, at least partially as an experiment.
Today we thawed it out again, discarding large quantities of water. What was left wasn't worth eating. At least we now know the results of the experiment.
| Tuesday, 14 April 2020 | Dereel | Images for 14 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Powercor problem resolution
|
Topic: general, technology, opinion | Link here |
Call this morning from Edward Barkla of Powercor. I had had difficulty accessing my power usage statistics on their web site: all queries produced the result “Customer response is empty”. Clearly it was a database problem, as I told him.
He didn't agree directly, but he asked if it would be OK if he removed my account and recreated it with a slightly different name (leaving out my middle name). Yes, not a problem, and it worked. But what kind of workaround is that?
In passing, it seems that Edward has grown up. I've had dealings with him for well over 10 years, and last time he was still Eddie.
|
Systems upgrade pause
|
Topic: technology, opinion | Link here |
For the past couple of days I've been too busy to continue with the pain of upgrading my computer systems. Today I had time, but most definitely not the inclination. Somehow all of this is far too hard, and I haven't even reached half-way yet: once the ports are all installed (or the ones I have identified, anyway) I still need to configure a whole number of subsystems from scratch. I don't see it happening in less than a couple of weeks.
But then, I've been trying for a couple of years, so what chance is there that it will be done even that quickly?
| Wednesday, 15 April 2020 | Dereel | Images for 15 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Galbi jjim for breakfast
|
Topic: food and drink, opinion | Link here |
After much cooking, I now need to do something with my galbi jjim (갈비찜). The plan was to eat it for breakfast. That means dividing it into individual portions. Started with a basis of 100 g of meat, but that proved to be too much; 80 g is better. But it tasted quite good, and now I have another 10 portions in the freezer to eat over the next few months.
How long did I cook it in total? About 16 hours. And the meat just fell off the bone—of the correctly cut pieces:
|
|||||||||||||||||||||||
But as I commented at the start, not all of the pieces were cut correctly, and there were joints in some of the pieces. And even after 16 hours, the meat didn't separate cleanly:
|
|||||||||||||||||||||||
I wonder how long it would have taken.
|
Wildflowers in Spearys Road
|
Topic: general, gardening, opinion | Link here |
Off down Spearys Road for a longer-than-usual walk this afternoon, way past Lorraine Carranza's place. To my surprise, there were a number of flowers there that I haven't seen elsewhere at this time of year:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
|
||||||||||
There was also a tiny 5-petaled blue flower, about 1 cm across, that I couldn't really get with the lens I had with me. Next time I'll take a macro lens with me.
And then there was this plant, not flowering:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I've seen this before in the Brisbane Ranges, but not round here. It appears to be a species of Pultenaea, like these ones taken last October:
|
||||||||||||||||||||||||
I'll have to keep my eyes open (and come down this way more often).
Finally, there was a plant that I had seen before, and which I'm still trying to identify:
|
||||||||||||||||||||||||
Something inside me wants to say “Laurel”, but I can't find anything similar online.
|
Mobile phone GPS: useless
|
Topic: technology, general, opinion | Link here |
How far did I walk today? I guessed about 3 km. But I had my Nokia 5.1 Plus with me, logging my progress. 8.55 km! (though the brain-damaged Wikiloc decides to convert it to archaic units of feet and miles):
|
|||||||||||||||||||||||
What a useless map! At best it gives a vague idea of where I have been. That confirms my decision: the phone goes back. And Google Maps gives a distance of 1.5 km each way, just what I had expected.
| Thursday, 16 April 2020 | Dereel | Images for 16 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Not untouched by COVID-19
|
Topic: health, general | Link here |
The current crisis with COVID-19 has left us largely untouched. I've been practicing social distancing for decades now, and Yvonne is not much more sociable.
But our daughter Yana is much more involved with other people. And as a casual worker (aren't they all?) in the restaurant (“hospitality”) business, she's doubly affected. Yes, she can get compensation from the government, but that goes through Centrelink, which is renowned for its slowness. So now she's out of money. Sure, we can lend her money, and we did so, but the least the government could do would be to act a little more swiftly.
| Friday, 17 April 2020 | Dereel | Images for 17 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Anniversary: Last hackers' barbecue
|
Topic: history, opinion | Link here |
It's been ten years since we had “The hackers' barbecue to end all barbecues”. I hadn't intended for that to be taken literally, but in fact we haven't had any since then.
I'm still in daily contact, via IRC, with a number of the people who showed up: Callum Gibson, Edwin Groothuis, Peter Jeremy and Juha Kupiainen. But I haven't seen them since then.
How time flies! I've been following my diary of 50 years ago, but this represents 20% of the time. And the halfway mark was also interesting: I surprise visited my father at the Beach of Passionate Love in Kelantan. At the time I had thought that it was the end of an era, and maybe it was.
|
GPS receiver accuracy
|
Topic: technology, general, opinion | Link here |
To my immense surprise, I got a positive comment about the track log of my walk with the dogs a couple of days ago:
Bonita ruta, felicitaciones.
Whatever gave him that idea?
Discussing on IRC, James Fraser suggested buying a Garmin smart watch: his wife has one, and it produces good results, as this map shows:
|
|
|||||||||
Of course, that's a relatively open route, and at a small scale; but it's clearly better than what I could do. James sent me a GPX log, but I didn't get round to looking at it today.
Daniel O'Connor, an Apple enthusiast, gave me another example done with some iPhone:
|
|
|||||||||
Admire the inverse video.
That, too, doesn't look bad, once you look past the colour scheme. But how accurate is it? Asked Daniel for a GPX file, which he was able to produce in only 15 minutes, much better than my failure last year. Put that through GPS Visualizer and Wikiloc, with surprising results. As expected, the GPS Visualizer was less smooth than Daniel's Apple stuff:
|
|
|||||||||
|
|
|||||||||
|
|||||||||||||||||||||||
But Wikiloc smoothed again. Here the second image:
|
|
|||||||||
Why do they do that? It's not so much smoothing as simplification.
In any case, both of these trails are with relatively good view of the sky. And the cheapest Garmin watch costs considerably more than most GPS navigators. The search goes on.
|
PV inspection
|
Topic: Stones Road house, general, opinion | Link here |
Daryl Buchanan of the Clean Energy Regulator's solar installation inspection program [sic] along today to, well, inspect our photovoltaic electricity installation, along with an assistant to hold the ladder (really!):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It wasn't clear from their communication whether there would be any disadvantages if I didn't agree (it appears to be voluntary), but I'm relatively happy with the quality of the installation, and wasn't expecting any issues.
But there were. He decided that there should be some protection to stop people driving into the battery. In principle that's not possible, since a fridge is in the way, but possibly he had a point. He also found the signage inadequate, with which I would certainly agree.
A bit of looking around outside, and I left him for a while. Found him back in the garage later with the inverter isolation switch in the neutral position: no power to the house! He hadn't told me that he was going to isolate the inverter, and certainly not that he was going to disconnect the house. But somehow he hadn't thought that he was going to do that. Grrr! If he had told me what he was going to do, I could have avoided two system crashes. And the more this sort of thing happens, the more I think that the UPS in my office is a good idea.
He also took many of the junction boxes apart to measure voltages, and came up with some strange values on the western battery array:
|
|||||||||||||||||||||||
113 V AC? Clearly that can't work. The inverter has a 480 V DC input. I got my multimeter and measured nothing. But his device is much cleverer than mine. What's the problem?
Simple: he was measuring the output side of the switch, which was turned off. What voltage he saw came from the inverter, quite possibly at a much higher frequency than my multimeter could handle. Power down the inverter (turn off the battery switch) and it disappeared. But I had to explain to him that he was clearly measuring the wrong side of the switch, which was evident from the wiring. The outside poles go to the PV array, and they showed a typical voltage.
Disconnecting the inverter was really just to isolate the PV array from the inverter; that's what this switch and the one next to it were for. It wasn't necessary to take the inverter out of the house circuit at all. He then went up and checked each isolation switch:
|
|||||||||||||||||||||||
That seemed to be to his liking. Came back and checked the inverter isolation switch:
|
|||||||||||||||||||||||
He didn't like that either. I can't blame him: I don't like this middle point that you have to get through quickly to avoid removing power from the house. Up goes Grid → Inverter → House, middle is all disconnected, down is Grid → House. He'll put that in his report, and maybe we'll get it replaced, though I have the impression that there might be a good reason for this middle point that he doesn't understand.
So: power up the inverter, put back in circuit (down → up). Quickly, please! Down → middle. Oh. Middle → up. Another two reboots! GRRRRRR!
Apart from that, he found the whole wiring strange. What would he have done differently? All this stuff in one box makes no sense. Why not a Selectronic unit? There the functionality is clearly delineated.
Well. As it happens, I've compared exactly these models and decided for the Ingeteam. More to the point, though, that wouldn't make any difference to the component layout: the inverter in the middle, connecting to the PV array, battery, load and grid either alone or (in the case of the Selectronic) with a separate battery controller). I'm left with the impression that Daryl doesn't really have a very good understanding of the systems.
He also doesn't have a good understanding of the current quarantine restrictions: we were spared an inspection of the ceiling space (where the cables from PV array run to the garage) because of the restrictions (not allowed to enter the house), but he came unnecessarily close to me in the garage. Clearly he hasn't accustomed himself to the 2 m rule (or is that 1.5 m? Or 1.8288 m?).
|
Alarm! Inverter problem
|
Topic: Stones Road house, technology | Link here |
After Daryl left, received a mail message from Ingeteam:
Serial number Start End Description Type
0AM172B16A22 (E10017220395) 4/17/2020 1:02:46 PM 4/17/2020 4:07:03 PM Grid voltage out of range Active for more than 1h
OK, the grid was disconnected. But was it more than 3 hours? No. In fact, it couldn't have been: this message was dated 17 Apr 2020 06:12:07 +0200, corresponding to 14:12:07 here. Where did it get the time “4:07:03 PM” from? Checked the inverter, which showed the correct time (now something like 15:30), but a gap in the readings:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
This was taken at 15:40. Things stayed like that until the time was past, and then carried on normally. What happened? My guess is that the inverter had a glitch on power up and got the wrong time, and then recovered. But the data had already been stored with the incorrect time.
|
More ports configuration issues
|
Topic: technology, opinion | Link here |
Mail from Yvonne today: would you please print out this PDF document?
No. Not needed. The real issue is that her MUA (mutt) won't display PDF. Didn't take long to find out why: the file /usr/local/etc/mailcap was missing. I must have forgotten to check it out after a recent upgrade.
|
Chasing the teevee X config problems
|
Topic: technology, opinion | Link here |
Why can't I start X from login on teevee? So far all attempts to catch any error message have failed. Today, thanks to Daryl Buchanan, I had to start it at least twice. And this time I did it from a shell on eureka. There was the message:
/usr/local/bin/X: Only console users are allowed to run the X server
Hmm. I think I've seen this before. Where did it go? Not even /dev/tty seems likely. What does it mean? Clearly not an interactive shell, since that was I was using. Off to search on the web, and came up with this article. Create a file /etc/X11/Xwrapper.config with the contents:
allowed_users=anybody
needs_root_rights=yes
Also read the man page Xorg.wrap(1) (admire the difference from the file name) for further details. And there I discovered that the file itself is relatively recent—it doesn't exist on eureka—and that these are the only two valid lines, and that the first one is the important one.
OK, off to create it. It's there! But X doesn't access it during startup. The last access timestamp doesn't change:
=== grog@teevee (/dev/pts/4) /var/tmp 11 -> l -uT /etc/X11/Xwrapper.config
-rw-r--r-- 1 root wheel 45 17 Apr 16:06:06 2020 /etc/X11/Xwrapper.config=== grog@teevee (/dev/pts/4) /var/tmp 12 -> startx
xauth: file /home/grog/.serverauth.1774 does not exist
/usr/local/bin/X: Only console users are allowed to run the X server
^Cxinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: unexpected signal 2=== grog@teevee (/dev/pts/4) /var/tmp 13 -> l -uT /etc/X11/Xwrapper.config
-rw-r--r-- 1 root wheel 45 17 Apr 16:06:06 2020 /etc/X11/Xwrapper.config
OK, off with stronger tools, in this case ktrace. That shows me:
1813 Xorg.wrap CALL open(0x2004be,0<O_RDONLY>)
1813 Xorg.wrap NAMI "/usr/local/etc/X11/Xwrapper.config"
1813 Xorg.wrap RET open -1 errno 2 No such file or directory
1813 Xorg.wrap CALL getuid
1813 Xorg.wrap RET getuid 0
1813 Xorg.wrap CALL access(0x7fffffffdd70,0x1<X_OK>)
1813 Xorg.wrap NAMI "/usr/local/bin/Xorg"
Wonderful! The name has changed. Copy the file to /usr/local/etc/X11 and all is well. Of course, if I had read the man page more carefully, I would have discovered that, but the article on the web and the presence of /etc/X11/Xwrapper.conf misled me. There's also the question of where /etc/X11/Xwrapper.conf came from. I certainly didn't install it, so it could possibly have come from the installation, though pkg doesn't want to know about it.
Well, that issue has been solved. But I still can't get my remapping done right. It seems that the default keymap is different depending on how you start X. Or more to the point, it seems to change somewhere between starting and the first xterm. So I need to set the keymap explicitly rather than as a series of changes.
Why is this all so hard?
|
Donald Trump's signature
|
Topic: politics, opinion | Link here |
It seems that the relief cheques (or should I write “stimulus checks”?) for people suffering from the US COVID-19 shutdown have been delayed: Donald Trump wants to have his signature on each of them.
Apart from the obvious questions about why they should be cheques rather than bank transfers, and why Trump should sign them, that's only half the story. As The Shovel reports, he's also putting his name on gravestones:
Department officials said it was unusual for government-issued gravestones to be altered in this way, but they said it was important for victims’ families to know that Mr Trump had played a key role.
The plan was initially to include President Trump’s name in small lettering at the bottom of the gravestone. But upon seeing the design, the President insisted that his name be much larger, and in gold type.
| Saturday, 18 April 2020 | Dereel | Images for 18 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Reshaping the garden
|
Topic: gardening | Link here |
Mick Solly along today to do some garden work, including what I think is the latest lawnmowing in the year. But there was more to be done, including removing the dead Buddleja (“it'll recover, won't it?”) and replacing it with a plant that I haven't been able to identify, which had planted itself almost against the south wall of the house:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Interestingly, it had very long lateral roots and almost none with depth:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It didn't occur to me until later that that might be ideal for the position: I had suspected that the plants had hit clay and given up.
Also removed the Boobialla tree from in front of the shade area in the north-west garden:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
Or at least that's what it should have looked like. In fact Mick started cleaning away the branches and then found something more interesting to do, so what I saw later was:
|
|||||||||||||||||||||||
Also removed an Arum clump that was growing directly outside the laundry door, and planted it both where we removed the mystery plant and also along the row where the Buddlejas died:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
If anything can live there, it must be Arum lilies, though possibly we should plant some Cannas too.
As if that wasn't enough, Mick also replaced the drain pump in front of the house, not without minor issues:
|
|||||||||||||||||||||||
Those are the two pumps (new above, old below) lying on their sides. They're very similar in construction, but the output is at a different height.
|
New gates
|
Topic: Stones Road house, general | Link here |
CJ Ellis also came along this morning—it was he who cut down the Boobialla—mainly to install gates. Now the front gate looks better, and we can also get from the driveway to the south of the house:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
We had planned to attach the gate opener as well, but discovered that we needed a 10 mm drill bit to attach it to the gate, so we'll have to postpone. But somehow we made a lot of progress today with the garden and the gates.
| Sunday, 19 April 2020 | Dereel | Images for 19 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Neighbourly contact in times of COVID-19
|
Topic: general, animals, opinion | Link here |
Linda Swift, our neighbour, along today bearing dog food. It seems that two of their dogs have died, including blue-eyed Beethoven (he was 16), and only Gizmo (whom I can't place) survives. And Gizmo is too small to eat the food, so ours got it.
|
Nikolai's paw
|
Topic: animals | Link here |
Nikolai injured his paw some time ago now, and it has been healing slowly—Pene Kirk tells us that that's normal. Here's the current status, along with the baby sock that Yvonne had found for him:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Cooking rice
|
Topic: food and drink | Link here |
How much water do you put in rice when boiling it? I've been cooking rice for well over 50 years, but I can never make up my mind. My old rule of thumb was a finger's width above the level of the rice, but clearly that depends on the depth of the rice. Others have said “one part rice, two parts water”, without even clarifying if by weight or (shudder) by volume. But that seems too much to me. On the other hand, my last batch came out a little undercooked.
Enough! Today I cooked 750 of normal long-grain rice with 1050 g of water, because it looked right. That's 1.4 times as much. The results were OK, but possibly a little more would do no harm. 1.5 next time?
This was normal long-grain rice. I suspect that Basmati would need a little more, and today's results are no help for short grain rice. Time for a web page on the subject.
Oh. I have one, which I set up only eight years ago. I really should read this stuff first. Now at any rate I have updated it.
| Monday, 20 April 2020 | Dereel | Images for 20 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More kitchen lights
|
Topic: Stones Road house, general, food and drink, opinion | Link here |
Chris Bahlo back from Ballarat this afternoon with a carton of wine and some adhesive tape that the post office couldn't find when she was there on Friday.
The tape was more important: it was for attaching the third strip light in the kitchen, underneath the range hood. Why not just repair or replace the range hood? The repair wouldn't last long, and I just can't find a range hood that meets my (normal, I would have thought) requirements. In particular, the light should be on the wall, not at the front, and the globes should be easily replaceable. And since I haven't given up hope that I'll find a suitable range hood, I don't want to drill into the tiles to attach this light.
So: connected up the light with the double-sided adhesive tape. It works! Now I just need to find a way to tidy up the cable mess. Mañana.
|
Spreading fertilizer
|
Topic: gardening | Link here |
Found some time to spread a bit of fertilizer this afternoon. I certainly didn't exhaust myself: the Grevillea robusta that I planted in the sewage paddock nearly a year ago hasn't died (unlike a couple of other plants that we bought at the same time), but if anything it's smaller than it was when we bought it.
|
|||||||||||||||||||||||
Yes, Proteaceae don't need much fertilization, but somehow nothing grows here. Hopefully the (“native”) fertilizer will change that.
And then a couple of Azaleas are already starting to flower in front of my bedroom window. They will benefit from fertilizer, as (hopefully) the rather sickly Camellia japonica in front of the house.
|
|||||||||||||||||||||||
| Tuesday, 21 April 2020 | Dereel | Images for 21 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Kitchen lighting, continued
|
Topic: general, food and drink, opinion | Link here |
I had put off taking a photo of the new strip lights in the kitchen until today. But I wasn't expecting what I saw when I came in in the morning:
|
|||||||||||||||||||||||
The adhesive strip had come unstuck. That in itself isn't that surprising, but the way it did was. Firstly, it had stayed (quite firmly) stuck exactly in the middle, and also on the ends of the light, and secondly it had stretched far beyond what makes sense. It had also done so completely symmetrically.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The completely unstuck part must have stretched by a factor of 4. Why do they make things like that?
What do I do now? The part that is still on the wall was across the gap between two tiles, which suggests that the adhesion to the tile was inadequate. Dirty? Uneven? I'll have to investigate.
|
Luminar again
|
Topic: photography, technology, opinion | Link here |
Note on the M43 Tech Talk Facebook group today: Luminar 3 (the previous version) is available for free. Haven't I tried that before? Yes. And I didn't seem to come to much of a conclusion. OK, download again, once again facing a lot of confusing information. But finally:
|
|
|||||||||
Restart your computer! Yes, that's normal enough in the Microsoft space. But dischord had been up for 34 days (even surviving Daryl Buchanan's power failures). Sigh. OK, reboot, try to find my way around this horrible tree climbing that the Microsoft space expects. And then:
OpenGL 3.3 and later is required for this application to run. Try updating your graphics driver.
I've seen that before. What it's really saying is “You're running remote desktop access,
and we're too stupid to handle it”. And then it crashed stopped working.
What do I do now? I suppose I could go to a direct display at least long enough to see if it's worthwhile. But somehow the promise doesn't make it worthwhile.
|
Robinia? Locust?
|
Topic: gardening | Link here |
What is the shrub that we transplanted on Saturday? Yvonne had thought that it was an Acacia, which it clearly isn't, and then came up with the German word Robinie. And Margaret Swan confirmed: it could be a Robinia. Off to take a look, and discovered that it's also called Locust. We've had that before, but it doesn't look anything like our current tree:
|
|||||||||||||||||||||||
Different species, maybe? In my diary entries I had identified it with Robinia pseudoacacia. Maybe the leaves will change.
|
Take my oil, please! I'll pay you
|
Topic: politics, general, opinion | Link here |
One of the side effects of the COVID-19 lockdowns has been serious issues in the transportation industry. Virgin Australia, an airline, has gone into voluntary administration as a result of their losses, and West Texas Crude oil, which was trading at $61 per barrel at the beginning of the year, briefly dropped well below zero. How much? $-40.32 per barrel?. Or maybe $-$37.63, as Wikipedia says? That's right, paying people to take it! It seems that the issue is storage and the cost of shutting down production. By the time I heard it, the price was round $2 a barrel, only a token sum.
Where will it lead? Time to start considering what will happen after COVID-19. Much could be positive: in principle, using less oil is a Good Thing. But on the way a lot of people will suffer and have to adjust. What will happen to the hospitality industry? Fast food will presumably be able to adapt to more deliveries and less in-house eating. And real restaurants? Yana will have to reconsider, for example.
|
Goodbye Nokia
|
Topic: general, technology, opinion | Link here |
I've pondered for a while, but there's really no reason to keep the new Nokia 5.1 Plus that I bought last month. Packed it up, and Yvonne can take it back to ALDI tomorrow.
| Wednesday, 22 April 2020 | Dereel | Images for 22 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden flowers in mid-autumn
|
Topic: gardening | Link here |
It's a month after the Equinox, time for the mid-autumn garden flower photos.
The existing plants in the garden are doing differently. The Schinus molle (pepper tree) that we transplanted 4 months ago is now well on the way to recovery:
|
|||||||||||||||||||||||
And the other one that we planted to the west of the garden is also coming along:
|
|||||||||||||||||||||||
The Hibiscus plants to the north of the house are also doing well. Here the Alyogyne huegelii:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
Then the Hibiscus rosa-sinensis “Uncle Max”:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And the Hibiscus syriacus:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
The trees are doing reasonably well. The one surviving birch in the east garden is now looking more like a tree:
|
|||||||||||||||||||||||
In previous years it had started to play “autumn” from January on. I didn't even take a photo this time last year,but here's a photo from January 2019, with noticeable yellow leaves:
|
||||||||||||||||||||||||
The Paulownia kawakamii is also now looking better:
|
|||||||||||||||||||||||
My guess is that in each case the combination of significant fertilization and irrigation has made the difference.
And the newly planted Corymbia ficifolia has barely stopped flowering since we planted it:
|
|||||||||||||||||||||||
There are plenty more buds. My guess is that this is a one-off because it was planted so late: normally they flower in the summer.
As last year, the Banksia integrifolia is producing multiple inflorescences:
|
|||||||||||||||||||||||
It's doing so well that I wonder if we shouldn't plant another.
The American Oak that we planted in the east garden now has very reddish-brown leaves:
|
|||||||||||||||||||||||
But maybe that's what it should look like at this time of year. Like the birch, I didn't take any photo this time last year, but the photo from March 2019 doesn't look anything like as healthy:
|
||||||||||||||||||||||||
The Clematis that we bought over 2 years ago are now both gradually flowering:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Again, I think a combination of irrigation and fertilization were the cause of the delay, though in this case ill-treatment also plays a role.
And the chili plant that I planted in October is bearing fruit, though strangely it hardly seems to have grown:
|
|||||||||||||||||||||||
The Buddleja bushes are all dead, but two of the cuttings I took of the last dying bush are still alive. This one is growing well:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The leaves at the top are all new growth. The other one is not as active:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Here nearly all the leaves were there when I took the cutting. But at least it isn't dead.
One bush that has never done well is the Camellia japonica in front of the house:
|
|||||||||||||||||||||||
It flowers, but I can't get the leaves green, despite plenty of irrigation and masses of fertilizer.
It's a similar story with the Gingko biloba in the north paddock, barely larger than when we bought it:
|
|||||||||||||||||||||||
I'm going to try lots of fertilizer on this one. It should be getting enough water.
One plant that I had almost given up on was the Fuchsia triphylla in the shade area. But at least I am now getting a couple of flowers from it:
|
|||||||||||||||||||||||
And the Fuchsias in the hanging baskets in front of the house are also flowering, but not profusely:
|
|||||||||||||||||||||||
Inside, Yvonne moved the Syngonium podophyllum from the bathroom to her bedroom last month. It's certainly not doing badly in its new position. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
My experiment planting them outside has not looked overly positive, but I'll give it time. In the meantime, the bath is empty. We also have two Spathiphyllum boxes, one in Yvonne's room and one that had been in the dining room. The one in Yvonne's room looked fine, while the other looked half dead. Trimmed off the dead leaves and moved it to the bath:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The only issue with Yvonne's plant is that it doesn't flower.
| Thursday, 23 April 2020 | Dereel | Images for 23 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Heating nasi lemak components
|
Topic: food and drink | Link here |
Nasi lemak for breakfast today. Heating the ikan bilis and peanuts in the “coffee machine” (small “air fryer”) seems to work well: 5 minutes at 160°, though potentially even less would be sufficient.
|
ISO sensitivity measurement continued
|
Topic: photography, technology, opinion | Link here |
Last week I looked at the first of two articles by Richard Butler of DPReview about the meaning of the ISO 12232:2019 standard on (to quote) “Digital still cameras — Determination of exposure index, ISO speed ratings, standard output sensitivity, and recommended exposure index”. Now the second instalment has appeared, and it, too, is very interesting.
As I noted last year, the original ISO 6:1993 standard based on the Hurter-Driffield curve describing film sensitivity:
Specifically, the sensitivity related to minimum exposure:
the "speed point" m is the point on the curve where density exceeds the base + fog density by 0.1 when the negative is developed so that a point n where the log of exposure is 1.3 units greater than the exposure at point m has a density 0.8 greater than the density at point m.
Last year I grumbled that I couldn't get ISO 12232:2019 without spending a lot of money, but the second half of the DPReview article helps. In particular, it shows this graphL
The green line (almost identical with the black one) shows the traditional tone curve, going linearly all the way to saturation, while the (new) black one takes advantage of changes in ISO 12232:2006 to effectively double the sensitivity of the camera without any further changes in the camera (in this case an Olympus E-620).
How does this work? As I see it at the moment, it's smoke and mirrors. The camera sets the exposure for the highlights, and this enables catching the same highlights with half the exposure. But the standard defines sensitivity in terms of the mid-tones. Something very similar happened in the 1950s with the ASA and DIN standards, where the conversion from one to the other changed several times due to different methods of measuring the sensitivity.
But Olympus was just the tip of the iceberg. Since then other makers have come up with different tone curves, in some cases (Fujifilm) even selectable from a menu:
That's still only part of the picture, but it's interesting enough in its own right. Now I need to think through the rest of the article.
In passing, I wonder whether ISO 12232:2019 has anything to say about silly sensitivity values like ISO 32,768.
|
Nikolai's paw
|
Topic: animals | Link here |
Nikolai's paw still hasn't healed completely, but it's making good progress:
|
|||||||||||||||||||||||
Yvonne is playing it carefully, and she'll leave his sock on for another few days.
|
More fertilizer
|
Topic: gardening | Link here |
Spread some more fertilizer in the afternoon. I really should try to spend more than 10 minutes at a time on the garden. But even so it seems clear that I need to apply much more fertilizer than I ever thought necessary.
| Friday, 24 April 2020 | Dereel | Images for 24 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Understanding adhesive strips
|
Topic: general, food and drink | Link here |
The adhesive strip that I bought to attach the strip light under the range hood did not exactly convince me of its usefulness:
|
|||||||||||||||||||||||
I've been doing some thinking about what the cause could be. It's fairly clear that it stuck better to the light than to the wall, which was confirmed when I removed the remains. I had half expected that: the back of the light is almost perfectly flat, while the wall is surfaced with ceramic tiles, which are slightly uneven. That's why I bought a foam based tape rather than a thin non-flexible version, though the latter probably would stick better on perfectly flat surfaces. So a good approach might be to first stick the tape on the wall and press firmly to adapt to the surface, and then remove the protective film and attach the lamp.
But after removing bits of tape I found:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
What's that white layer? I can't recall seeing that.
|
Shrinking the world
|
Topic: technology, opinion | Link here |
Saw on the overnight transcript of IRC: Jari Kirma was complaining about the ping time from Helsinki to Sydney, about 450 ms. OK, it is two thirds of the way round the world—isn't it? Try a traceroute:
=== grog@eureka (/dev/pts/8) ~ 111 -> traceroute www.hut.fi
traceroute to www.hut.fi (130.233.220.240), 64 hops max, 40 byte packets
1 radiation-tower.aussiebb.net (167.179.136.1) 36.516 ms 29.910 ms 38.899 ms
2 cmslive.aalto.fi (130.233.220.240) 310.082 ms 312.576 ms 310.011 ms
Huh? What's that? Yes, the first hop is the National Broadband Network link, but that's in Melbourne. Where are the intermediate hops?
A bit more experimentation and discussion showed that it was related to Aussie Broadband and sites in Europe. The same thing happened with my server in Frankfurt am Main:
=== grog@eureka (/dev/pts/8) ~ 112 -> traceroute ffm
traceroute to ffm.lemis.com (45.63.116.55), 64 hops max, 40 byte packets
1 radiation-tower.aussiebb.net (167.179.136.1) 35.984 ms 32.726 ms 27.059 ms
2 ffm (45.63.116.55) 300.089 ms 299.871 ms 290.954 ms
But the server in Los Angeles showed more normal results:
=== grog@eureka (/dev/pts/8) ~ 115 -> traceroute lax
traceroute to lax.lemis.com (45.32.70.18), 64 hops max, 40 byte packets
1 radiation-tower.aussiebb.net (167.179.136.1) 29.991 ms 29.765 ms 29.943 ms
2 be1.bdr1.coresite-sv1.sjc.aussiebb.net (180.150.2.107) 190.218 ms 189.664 ms 190.070 ms
3 HundredGigE0-0-0-8.core1.vdc03.mel.aussiebb.net (202.142.143.224) 178.982 ms 179.720 ms 180.060 ms
...
12 www (45.32.70.18) 194.859 ms 189.882 ms 190.789 ms
What is that? It seems that something is not adjusting the TTL values. Is this good? Bad? It certainly doesn't help understand the net topology. But at the same time as these (relatively consistent) times to Europe, others were getting slower results. Callum Gibson gets 340 ms; my times include the 30 ms latency to the next hop, so without that they would be round 280 ms, which is close to the 290 ms that James Fraser (also with Aussie) got: 40 ms better than Callum's. Is this some new kind of link? Is there maybe now a link to Europe that doesn't go via the USA? Why is the obfuscation needed?
|
Another PV recalibration
|
Topic: Stones Road house, general, technology, opinion | Link here |
Nice sunny day today, perfect for generating lots of electricity. But when I looked at the “dashboard” in the afternoon, the PV array was generating nothing.
Take a look at the database. No update since 25 March! That was the day that Fred installed the new firmware. Did I somehow mess up my sniffing code? Took a look, and all looked well:
Insert into powerstats set Id = "127.0.0.1:502:1", Pac = 640, SetPoint = 0, Alarms = "", Status = "On-grid (calibrating batteries - 2)", StatusCode = 3, SOC = 68, VBat = 264, PacBat = 750, PacPV = 0, FromPV = false, Codes = "", W1 = 7, W2 = 0, W3 = 0;
It also confirmed my suspicion that the inverter was calibrating again. OK, insert into the database. No column StatusCode. The new firmware update included a new database layout. I don't suppose I can complain about that, since it's not really intended for use by Mere Mortals, but clearly an indication that I should check things after firmware updates.
Added the column to the table, in the process also widening the Status column from 20 to 36 characters, and things went their merry way. This time, though, it didn't even do a complete cycle: it started round 14:18 after quite a period at 100%, and only went down to 0% before resuming normal operations at 17:11.
|
|||||||||||||||||||||||
I can live with that; I've already said that once a month seems acceptable, and that's exactly how long it has been since the update. In addition, the relatively short duration this time might be indicative of the presumed fact that they have fixed something (and thus that there was something to fix). It's also interesting to look at the battery power line (brown): during the day there were short bursts of charge of up to 1.15 kW, which could be indicative of incorrect calibration.
|
|||||||||||||||||||||||
Now they just need to feed the excess PV power through to the grid. To be observed, but for the moment I'm relatively happy.
|
Infrared photography revisited
|
Topic: photography, opinion | Link here |
It's been over 5 years since I bought an infrared filter and did some experiments with it. I wasn't very satisfied with the results, and I didn't look any further.
But in one circumstance it could be of use: my series of analemma photos. The sky is not always clear at noon, but a thin veil of cloud doesn't obscure the position of the sun. Could it be that the infrared filter would produce clearer images? Off to try things out today, also with images unrelated to the analemma, and learnt a number of things:
Use white balance! I didn't with the photos 5 years ago, so they all looked red. With a white balance adjustment they came out mainly greenish.
The E-30 no longer seems to be to most sensitive. The E-M1 Mark II was round 3 EV more sensitive.
The colour rendition of the E-30 and the E-M1 Mark II is significantly different. Here the same scene, first with the E-30, then with the E-M1 Mark II:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In passing, the exposure of the E-30 image was 3.4 EV more than the E-M1 Mark II.
The infrared filter I have has horrible reflections. Here a comparison with the grey filter, also showing the difference in exposure (both 2 s at f/8):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In passing, it's sad that Olympus has disowned the E-30. They used to have a page on their web site, but now it's gone. Clearly they don't have the space for such old equipment.
| Saturday, 25 April 2020 | Dereel | Images for 25 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
ANZAC Day sunrise
|
Topic: general | Link here |
By chance, up this morning at 6:40 to let the cat out. Beautiful sunrise:
|
|||||||||||||||||||||||
And today was ANZAC Day, the celebration of a great defeat in the First World War 105 years ago. For some reason, it's traditionally (for the last few decades, anyway) celebrated with a dawn service. Not this year, due to the COVID-19 outbreak. But somehow I was faced with it, apparently 16 minutes before dawn, though the services (presumably based on dawn in Sydney or some such place) were already over.
|
No boobialla extraction
|
Topic: general, gardening | Link here |
While I was taking my weekly panoramas, CJ Ellis showed up unexpectedly with a small tractor and chains to pull out the stump of the Boobialla that we cut down last week:
|
||||||||||||||||||||||||
Problem: the chains weren't long enough. So off he went again, planning to come back some time to poison the roots first.
|
Infrared or not?
|
Topic: photography, opinion | Link here |
Yesterday's infrared photography experiments left a number of questions open, in particular the original one: does infrared photography cut through clouds better? Today tried a comparison:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Somehow it doesn't really tell me anything.
|
Making fun of the US American people
|
Topic: politics, opinion | Link here |
On many occasions I have wondered whether Donald Trump can be as stupid as he appears to be, or whether his candidacy for President of the USA was not simply an elaborate joke to see how far he can get. Today he suggested injecting or ingesting disinfectants. That way death lies, as the makers of the products were quick to point out.
Ah, but he didn't really mean that. He was just being sarcastic. And here's me thinking that he had no concept of humour or anything similar. The perfect straight man! I wonder if even he thought that so many people would believe his nonsense.
Of course, there's no real danger, as The Shovel points out.
| Sunday, 26 April 2020 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The liar tweets tonight
|
Topic: politics, opinion | Link here |
One of the more amusing parodies I've seen lately:
|
Buying cheese online
|
Topic: food and drink, opinion | Link here |
It's not just COVID-19 that is keeping us from going to Melbourne and buying various groceries that we can't get in Ballarat: we hate the journey, and the last two times (both in 2018!) were because Yvonne needed some medical examinations. In the last 4 years we have been shopping there exactly 3 times. And apart from the distaste, there's the cost: 300 km there, around town, and back, and a day spent mainly on the (congested) roads. Why not pay a little more for delivery and buy things online?
Spent some time today looking for cheese for Fondue. What do I need? A good Gruyère, Appenzeller, maybe Vacherin? Off looking for online suppliers, and found no fewer than 8, creating a groceries page to store them in.
How good are they? As far as I can tell, useless. They all refuse to give you a list of their stock, offering silly photos of what they think you want (“Gabrielle Coulet Roqueforte” [sic], “Mini Cheese Tower”, “More than Cheese Gift Box”, subscriptions). Only about half of them deliver, and their prices are all significantly higher than I'm used to. Tried searching on some of the more promising ones, and found no supplier with both Appenzeller and Gruyère. No Gruyère? What kind of cheese shop is that? My guess is that they do have it, but can't be bothered to maintain their web site.
OK, alternative: try to order the cheese we're looking for at Campana's, (“located locally”), which is also our local cheese shop.
Who says that Australia is a culinary paradise? You can buy 100 varieties of “Tasty” and fake Cheddar, but there's no variety. We've completely given up on French cheeses.
| Monday, 27 April 2020 | Dereel | Images for 27 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Tracking COVID-19
|
Topic: health, technology, general, opinion | Link here |
COVID-19 is dangerous for many reasons, but probably the worst is the combination of extreme contagiousness and the difficulty identifying carriers. That's the main reason for the quarantines that have crippled the world economy.
Now that the number of infections here in Australia has dropped almost to zero, the question is how to keep it that way but to allow people out of their houses.
Mobile phones to the rescue! The government has brought out an app called COVIDSafe, which tracks nearby users via Bluetooth and logs (on the phone, for privacy reasons) any such encounter that lasts more than 15 minutes. If then anybody involved then “tests positive” for COVID-19, you can find out somehow that you could be in danger.
All sounds very good, but there are down sides. The big one is that people are concerned about privacy. That's why the data is only stored on your phone. I still don't quite understand how you can be notified. From the information page:
The information is encrypted and that encrypted identifier is stored securely on your phone. Not even you can access it.
When someone is diagnosed with COVID-19, state and territory health officials will ask them or their parent/guardian who they have been in contact with. If they have the COVIDSafe app and provide their permission, the encrypted contact information from the app will be uploaded to a highly secure information storage system. State and territory health officials will then... use the contacts captured by the app to support their usual contact tracing.
But how? If the information is only available after somebody tests positive and provides permission, how can they track anybody?
And how long must the encounter last? When I installed it I saw something about 15 minutes,
but the information page doesn't specify a duration. 15 minutes seems very long to me; it's
a very different viewpoint from “keep your 2 m 1.5 m distance at all times”.
And then there's this:
The COVIDSafe app does not collect your location.
Why not? Consider the perfectly possible scenario where two users go shopping somewhere, and somebody in the shop, not using the app, is spreading the virus. One gets infected, and if somebody can answer my first question, the second user gets warned in time. And that's all! If there are 10 people affected, there's a good reason to believe that it's related to the location, but nobody can remember what that is, so the infections continue as long as the carrier is healthy enough to stay there.
Yes, privacy is good. Survival is better. Wouldn't it be a good idea to have a more draconian approach: you can only leave quarantine if you're running the app, and you must take into account that your location will be tracked. And if you can't afford a mobile phone, you can get a very basic one, without SIM card, cheap from some government stockpile (old ones, for example, that nobody would really want to use if they had the choice).
Still, it seems that Australia is the first country to introduce something like this, and it has made people sit up and listen. There have been about 2 million downloads so far, roughly 10% of mobile phone users. The new took centre stage on the German Tagesthemen news programme today, and it seems that Germany is thinking of doing something very similar. The programme also explains what the Australian page doesn't: the upload of infection information generates a signature that the individual phones need to check. What an effort in name of privacy! And it's not clear that that's the way we do it in Australia.
|
Installing the new server, continued
|
Topic: technology, opinion | Link here |
It's been nearly two weeks since I paused my system installation efforts. It wasn't just the pain; I had other things to do. But today I finally got round to continuing.
In fact, installing ports wasn't that unsuccessful. My simplistic ports identification attempts didn't help, but the only ones that failed were: apache, ghostscript9, lame, nvidia, p5-, php, samba, vbox, xkeyboard, xtset, xmlprprxmlrpc, xproto. I forgot to catch the error output, so my logs contain the relatively uninformative:
=== apache
Updating FreeBSD repository catalogue...
FreeBSD repository is up to date.
All repositories are up to date.
===
The important thing missing is the information that the port was installed.
Clearly there are issues there. p5- is the prefix for far too many Perl-related ports, potentially hundreds. I have a complete list of what's installed on eureka, so a good start might just be to ignore them and install them as they become necessary. nvidia is a set of ports that I know only too well, but before I install them I need to know what graphics card I'll put in the box.
Today I installed apache24 and went looking for PHP ports. How many can there be?
=== root@dereel (/dev/pts/2) /home/grog/Port-upgrade 47 -> pkg search php | grep -i ^php | wc -l
1849
1,849 ports starting with php, but none called php. How do I decide? In the end installed php74. From memory, there will be more needed, but how do I find them? Probably it will involve reading back in this diary.
|
Simple, obvious, ... wrong
|
Topic: technology, opinion | Link here |
While installing things on dereel, decided that it would be time to upgrade the ports on teevee. That's straightforward enough: pkg upgrade.
Oh:
New packages to be INSTALLED:
ImageMagick6: 6.9.10.90,1
...
Installed packages to be REINSTALLED:
emacs-26.3_3,3 (direct dependency changed: ImageMagick6)
Callum Gibson could be right: pkg knows nothing of the fact that my Emacs depends on ImageMagick 7, not 6. OK, once and for all, let's fix that. I have already modified the Makefile, so apart from checking that it does now build and run on teevee (last month, for some reason, it didn't), and then commit.
Oh. Various complaints. Firstly, there's an emacs group who do this stuff, so I should have asked them. And they had specific reasons to keep ImageMagick 6, including the fact that all KDE ports still use it (why?). It looks as if the problem is much more general than I thought.
And the pkg upgrade? In the end I upgraded those ones that didn't want to change my view of the world. Now I only have a couple of minor library upgrades, and they can wait until the pkg database is updated.
| Tuesday, 28 April 2020 | Dereel | Images for 28 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
COVIDSafe: Only for people with money
|
Topic: health, technology, opinion | Link here |
Fired up my old Samsung GT-I9100T mobile phone today, with intention for Yvonne to use it for the COVIDSafe app. No expectation that she would use it as a phone; like all Android phones, it's far too clunky, and her old steam phone works just fine.
How I hate the user interface on mobile phones! Somehow I didn't manage to spell “COVIDSafe” correctly, and all I got was this message:
|
|
|||||||||
Well, work harder then! Tried again in a different way with the same result. Finally checked at the toy shop, where I read:
This app is compatible with some of your devices.
Some? Which ones? Ah, that would be telling, just like saying on the Samsung “We've found an app of that name, but it's too polite to work on your crappy old phone”.
Admittedly, I later discovered that you have to prod the page in the right place, after which you get the information:
|
|
|||||||||
People, this is supposed to be a life-saver. Even if you only want to save the lives of people who can afford newer phones, the dying paupers can infect you too. Ugh.
|
Mushrooms out of season
|
Topic: gardening, opinion | Link here |
Seen in Harrisons Road today:
|
||||||||||||||||||||||||
For some reason I thought that they were out of season, but looking back through my photos it seems that they grow from April to July, so the season is just beginning.
| Wednesday, 29 April 2020 | Dereel → Ballarat → Dereel | Images for 29 April 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another grid power outage
|
Topic: general | Link here |
Another short (1 second) grid power failure this morning at 2:46:34. It's interesting because it's the first outage of any kind in over 6 weeks. I don't know when that last happened.
|
Shopping in Ballarat
|
Topic: general, opinion | Link here |
Into Ballarat again today for shopping, the fourth time since the curfew started. That's only marginally less frequent than in the 2nd half of 2019 (23 times). But I had to buy some stuff that Yvonne would not be able to decide on.
When can I get my hair cut again? On the way into town took a look at the barber's shop. Open! Somehow the restrictions in Australia, if I understand them correctly, are not nearly as draconian as elsewhere. I didn't have time to have my hair cut then, but it's good to know.
First to The Good Guys for a beard trimmer. The place looked pretty much as usual except for the yellow markings on the floor at roughly 1.5 m intervals, some of which the people respected.
On the roads the obvious things were the petrol prices. Before the pandemic started, a litre of petrol cost about $1.45, and up to $1.60 in some places. Today I didn't see any over $1, and this was the cheapest:
|
||||||||||||||||||||||||
That's only about half of the top-of-the-range before the pandemic.
Then to Bunnings, braving multiple barriers to entry:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
They have even beefed up their entry maze to make even more sure that you come in contact with some surface that could have had contact with one of the 9 cases of COVID-19 in the Ballarat area:
|
||||||||||||||||||||||||
The entrance is in the middle, and you have to walk all the way to the left and then back again. What makes people think that this can help at all with protecting from infection?
At Bunnings I tried out my new face mask, the kind that is recommended for everybody. I was the only one. And I can understand why people don't like them. Apart from the fact that mine kept slipping, my breath came out under my glasses and misted them up. Not a thing that I would want to do frequently.
I had intended to replace my cordless drill. The real reason is that I have lost the battery charger for the drill. It's possible that with much searching I could find it again, but why not buy a better drill?
I found out why. The power of the drill is reflected in the voltage of the battery packs. The cheap ones are 12 V, intermediate 18 V (like my current drill) and high-end ones are 24 V.
I later discovered that my drill is a 14.4 V model, something that I didn't see at Bunnings at all.
I didn't get to look at the high-end ones, which run at up to $500, far more than I expected to pay. And all the others were 18 V or 12 V. Finally found one, rather more expensive than I had wanted to pay (round $89, from memory):
|
||||||||||||||||||||||||
But look at that fine print:
|
||||||||||||||||||||||||
I suppose that makes perfect sense. Why buy multiple batteries and chargers for multiple tools when a single one would do? But that took the whole thing beyond what I'm prepared to pay. Maybe I can find a charger for my current drill.
| Thursday, 30 April 2020 | Dereel | Images for 30 April 2020 |
| Top of page | ||
| previous day |
|
Infrared comparisons
|
Topic: photography, opinion | Link here |
Last week I took a couple of infrared comparison photos with the Olympus E-30 and the OM-D E-M1 Mark II. The colour rendition and sensitivity were very different. And somehow they didn't match my experiments of 5 years ago. On that occasion I decided that the E-30 was more sensitive than the E-PM2 and the E-M1 Mark I. But of course I didn't compare those cameras last week.
OK, I now have 6 cameras at my disposal. Time to compare them all. Out into the bright sunshine to take a photo of the north of the house:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But that's as far as I got before the bright sunshine turned to rain. So a bit later I positioned the tripod just inside the doorway of the terrace door and tried again. It's not worth displaying the results here. All of the mirrorless cameras behaved as good as identically. Here without the filter, E-30, E-PM1, E-PM2, E-M1 Mark I, E-M1 Mark 2, Olympus OM-D E-M5 Mark III:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The exposures were also interesting. The E-30 was 4 s at f/5.6 (EV 3 at 200/24° ISO). The others were all more sensitive. Some were on auto-ISO, but reduced to 200/24° they were 4.4, 4.4, 4.9, 4.5 and 4.4 EV, an average of about 1.5 EV more than the E-30.
So why did I get the opposite results 5 years ago? One possibility is that the E-30 I used today is not the same camera as 5 years ago; maybe they changed something in that area between building the two cameras. For the sake of comparison, it's interesting to note that the infrared filter reduces the sensitivity of this E-30 by about 9.5 EV, and about 8 EV for the others.
|
One Amanita muscaria? Many!
|
Topic: gardening | Link here |
As I noted on Tuesday, Amanita muscaria are just starting to come into season. But they're doing it with a vengeance, maybe as the result of the cool, moist summer:
|
||||||||||||||||||||||||
In the process, got a photo of one with the infrared filter in place:
|
||||||||||||||||||||||||
|
Frikadellen
|
Topic: food and drink, opinion | Link here |
We make our own hamburgers, but there are other ways of cooking mincemeat. When I was a lad we ate rissoles. Or did we? Looking for them today, I discover that a rissole is “a small patty enclosed in pastry”. That's clearly US language, but my British cookbooks agree. But my recollection is just mincemeat and onions. Why is it different?
In any case, what I was really thinking of was a mincemeat lump thicker than a hamburger, and served (cooked?) with gravy. German frikadelle? Off looking for recipes and found many. This one looked like a good starting point, so I tried that, modifying the recipe somewhat:
| quantity | ingredient | step | ||
| 500 g | beef mince | |||
| 90 g | dry white bread | |||
| 150 g | onion | |||
| 1 | egg | |||
| 1.3 g | oregano, dried | |||
| 1.2 g | tarragon, dried | |||
| 10 g | salt | |||
| 20 g | Thomy hot German mustard | |||
| 7.5 g | chopped parsley |
The complete details (updated), are in the recipe, but basically it was a matter of mixing everything together, forming into flattish balls (thicker than hamburgers) and frying. Here a deep frozen hamburger patty, a deep frozen and then a cooked frikadelle:
|
|||||||||||||||||||||||
But where's the gravy? Yvonne says there should be no gravy, but I recall from my time in Hamburg that I had had them like that—or were they something else than Frikadellen? I couldn't do a web search: “Hamburger Frikadellen” generates a ridiculous amount of false positives that have nothing to do with Hamburg. In the end I thawed out some left-over gravy and poured it over the Frikadellen, and I think it improved them. But it's still not what I have in my mind's eye.
Still, they tasted remarkably good, better than I had hoped for. In particular the texture was much lighter than a hamburger, presumably because of the bread. I'm reminded of the story:
A butcher lay dying, and called: “Hans, my son, are you there?”. Hans: “Yes, father”. Father: “Hans, you inherit the shop. I have an important secret for you: if all goes wrong, you can make sausages with meat”.
Next time I'll probably use more spices. And with any luck I'll also find the recipe I was really looking for.
In passing, it seems that my recollection of rissoles from my childhood was not so incorrect: the Wikipedia page that mentioned the pastry also has the Australian and New Zealand variant which is pretty much what I remember. Andrew Perry and James Fraser on IRC also confirmed.
|
Another teevee hang
|
Topic: multimedia, technology | Link here |
X hung again on teevee this evening, again with messages that suggest hardware problems, either with the system itself or with the display card:
Apr 30 18:58:52 teevee kernel: NVRM: GPU at PCI:0000:01:00: GPU-85983119-e9ef-ac66-6817-fbed5657b871
Apr 30 18:58:52 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 0000045c 0001002a 00000007 00000000
Apr 30 18:58:52 teevee kernel: NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000088 0001000a 00000007 00000000
Apr 30 18:59:03 teevee kernel: NVRM: Xid (PCI:0000:01:00): 39, CCMDs 0000000c 000090b5
It would be really nice to know what that means, but it looks almost identical to the hang I had three weeks ago. Time to think of hardware replacement, but which? My guess is the display card.
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 |
