|
| Greg's diary |
| June 2013 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Saturday, 1 June 2013 | Dereel | Images for 1 June 2013 |
| Top of page | ||
| next day | ||
| last day |
|
Firefox on hold for now
|
Topic: technology | Link here |
Into the office this morning to find that the debug build of firefox had finally completed. Ran it:
=== grog@monorchid (/dev/pts/1) ~ 2 -> firefox
=== grog@monorchid (/dev/pts/1) ~ 3 ->
Well, at least it didn't crash. But it seems that the debug build is more than just debug symbols, and it's not clear why it didn't continue. But I had other things to do, so left it at that.
|
Insight into Hugin stitching modes
|
Topic: photography, opinion, technology | Link here |
It was the beginning of winter and coincidentally house photo day today. Despite the forecast, it wasn't raining, and there was almost no wind, so I ended up with some relatively good if dreary images:
Once again I experimented with the various stitching methods that Hugin offers. I'm gradually understanding the different stitching methods that I looked at last week.
Of course, I have to confess that I didn't RTFM. Not that that helps much: the page shows the stitcher menu and claims “If you click the Create panorama button the new stitch popup sub-screen will be displayed. This will give you an easy to understand overview of the options.” In other words, no explanation. There's more in the documentation of the Stitcher Tab, though. In particular, it seems that the “Assistant” is doing the right thing by exposure fusing from “any arrangement”:
If Exposure fused from any arrangement is enabled then hugin will seam blend images with similar exposure with enblend and than it will exposure fuse them using enfuse. This variant is often much more successful than Exposure fused from stacks in two situations:
...
When the panorama has been shot entirely on automatic exposure, in this situation it is useful to seam blend adjacent photos with small EV differences, but then exposure fuse larger EV differences.
And that's exactly what I've been doing. It also explains the intermediate images. Today I did a full spherical panorama of the garden centre view and ended up with no less than 6 intermediate images. Here a couple. Move the mouse over each image to change it to two others:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
How well does it work? I still can't make up my mind. The exposure fusion gives greater apparent dynamic range, but it can appear a little strange. Here the garden centre again, first with normal output and then with exposure fusion. Again, move the mouse over each image to change it to the other:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
As I noted last week, somehow the Salvia microphylla on the right is too bright, and the sky looks too dark. This might be the case, though; it was very overcast.
Now that I have a second real machine (monorchid) up and running, did some of the panorama processing on it. It helped, but I ended up with a surprising number of broken images. It's not clear how that happened; it wasn't repeatable.
|
Old, forgotten slang
|
Topic: language, history | Link here |
A couple of days ago Jari Kirma came up with some Finnish slang on IRC, and it reminded me of some of the slang that I learnt from my father when I was young. I don't know where he got it from, but I suspect from the RAN in the Second World War. A couple I recall are “Up in aunty Aggie's room, behind the tango clock”, presumably a location for something lost, and “Black as the arsehole of death”, a simile that even at the time seemed inaccurate. But I mentioned the latter, and Jari went off looking for it on Google. Only a single hit! And that in my own diary for June 1967! Possibly there are other references in the parts of my diary that I haven't put online, but I'm surprised that I couldn't find any reference at all from other people. I must keep a page on this sort of thing.
| Sunday, 2 June 2013 | Dereel | Images for 2 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Firefox: success?
|
Topic: technology | Link here |
Tried my latest firefox build again today. It worked! Well, sort of. It might be that I forgot to specify a DISPLAY environment variable yesterday, though when I did so today I got a corresponding error message. But it produced voluminous debug output and... worked. Another Heisenbug chased away by debug code. I have better things to do than chase this stuff, so I did them.
|
HDTV, Australian style
|
Topic: multimedia, opinion | Link here |
A couple of months ago Malcolm Turnbull, while presenting the Government Opposition's castration of the National Broadband Network, made the claim that you could stream high quality HDTV at only 6 Mb/s. That probably says more about your definition of “High Quality” than anything else. SBS seem to be out to prove it. Their presentation of Al Jazeera news is a good example. This is a 1080i ”high definition“ stream which arrives here at a rate of 10 Mb/s:
|
||||||||||||||||||||||
|
|
||||||||
|
|
||||||||
To be fair, it's not all that bad, but it's amazing what people think is acceptable. Time to find another feed for Al Jazeera.
|
Garden in early winter
|
Topic: gardening | Link here |
I haven't been paying much attention to the garden lately, but it has recovered well from the summer (especially the weeds), and a number of flowers are still doing well. Here some that may not make it to the middle of the month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And of course individual Narcissus keep popping up:
|
|||||||||||||||||||||||
| Monday, 3 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Al Jazeera issues
|
Topic: multimedia, opinion | Link here |
After yesterday's disastrous rendition of Al Jazeera news, I recorded it on ABC News 24 instead. Not an improvement. The actual resolution was lower (720p instead of 1080i), it was twice as long (different news programme), it was in the middle of the night, and thus 12 hours old, the bottom of the image was overlaid with ABC information, thus obscuring part of the content—and the resolution was no better! Clearly the real issue is with Al Jazeera and not with SBS or ABC. So back to SBS, at least until I can find a better feed.
|
More slang research
|
Topic: language, history | Link here |
So the term “Black as the arsehole of death” only has two hits on Google, as I expected. But I have a dictionary of slang, an older edition of Eric Partridge's A Dictionary of Slang and Unconventional English, so looked in there. No hit, of course, but some similar ones: “Black as the Earl of Hell's riding-boots (or waistcoat)”, and directly below “Black as Toby's arse”, both implying a pitch-black night. My expression lies somewhere between the two of them, and the latter comes from a naval or nautical background, so it seems that my assumption is correct. I'm still surprised it hasn't surfaced elsewhere.
|
Tidying up the garden
|
Topic: gardening | Link here |
I'm still paving the way to hell with good intentions in the garden, but today I got a little work done, mainly rearranging plants and throwing out dead pot plants. Our Meyer lemon had a moment of glory 2 years ago, but since then it hasn't borne any fruit, presumably because it wasn't getting enough sun. So now I've put it on the north verandah, where other plants died because of too much sun. It's not likely it'll bear anything at this time of the year, but we'll see.
| Tuesday, 4 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Firefox: works
|
Topic: technology, opinion | Link here |
Back to the firefox problem today. I have established that I no longer get a SIGSEGV on startup if I build with debugging support, which goes beyond mere symbols in the binaries. It also produces lots of debug messages. The fact that the resultant binary doesn't fail suggests that the problem could be a race condition of some kind. So: back to a build without symbols.
Problem: it worked. I can no longer reproduce the problem. How I hate that! It usually means that the bug has gone into hiding, but hasn't gone away. Potentially some minor change in the system software in the nearly 2 weeks since I first ran into the problem might have “solved” the problem.
Then Jürgen Lock pointed me at the freebsd-ports mailing list, to which I really should have been subscribed. There are at least two other bugs that have been reported there, one compiler-sensitive (works with clang or gcc version 4.8, fails with gcc 4.7), the other library-sensitive (works with libc.so.7, fails if libc.so.6 is present). Neither appear to have anything to do with my bug. But what kind of situation is that? Somehow stability seems to be getting worse in the course of time.
|
Chicken liver, Indian style
|
Topic: food and drink, opinion | Link here |
Yvonne wanted to cook chicken liver as a main course today. Generally we just fry it, but that's a bit boring, and I went looking for more interesting recipes. Finally found one in Wendy Hobson's “The classic 1000 Indian recipes”. That's one of those books, like “500 Resep Lezaaat [sic] Masakan Indonesia”, where the number of recipes seems to be the most important factor. Finally put together a recipe which proved to be rather more than boring. I still don't have a good interesting recipe.
| Wednesday, 5 June 2013 | Dereel | Images for 5 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Washing machine recall
|
Topic: general, opinion | Link here |
The Sλmsung washing machine that we bought two years ago has done its job well, and we have no reason to complain. But Sλmsung sees it differently: it seems it has a potential electrical fault, and they've issued a recall for a number of models, ours possibly included. The instructions stated to check whether it's the right model and then call a free phone number.
Checking the number seems trivial. It's written on the top of the machine: SW75V9—what a mouthful. But the closest model number I could find was SW75V9WIP/XSA, twice as long. And that's on the label somewhere on the machine. They gave a convenient diagram of how to find it:
|
||||||||||||||||||||||
What does the diagram on the left represent? They don't say. Dug around the back of the machine, and found the label, more despite the diagram than because of it. My guess is that the diagram on the left is intended to represent the back of the machine, but they don't say so, it doesn't look like the back of the machine, and the aspect ratio is wrong.
Still, the label said that yes, indeed, I had SW75V9WIP/XSA, serial number 18795DBZB00906H (though they don't ask for the serial number)—I'm tempted to think that WIP stands for “Work In Progress”. So I called up Sλmsung on 1 800 239 655 and spoke to somebody who, on the third request, finally gave me her name clearly enough for me to write it down: May. And of course she wanted the serial number, though the recall notice didn't say so, and also the date of purchase. I didn't give either, and she processed the request anyway. Not without difficulty: I could barely understand her. My guess is that she's a Filipino, and her English isn't very good. Somehow that seems to be almost a job requirement for these people. On the other hand, I suppose people in the Phillipines or India who do speak near-perfect English can get better jobs than in a call centre.
So now I have both a reference number 8212699259 and a Service Order Number 4153015762, and I must wait for a call in the next 72 hours (3 business days), as if it were a duty. It'll be interesting to see how long it really takes; but the last Sλmsung repair (to a monitor) happened quickly.
|
Finally the mixer tap
|
Topic: general | Link here |
Last August I finally bought a new mixer tap for the kitchen sink. Installation wasn't easy: everything had different dimensions from the old one, and finally I got Bryan Jackson to install it for me—and he wasn't able to tighten the screws either, because the thing has 9 mm nuts which can only be turned with a box spanner. Most such things have 8 mm nuts. And you can't easily buy a 9 mm box spanner. The result was that the tap wasn't attached to the sink as tightly as I might have desired. And over the course of the months it has loosened considerably.
A couple of days ago I finally bit the bullet: bought a set of box spanners on eBay for $26, a considerable proportion of the original $58.50 cost of the mixer itself. They arrived today, and finally, after nearly a year, the thing is properly attached. What a time it takes me to do things nowadays.
|
Web search: brute force and ignorance
|
Topic: technology, opinion | Link here |
Should I keep my diary in text files or in a database? Clearly the professional approach is a database, but I've never quite got round to doing it. One of the nice things about keeping my diary in files is that I can use standard tools like grep to search for things, something I do relatively frequently.
But that only works at home. Two weeks ago I was at Jenny Bartlett's and wanted to find a diary reference to Android using her Android tablet. A complete pain, especially since I didn't recall the date. What I need is a web-accessible grep.
Over the last couple of days I've written something like that. It's a brute force “read all the files” approach, hardly the most elegant. But it has the advantage that I can do it quickly and play around with it if I don't like it. With a database I would feel more constrained. On the other hand, it's ugly enough that I'm not going to point at it. I really should consider more elegant solutions.
|
Flowers in winter
|
Topic: gardening, opinion | Link here |
The summer was pretty terrible, and lots of things didn't flower at all. But now things are coming out again. The Hibiscus rosa-sinensis that looked so unhappy two years ago is now flowering:
|
|||||||||||||||||||||||
So is the Mandevilla:
|
|||||||||||||||||||||||
Admittedly that's in the greenhouse, but the temperatures are still quite low, and it's a far cry from 2 years ago.
| Thursday, 6 June 2013 | Dereel | Images for 6 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dereel doesn't want a mobile tower
|
Topic: general, opinion | Link here |
Astonishing news from Julie Lee on the Dereel Facebook page today (my links):
The Premier was in town today and produced a letter from Telstra saying that residents of Dereel did not want a phone tower. The Mayor has asked that anyone in Dereel that would like a telstra phone tower to send a letter into the Golden Plains Shire stating that you require a phone tower. Please mail to Golden Plains Shire Council, PO Box 111, Bannockburn Victoria 3331 or you can email to the mayor cr.jblake@gplains.vic.gov.au Please state that it is about a phone tower for Dereel, please tell all your neighbours and friends who may not be on facebook. Please do this as soon as possible
OK, this is Telstra's claim. We know it's not true. But how did they get that kind of idea? Sent in my message as requested, copying Denis Napthine, whose staff seem to confuse web forms with email. But despite the web site we dragged up two email addresses: premier@dpc.vic.gov.au. and denis.napthine@parliament.vic.gov.au.
|
Subdivision after all?
|
Topic: general, opinion | Link here |
We're still looking for land to build a house on. Yesterday Yvonne returned from shopping to say that there was a block of land for sale just round the corner on the Rokewood Junction Road. That proved to be incorrect: there's a house on it. But what did I read in the description?
With two road frontages and being in a Rural Living Zone it presents a perfect opportunity for subdivision (STCA).
How can that be? It's in the same zone as we are, and according to the council the minimum lot size is 8 ha (or 20 acres, as we say in metric parlance). This property has only 6.2 ha, while ours is one of the few with the requisite 8 ha. Called up Jarrod Hodgson, whom we met a couple of months ago in Snowgum Road, and he confirmed that he had seen a surveyor's report stating that it could be subdivided. He promised to find out more and call me back tomorrow. In the process discovered that STCA means “Subject To Council Approval”, which, I fear, will not be forthcoming.
|
Retirement: rejected
|
Topic: general, opinion | Link here |
Astonishing letter to Yvonne in the mail today:
|
||||||||||||||||||||||
|
||||||||||||||||||||||
“On 20 September 2012 you applied for a retirement pension. You have not replied to our letter of 27 May 2013, so we have cancelled your application”.
What kind of nonsense is that? It's not clear what letter they're referring to; the date they state is the date of this letter, written differently to be confusing. Yvonne did respond to the last letter; it's not clear what has gone wrong, but it makes it clear that incompetence isn't limited to our immediate surroundings. The good news is that she has 2 months to appeal the decision, and clearly the appeal will be accepted.
|
Piccola in catfight
|
Topic: animals, opinion | Link here |
Watching TV the other night, we paused playback for some reason or another. And then I saw Piccola on the floor with a tail three times the usual diameter. The reason: a tortiseshell cat, the size of a leopard, looking straight at her from the TV.
That's surprising for a couple of reasons. Firstly, most animals don't recognize things on TV. I know that cats do, but conventional wisdom has it that they only respond to motion. Even more surprising is that today she's still wary of the TV, in case the cat might still be hiding inside.
|
Interrupt timeout, channel dead
|
Topic: multimedia, technology | Link here |
Watching TV today, the playback tripped over some kind of data corruption, after which I couldn't play anything back. It's not the first time, but this time I looked at the system log and discovered:
That's the audio output channel. And there seems to be no way to revive it. Reboot time. And clearly, since it has happened more than once, time to investigate more carefully. It seems that there have been a number of reports of this problem, most recently probably this one. But that's been committed.
| Friday, 7 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
PHP: planned obsolescence
|
Topic: technology, opinion | Link here |
About a week ago the Oly-e web site went down, with messages like:
No prizes for guessing where that came from. Reported it to Reinhard Wagner, who runs the site, and he told me that his ISP had sprung a PHP update on him without warning. It was clearly a big problem for him, and today he sent out a cry for help, including a warning that it might be enough to shut down the site for good.
Fortunately he found somebody to help, and the site is back online. But what's the problem? PHP's attitude “use the latest and greatest”. I ran into this problem a year ago: the latest versions of PHP deprecate POSIX regular expressions. This kills phpMyEdit, a program that I use for quick and dirty MySQL table editing. I haven't been able to fix it: as a result, I run an old version of PHP in a virtual machine, and it'll stay that way until somebody updates phpMyEdit or I find a replacement.
But why does the PHP team do this? Yes, it's good to recommend moving away from old cruft, but that doesn't mean you should break things. The FreeBSD project maintains compatibility for just about any program ever written for FreeBSD—certainly I have a couple for which I have lost the source and which are coming on 20 years old. Why the difference with PHP? And why (in this case) deprecate POSIX regular expressions anyway? They're a standard, after all.
|
Healthy eating
|
Topic: food and drink, opinion | Link here |
We eat too much fried and deep-fried food. I'm not a real health nut, but there must be some alternative accompaniment to food than the eternal chips. Today we ate stuffed sardines, which we would normally fry, and the normal accompaniment would be, well, chips. What else? Pasta seems wrong, so do beans. Rice is too dry. Some other form of potato? Gratin dauphinois maybe? A little heavy. But how about grilled potato slices, and also grilling the fish?
That's what we did. It wasn't bad. It just wasn't good. I think the best solution next time would be to make it with some kind of sauce, which would make at least rice a viable option, and maybe pasta too.
| Saturday, 8 June 2013 | Dereel | Images for 8 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another power failure
|
Topic: general | Link here |
Another short power failure this morning at 5:04. That's the 20th this year.
|
Alternative panorama processing
|
Topic: photography, technology, opinion | Link here |
More playing around with panorama processing today. The sun was shining, so I took more HDR images than normal. Normally I process the images by first converting them to tone-mapped images, and then running Hugin to stitch them together. But Hugin can handle the HDR conversion too, so today I tried that (“Exposure fused from stacks”). What a time it takes! The tone-mapping step takes quite a time too, but here I had a total of 63 images, and cpfind alone took over an hour of CPU time to find its control points. Stitching took about another 90 minutes, and in total the whole thing took 3½ hours. For that, the results would have to be worthwhile. And so (unfortunately!) they are. Here the old style and then today's method. Move the mouse over each image to change it to the other:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
These two panoramas were made from identically the same set of input images. The minor differences in the roof are the result of the difficulty of positioning the centre point exactly.
In particular, the dark areas on the left are much better. But there are some significant downsides:
I've already meantioned that it takes a very long time. If I make both a normal image and an animated panorama, it's even more extreme.
How do you add manual control points? With the old method, adjacent images are single images, so there's only one set of control points between them. With the new method there are 3 of each image, so you have 9 different sets of control points. If you need to manually fix something, like this view really needs at top left, you need to add each new control point 9 times, at least in theory.
The same considerations apply for masks. At least with the GUI there's no way to apply the same mask to more than one image, and doing it manually would be almost impossible. I suppose I should look at the project files and see if there's a way there.
In any case, an interesting improvement in total image quality. And probably yet another reason for me to buy a faster motherboard capable of taking much more memory.
| Sunday, 9 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More alternative panoramas
|
Topic: photography, technology | Link here |
On with my panorama processing experiments today. One view that is particularly challenging is the 180° view south-east from the north-east corner of the house. The left-hand end is a shade area, and to the right of centre it's open, so the difference in illumination is marked—today, surprisingly, it was only a difference of 3.4 EV from the darkest to the brightest. Still, an interesting motive for comparing the three relevant methods. In each case I started with the same 18 images, representing the following 6 views:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In fact, because of the way my camera works, I took 5 images of each view, bracketed at +3 EV, +2 EV, +1 EV, 0 EV and -1 EV, and then processed the three bracketed at +3 EV, +1 EV and -1 EV. I've deliberately shown those taken at relatively the same exposure. Stitched together with no corrections at all the result looks like the first image. My best attempt is like the second. Move the mouse over either image to see the difference:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
To do better I must, at the very least, adjust the exposure. There are three ways of doing this:
First create tone-mapped images of each view, using align_image_stack and enfuse. This is the way I currently do it. It means I only have to feed the 6 resultant images to Hugin.
Align all 18 original images and then build the panorama using the “Exposure fused from stacks” stitching, which creates intermediate fused images, effectively what I did in variant 1, and then stitches them together.
Align all 18 original images and then build the panorama using the “Exposure fused from any arrangement” stitching, which creates intermediate, possibly partial images with corresponding exposure, and then fuses them.
The results? All different. Once again I'm not sure which is best, but I don't think it's variant 1. Here they are in sequence. Move the mouse over each image to change it to the next:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On the whole I think the third looks best, notably the sky on the right.
To take the originals, I had the issue of the sun in the image. That's very clear in the uncorrected image, where lens flare shows up at the bottom left of the image. I solved that by taking images with my hand blotting out the sun, which then needed masking out again:
|
||||||||||||||||||||||
That's fine for a single image, but here I needed the same mask on each of three images. How do you do that? The Hugin GUI doesn't provide for it. But looking into the project file (.pto) for version 1, I found:
I can't find a description of the file format (though I'm sure I've seen it before), but it's relatively clear that the second parameter on that line is the image number. So it's a simple matter to change the image numbers and repeat the process:
=== grog@eureka (/dev/pts/6) ~/Photos/Hugin-build-eureka 22 -> diff -wu 00-17.pto 00.pto
And how about that, it worked! I need to learn more about the project files.
|
Turnbull speaks
|
Topic: technology, general, opinion | Link here |
Reply to the mail message I sent to Malcolm Turnbull last month:
Greg, we have every intention on delivering on our plan. Thanks for your tip on the fasterbroadband survey site, we will look into that.
Nothing world-shattering, and I remain to be convinced, but since at the time I voiced an opinion that he might not reply, it's only fair to note that I was wrong.
|
Pizza dough timing revisited
|
Topic: food and drink, opinion | Link here |
Pizza for dinner tonight, which of course involves making dough. How long does it take? In the past I've started 2½ hours before planned eating, but I've come to the conclusion that that's not enough. You really need:
15 minutes to put the ingredients together and knead them.
At least an hour, maybe 1¼ hours to let it rise in the oven at 40°.
10 minutes to mix it up again and then roll it out to shape on the plates.
30 minutes to let it rise again.
5 minutes pre-baking with paper under the plates to get it to the point where it won't stick to the plates.
15 minutes put the topping on.
15 minutes to bake them.
That adds up to 2¾ hours. Not much more, but enough. You can leave a lot of time between steps 5 and 6, so I think I'll aim for 3 hours next time.
| Monday, 10 June 2013 | Dereel | Images for 10 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More Hugin project file frobbing
|
Topic: photography, technology | Link here |
Preparing comparison images of panoramas isn't easy. Each time the crop is subtly different, so they don't register correctly when combined on a web page: the images tend to jump when the mouse goes over them. What I need is a way to crop different images identically. And that information, too, should be in the Hugin project file. But where? I still can't find the documentation. Still, inspection can be useful. Tried comparing the project files of the “normal” panorama and the HDR version of yesterday's garden to the south-east investigation. They differ in many ways, of course, notably in the number of images, but right at the start I got:
No idea what the individual parameters mean, but it's reasonable to assume that the w and and h parameters describe the size of the crop. So I replaced this one line in the project file for the “normal” panorama and tried again. How about that: the crop was right. And the intensity was all wrong. Useless. Tried cropping an image differently and saving the new project file to see what changed:
So w and h are something else. What changed was S. Irritatingly, so did the masks, though I can't work out what the difference is. Ignored that and stitched a new panorama—this time with perfect registration. But I really must find that documentation of the project file.
|
Bach and pianos
|
Topic: music, opinion | Link here |
One of the most irritating things I know is the way many broadcasters, ABC very much included, play Johann Sebastian Bach's harpsichord music on pianos. It sounds terrible! But they argue that Bach would have used pianos if he had had them.
It seems that that's not the case. In a memorable visit to Sanssouci on 7 May 1747. Alter Fritz demonstrated several of his fortepianos to Bach. When Bach returned to Leipzig and wrote his Musical Offering, it seems that he might have written some parts for fortepiano, others for harpsichord. At least that's the way they're performed in the Edition Bachakademie recording that I listened to today. The initial ricercar was played on a fortepiano, and other parts on a harpsichord.
Is this historically accurate? You'd think there would be a lot of information available, but so far I haven't found much. The booklet that came with the CDs doesn't mention it, and the score doesn't mention any instrument at all for the ricercar. There's a newspaper article on the subject, but it even has difficulty differentiating between fortepiano and pianoforte. But it does suggest that Bach didn't like the fortepiano, though it's not clear whether that would have been the case had the design been a little more mature. But it's certainly an interesting point in the discussion about the harpsichord vs. piano arguments.
The other thing that surprises me is the fascination that many people—all, it seems, American—have for the Musical Offering. Chris Bahlo sent me a link to an article likening the first Canon à 2 to a Möbius strip. I don't know if that's an appropriate way to look at it, but it's amusing to note that the score has an inverted key signature at the end, though things don't quite seem to work when you invert it. But I don't see the connection with the Möbius strip.
|
What good is this diary?
|
Topic: history, opinion, general | Link here |
Question today from Mohamed Ifadir: why do I keep this diary? Mohamed is good for this kind of question, simple to ask and almost impossible to answer. But it got me thinking.
I started keeping a daily diary over 50 years ago, and in those days I kept it up until the end of September 1970. Looking back—which I find fascinating—I'm glad I kept it, and also sorry I stopped: I didn't start again on a regular basis for nearly 30 years.
So, what good is it? I consider myself to have a very good memory, and I can often surprise people by putting a date on things that happened 45 years ago, and then being able to prove the claim with documentary evidence. But the diary also shows the things I forgot, particularly attitudes. My behaviour and attitudes have changed considerably since I retired, and sometimes I think it's because I'm getting old. But no, look at that, I behaved just the same way 45 years ago, confirming that work just gets in the way of Having A Life. Strangely, even the style of the diary hasn't changed nearly as much as I would have expected.
So, what do I get out of keeping a diary? It's great for keeping notes, of course, and it helps me think more carefully about things before I write them down, but it also helps me maintain perspective. I wonder if that answers Mohamed's question.
|
No washing machine repair
|
Topic: general | Link here |
Sλmsung had promised to repair our washing machine within 3 working days, and today Sam called up to say that he wanted to come and do it—at 19:00 on a public holiday (Queen of England's birthday as celebrated here). Decided we weren't in that much of a hurry and asked him to defer.
| Tuesday, 11 June 2013 | Dereel | Images for 11 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Eastern Rosellas
|
Topic: animals | Link here |
We have a lot of varied bird life here, including lots of crimson rosellas:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Today Yvonne found three eastern rosellas in front of her office window:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In fact, it seems that we do have them from time to time, as my photo record reveals, in one case together with a crimson rosella and some galahs:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
These are the best images. For some reason they haven't come as close as the crimson rosellas.
|
Bach and pianos revisited
|
Topic: music, opinion | Link here |
Surprisingly, my ruminations on playing Bach on pianos roused some interest. Callum Gibson thought I shouldn't blame ABC for broadcasting Bach on pianos: I should blame the people who do the recordings.
Well, that too. But ABC has a choice, and in my view they make a suboptimal choice. Callum wanted to know how many recordings were made with pianos and how many with harpsichords. I have access to the Naxos Music Library via the State Library of Victoria, so off to have a look. Chose the Goldberg Variations as a better example.
What a catastrophe! The library has many recordings, of course—that's why I chose it. There even appears to be one in the five most recent additions, though looking at the others I wonder if it hasn't been adapting to my searches:
|
|
||||||||
But the search functions are an incredible pain. Searching for “Goldberg Variations” finds a total of 256 hits, shown 20 per page. No detail information on that page, of course:
|
||||||||||||||||||||||
To find any information at all, you need to follow the links. Here's the information on the newest one, by Watanabe:
|
|
||||||||
There's no information about the instrument. Many of them have a booklet link, but this one doesn't. The cover image doesn't say anything. There's also a link to the back cover, in PDF format. In this case there's an additional hindrance:
|
||||||||||||||||||||||
That's Japanese and German. The German doesn't say what kind of instrument they are, just that they were both built by Martin Skowroneck in the German and French style of the 18th century. But they built fortepianos in the 18th century too. Only on the side does it say “Yoshio Watanabe (Cembalo)”.
Finding out one of these is enough pain. I went through the first 20 hits and checked the easier ones. The result: 2 recordings on piano, 10 on harpsichord, 2 for string trio (it seems that Sitkovetsky wrote an arrangement) and what appears to be one for harp. So it's clear that the piano renditions are the exception, not the rule.
| Wednesday, 12 June 2013 | Dereel | Images for 12 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
eBay strangenesses
|
Topic: general, opinion | Link here |
I've been using Oral-B interdental brushes for some years now. They make three kinds of brushes: too thin, too thick, and just right. For some reason I haven't been able to find the normal size for months now, but finally I found some in eBay, in Thailand, so I ordered some. They wanted $5 postage, which seemed reasonable for two packs, but I wanted to order 8, and they wanted another $7.50 for that—an increase in weight of 20 g. Finally we agreed on $3 extra.
They came today, posted
in JesseltonKota Kinabalu,
with a total postage of MYR 30 (about $10). Why so much? The packaging:
|
|||||||||||||||||||||||
Those are the brushes in the foreground. No idea why they were in such a big plastic container. eBay sellers continue to amaze me.
|
Bad weather, fallen branches
|
Topic: general, gardening | Link here |
Lots of rain today, so at least I had an excuse not to do anything. Watching TV, heard a loud noise just outside the room. Another branch fallen from the Cypress trees outside, barely missing the house:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Another task for CJ.
| Thursday, 13 June 2013 | Dereel | Images for 13 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Hibiscus in winter
|
Topic: gardening | Link here |
How hardy are Hibiscus rosa-sinensis? I know them from the tropics, and the Wikipedia page (currently) states that they can't survive temperatures under 10°. But my uncle Max in Melbourne has one growing near his front door, and I have grown a cutting from it that is currently flowering for the second time this month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The one on the right is as good as dead, of course, but I first took a photo of it over a week ago. And some people say that they only flower for a day.
|
Subdivision again?
|
Topic: general | Link here |
I've now heard back from Jarrod Hodgson with details of who claimed that the property round the corner could be subdivided: Julie Lee, whom I know from Facebook. Round to talk to her about the details, but unfortunately it was no longer clear how it could be done. She's going to follow up and get back to me.
|
Gearing up to the election
|
Topic: opinion, general | Link here |
There will be a Federal election on 14 September, the first time we've known so long in advance, and gradually the politicians are gearing up, making speeches and sending lots of paper to voters. But the only flyers we have received are from candidates in the electorate of Ballarat, and we're in the Corangamite electorate. So why the flyers? Have they changed the electorate boundaries without telling us?
To be sure, I went to the Australian Electoral Commission site, which tells me that Dereel is in the electorate of Ballarat. And also in the electorate of Corangamite. How can that be? Clearly a reason to contact them. How do you do that? Only by phone! Even if you're deaf or have a speech defect, it's the same telephone number! There's even a statement:
Internet relay users connect to the NRS then ask for 13 23 26.
I wonder if they mean IRC. There's more on NRS at that link, of course, but the instructions are a 4 minute video, which is impractical with my network connection. So, basically, if I were deaf I couldn't contact them. What's wrong with an email address? Maybe there is one, but I don't see it on that page.
So off to look at other information on that site. With some difficulty found official maps of the Ballarat and Corangamite electorates. They're of appalling quality, but it seems clear that Dereel is in the Corangamite electorate. At least I can understand the confusion of the candidates for Ballarat.
What's wrong with this picture? The AEC is an important government agency, and an important part of their mission is to inform voters and candidates. Clearly they've failed miserably:
Their search engine is Just Plain Broken. It seems to search only by post code, not by locality. Since post code 3352 is partially in the Ballarat electorate and partially in the Corangamite electorate, it claims—incorrectly—that Dereel is in both electorates.
The only way to contact them is by phone, even if you're deaf or have a speech defect. I'm sure that it's marginally possible with the NRS, but why are they not contactable by email? I can't be bothered calling people on the phone, because the people I speak to are almost never able to help, and it's difficult to escalate a phone call; escalating an email message is trivial. Phone calls also tend to last forever.
The official maps are terrible! The small versions (A4) leave out even important main roads (Dereel is shown in the middle of nowhere). The full-size maps are enormous in size (Ballarat is 26 MB, Corangamite is 18 MB), and they bring acroread to its knees. Just the initial display took 17s CPU time on my machine, and scrolling to the Dereel area took another 40s. Expanding to 800% to see the area better took another 2 minutes CPU time, leaving me at top left. Moving it to the Dereel area took a total of 12 minutes CPU time. And what did I get? A really low-resolution map. There's nothing in the map to justify the enormous size; a little bit of optimization could probably reduce its size by 90% to 95%.
Here a comparison of the section for Dereel from the Ballarat map (the big one, not the reduced one) and the corresponding Google Maps map:
|
||||||||||||||||||||||
|
||||||||||||||||||||||
The red line at the top of the AEC map is the southern boundary of the Ballarat electorate. It's difficult to get an exact crop, because the AEC map is so bad, but it does seem that the aspect ratio is wrong. Certainly here most of the roads are not named, and it shows “roads” that just plain don't exist, along with leaving out roads that do exist. And that's the official map! I'm not a great fan of Google Maps' accuracy, but the AEC map appears to be the work of a rank amateur.
| Friday, 14 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
AEC followup
|
Topic: opinion, general | Link here |
After the previous article on the breakage of the AEC web site it was only appropriate to send them a message reporting the information and asking for correction. I also referred to the article and asked to check for any factual errors. The response, written in Microsoft-style reply-before-letter form, shows a typical problem with this approach. Here it is in a more sanitized version:
And that's all. No mention of the issues I raised. Sometimes I wonder if people read these messages. Certainly there's plenty she can do to be of “further” assistance, like addressing the issues. Replied accordingly. I doubt anything will happen. Why do people have such extreme trouble communicating?
|
Internode: our fault after all
|
Topic: technology, opinion | Link here |
My network problems are no longer quite as bad as they were two months ago. The connection quality is still so bad that VoIP is seldom practical, but I no longer have the extreme dropout rate that I had at the time. That's no thanks to Internode, who ignored my traces and just gave up. I asked them to escalate the matter, and they said yes, they'd do so. Silence.
And then I got a message:
That looks like exactly the kind of problem I suspected at the time—in the Internode network, not the Optus network. And they've “recently discovered” it. If they had addressed the information I supplied them, they could have fixed it nearly two months ago. Until a few months back, it was clear to me that I would stay with Internode when the radiation tower is up and running. Now I'm not so sure. The only question is whether any other ISP is better, not whether Internode is good.
|
Revisiting AEC map slowness
|
Topic: technology, opinion | Link here |
One of my complaints about the AEC web site was that the maps they provided were enormously bloated, and they took up to 15 minutes CPU time to render on my machine. We discussed the matter on IRC and discovered a number of interesting things. Firstly, one of the issues was the way I zoomed: increase the map size to 800% and then pan with the slide bars. I didn't realize that most maps have labels in text, so I could have searched with normal text search, which speeded things up considerably—only a couple of minutes. But then somebody tried running with Microsoft, and the map rendered orders of magnitude faster than under FreeBSD—so fast that I probably wouldn't have commented.
So why is FreeBSD so slow? There are a number of possible reasons: I'm using acroread version 8 on eureka, while dxo (the Microsoft box) is running the latest and greatest. Clearly Microsoft currently wins the race to keep computers up to date. On the other hand, xpdf does not do noticeably better. epdfview does, but it's still orders of magnitude slower than anything on Microsoft.
Another reason could potentially be the difference between X and other display environments. What speaks against that is that most of the CPU time seems to be spent when not rendering. Still, a very disappointing comparison. Once upon a time we ran rings round Microsoft. In this case at any rate we're way behind. Yet another indication that FreeBSD is not what it once was.
|
Timekeeping accuracy
|
Topic: general | Link here |
It's been a month since I got my new watch. At the time I set it to the exact time and left it to see how accurate it was. It seemed to gain a lot, but today I checked: 6 seconds in a month. I can live with that. Set it to 4 seconds slow; I'll see if it continues that way.
Interestingly, the old watch, which also gained significantly, now seems to be keeping exact time. Is it something about my wrist?
| Saturday, 15 June 2013 | Dereel | Images for 15 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden flowers in early winter
|
Topic: gardening, photography | Link here |
Middle of the month again, and also a Saturday. In principle time for both the garden flower photos and the “house photos”. In practice, it doesn't seem to make much sense to take a complete sequence every week at times when there's no change, so this week I took very few, not even the almost-obligatory 360° verandah panorama. But of course there were some flowers—many fewer than this time last year.
Why? The hot summer and insufficient irrigation were clearly two reasons. I haven't been paying much attention to the garden lately, but that shouldn't be the issue. In any case, there are positive signs: in principle, the Alyogyne huegelii is looking much happier than it has been:
|
|||||||||||||||||||||||
Some of the Eucalypts are flowering, and the Grevilleas are coming on:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Gentiana scabra, which is supposed to be winter tender, has finally recovered from the summer, and has produced a few shy flowers:
|
|||||||||||||||||||||||
And the Schlumbergera truncata in the kitchen is also in full bloom:
|
|||||||||||||||||||||||
Roll on the spring. The Acacias are preparing, and we've had various individual Narcissus in flower for a while now:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Acacia baileyana in the first image has so many buds that Yvonne thought it was already flowering. But it won't be long now.
By comparison to last year, though, there's a lot missing: this time last year the sweet peas and Hardenbergia violacea were flowering, as were the Arum lilies and the Senna aciphylla. Here last year's photos:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why the difference? The hot summer can only be part of it.
|
Taming DxO
|
Topic: photography, technology, opinion | Link here |
I've been processing my photos with DxO Optics “Pro” for over a year now, but I still haven't really understood all the details. I frequently use the “HDR Artistic” profile, which gives me the kind of in-your-face colours and shadow accentuation that I like, but I've only gradually come to realize that it comes at the expense of considerable burnout in the highlights. Today I had to adjust some of my flower images by up to 2 EV to get some detail back in the highlights. More to be learnt.
| Sunday, 16 June 2013 | Dereel | Images for 16 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another mystery plant
|
Topic: gardening | Link here |
The garden's full of weeds, but some of the things I find don't seem to be weeds. What's this?
|
|||||||||||||||||||||||
I fed it into Google Images, which matched it with a couple of unrelated plants and a number of jaguars. How do I find out what it is?
|
PayPal: Don't follow this link
|
Topic: technology, opinion | Link here |
Mail from PayPal today with some policy updates and the good advice not to follow any links in their messages, but to type in the URL! Yes, it's possible to have a disguised link, but the idea of cutting and pasting doesn't seem to have occurred to them. And then they provide links anyway—which do exactly the kind of obfuscation they're warning about (real URL at bottom left):
|
|
||||||||
What happens if you follow that link? An electronic rap on the knuckles?
| Monday, 17 June 2013 | Dereel | Images for 17 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cold and miserable
|
Topic: general | Link here |
It's cold and miserable at the moment, and in addition Yvonne had a cold and spent all day in bed. I don't usually get colds, but somehow the general misery of the situation washed off on me too.
|
Repairing the washing machine
|
Topic: general | Link here |
Sλmsung seem to have a desire to perform their recall action outside normal working hours. Last week they wanted to send somebody by after dinner on the Queen of England's birthday, and today Peter wanted to come at 8:00, a time when I normally am just thinking of getting up. Managed to put him off until 9:20. He arrived, took the cover off some of the electrical distribution, and pointed out the problem:
|
|||||||||||||||||||||||
It seems that the area can get very moist, and this connector can then start to burn, setting fire to the machine. Looking at it, it's amazing how many things are wrong: apart from the fact that it's not moisture protected, it's apparently a connector to the motor, as evidenced by the thick wires going out. But the input wires are an order of magnitude thinner. That looks very dubious to me. In addition, it suggests that the connectors are made of a potentially flammable material—most of the surroundings are metal. And I wonder if the moisture in the affected machines is due to condensation or leakage. As the “after” images shows, this thing is right next to the cold water input:
|
|||||||||||||||||||||||
You'd think they would have sorted that out better.
|
Reprocessing old panoramas
|
Topic: photography, technology | Link here |
It's been nearly two years since I started my current panoramic views of the garden, and in that time lots of things have changed. In particular, at the time I was using out-of-camera JPEGs, and now I process my images with DxO Optics “Pro”. The difference in appearance may be at least partially due to that. Here the verandah two years ago and now:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Spent quite some time reprocessing the images with DxO Optics “Pro”, but the difference wasn't that pronounced. Here a comparison, first the original from two years ago and then today's processing. Move the mouse over each image to change it to two others:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The difference isn't nearly as pronounced as I expected. Quite possibly part of the issue is that at the time I exposed all the images in the same way, like the instructions say, and I thus lost a couple of EVs of dynamic range.
| Tuesday, 18 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye Friends
|
Topic: technology, opinion | Link here |
Phone call from somebody at a company Wordsworth. He wanted access to the domain fbbg.org.au. I told him there was a web site, but no, it seems Wordsworth (or whatever) is the company doing the transition to a “professional” web site, and what he really wanted was information on how to update the DNS information. He had the registry key, but didn't know what to do with it.
Asked him to send me a mail message, which came from Adel, with whom I had spoken earlier this year, with an email domain address from a different domain. Sent her the information, and the DNS records were quickly updated. Off to look at the site. Nothing. Literally nothing: the web server delivered an empty document. Sent her a message to that effect; it seems that they still hadn't loaded the content. Later I found this page, which stayed there all day long.
Why? It looks as if they first bent the DNS and then started to install the web site. Why couldn't they have done it before switching the DNS? And why did it take so long?
|
Death to HTML!
|
Topic: technology, opinion | Link here |
While playing around with my photos a few days ago, I came across a strange problem: in my diary for 16 April 2011, my normal photo resizing stopped working correctly. Spent some considerable time investigating it, finally reducing it to a simple example. Normally I have five potential display sizes for an image: hide (size 0), “thumbnail” (67,500 pixels, size 1), “small” (270,000 pixels, size 2), and two “big”, both the native size of the image. Size 3 scales this image to the width of the window, while size 4 shows it in full resolution.
In this case, though, size 3 was smaller than size 2. Everything else worked. My image display functions are large and have evolved along with my understanding of PHP. All in all they're 2,500 lines long, and I dreaded debugging them. Was this maybe a result of the “lazy loading” of images that I implemented a year ago? In the meantime I've found some other reasons to want to disable it, so I added a request parameter lazy which defaults to 1, but when set to 0 disables the lazy load. That's particularly useful for printing things out, since printing doesn't enable the loading. And in this case it gave me the opportunity to confirm that it wasn't the problem with my web page.
What was it? Hacked around in the image and finally discovered that a <p/> tag was causing it. If I replaced it with a <p></p> pair, the problem didn't occur.
Why? It seems that <p/> is not allowed in all dialects of HTML, and browsers all interpret it in such a way that the implied CSS remains in effect after the tag. But it is in XHTML, which is what I'm using here.
That could change, but at the time of writing it's correct.
And no test tools complain about the usage. However, it seems that I'm sending out headers from the web server indicating that this is text/html. Not deliberately.
And then I hear a claim that the P tag is deprecated in HTML5. According to the W3 schools reference that's not the case, but who knows what might happen? In any case, these things differ from one dialect to another. Why didn't they introduce a completely different markup instead of trying to make a silk purse out of a sow's ear?
|
SCO: The pain that never ends
|
Topic: technology, opinion | Link here |
In other news, Jürgen Lock pointed me at this report. After only 10 years, a complete lack of evidence, and a bankruptcy, SCO (now spelt “XINUOS”) is still not giving up with their law suit against IBM. Why?
Went back looking and discovered that it was ten years ago today that SCO announced that they had terminated IBM's UNIX license. It's also the tenth anniversary of the publication of a Byte interview with Chris Sontag, unfortunately also no longer accessible. Went looking and found that most of the documents to which I had referred in my documentation of the case have since ceased to exist. Spent some time tidying up my own documents, which are sorely in need of it, notably my rebuttal of the Sontag interview. What a mess it all is, and how few of the web links there are still active!
|
Minimal garden work
|
Topic: gardening | Link here |
I've done almost nothing in the garden for months, and it shows. But gradually I have to do some things. For reasons that no longer make sense, I moved the Meyer lemon to the north verandah two weeks ago. That's not a bad place for it in the summer, but it's no good at all in the winter, when the greenhouse would be a much better place, so moved it there.
Also dragged out the Cyclamen plants that have been in the shade area since the summer. There are seven pots, and I had to extricate them from the grass that had grown around them. It looks as if three of them haven't survived. I should pay more attention to these things.
| Wednesday, 19 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cleaning up the fallen branch
|
Topic: gardening, general | Link here |
CJ along today to remove the branch that fell last week, along with another that had fallen in a paddock a little later, and also to finally to remove the locust that I had been leaving until the small birch next to it was big enough. But somehow the birch was losing the battle; it's still only about 1.8 m high, while the locust had reached 3 m, and the trunk was 25 cm in diameter. Clearly not a thing to allow to establish itself.
While I was at it, I asked him to remove another of the trees that had sprung up in what should be the birch grove, which I took to be one of the Pittosporums that once created the Cathedral. But then I saw buds:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's clearly not a Pittosporum. What is it? I have a feeling that I should know, but it evades me at the moment. At least it means it will survive until it has flowered.
| Thursday, 20 June 2013 | Dereel | Images for 20 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Coldest night on record
|
Topic: general | Link here |
By Australian standards last night was very cold, -1.2°, the coldest since I've started keeping records. By the standards of other countries, that's not particularly cold, but with the standard of insulation we have here, it was miserable. Roll on subdivision or some other way to get a new house.
|
Why we can't beat spam
|
Topic: technology, opinion | Link here |
Spam has been one of the most irritating aspects of online life since I first started using email on the Internet. There are all sorts of ways to try to limit it, but as time goes on, it's clear that we can't win. The basic problem is that too many people consider it to be just another legitimate form of advertising, and too many users just accept it. Yes, penis enlargements and Viagra are frowned upon, but that's because of the subject, not because of spam.
But who uses email any more anyway? Facebook is the way of the future (Oh brave new world, that has such people in't!). Up to recently it restricted its advertisements to a column on the right, but lately it's interspersing them in the normal messages:
|
|
||||||||
What's the difference between that and spam? With email, you can filter the spam. With Facebook, you can't. And it's not just the smaller people who do it. Yesterday I saw spam from the Commonwealth Bank of Australia, who should be above such practices. Worse still, people like this stuff. 2,146 likes so far for the North Queensland spam, for example. We've lost, at least on Facebook. One more reason not to use Facebook.
And not all advertisements are that harmless. On the right I find one encouraging me to have sex with married women:
|
|
||||||||
Can Viagra and penis enlargements be far behind?
|
Cooking hamburgers
|
Topic: food and drink, opinion | Link here |
Hamburgers for dinner tonight. I normally fry them or barbecue them, but looking at the recipe I find I had mentioned grilling them (1½ minutes per side). I couldn't recall ever having done that, but today I tried. It took a total of 12 minutes to get them cooked, by which time they had shrunk considerably. Not to be repeated. I suspect I meant to barbecue them for 1½ minutes. Recipe updated.
| Friday, 21 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye Facebook, hello Google+?
|
Topic: technology, general, opinion | Link here |
After my rant about Facebook yesterday, Peter Jeremy came up with a “solution”: use Google+ instead. It doesn't have inline spam (yet). I had actually had a better solution: ignore all these social media sites. But for the fun of it, I signed up with Google+.
What is it? Yes, of course I didn't read the description, but what I'm presented with looks nothing like what I expected. I'll revisit it some time when I'm bored, but for the moment the difference is clear: I'm on Facebook because it's the only way to communicate with lots of people I know. Few people have invited me to join them on Google+, and most of those are also on Facebook. So for the moment I see no advantage.
|
Friends web site still down
|
Topic: technology, gardening, opinion | Link here |
I noted a couple of days ago that the new web site for the Friends of the Ballarat Botanical Gardens is still down. It seems that they jumped the gun and changed the DNS information before installing the web site. Since the site has been down for days, it's reasonable to assume that they've run into trouble. So I put up the old site at fbbg.lemis.com and offered to Adel to redirect to it while she sorts out her problems.
But today I got a message which blew my mind. This isn't a bug, it's standard procedure, it seems, and they had warned the Friends that the site would be out of action for a couple of days: “there are some elements of the site that can't be implemented and tested until after propagation”.
I'm baffled. Who tests live web sites? What does she even mean by “propagation”? And why does anything like this take 4 days? She blames the registrar, but that's nonsense: the DNS changeover happened within minutes, and clearly “propagation” didn't start until it was complete. Clearly my lack of understanding is because I'm not a professional. I'll watch further development with interest.
|
AEC replies again
|
Topic: general, opinion | Link here |
Another reply from the AEC today, after I had given up hope, and almost exactly a week after the last reply. The content was surprising: she claimed that yes, indeed, the “suburb” of Dereel was split between the Ballarat and Corangamite electorates, since 2010. She pointed me to another map which she claimed was authoritative—but it doesn't even cover the electorates we're talking about. It's an electorate map for the State of Victoria, and it shows the boundaries between Polwarth and Ballarat East. As far as I can see they're geographically the same as the Federal Corangamite and Ballarat electorate in the area I'm interested in, running between Enfield and Dereel. But how can that be authoritative? It doesn't even mention the electorates in question.
Then she suggests that I do an electorate search—exactly the point I was complaining about. Finally she suggests that I use the low-resolution divisional map, the one that shows no roads around Dereel at all with the possible exception of the Dereel-Mount Mercer Road, which is shown at 45° to the correct direction. How can I get any information from that? That's another of the things I was complaining about.
I have now had two different people reply to this message. I'm left wondering if either read it. Certainly it's an indication of extreme incompetence somewhere.
|
Server down after nearly 5 years?
|
Topic: technology | Link here |
My external server is hosted with RootBSD, and I've been very happy with them. As of today it has an uptime of 1,726 days—4 years, 8 months and 24 days. And now they're going to move the data centre! To quote this page:
And his server had only been up for 2½ years.
|
Counting beans
|
Topic: food and drink, opinion | Link here |
Over the last few months we have somehow ended up with many more dried legumes than intended. I must have the best part of 10 kg of dried beans, peas, lentils and other dal. Today, in preparation for a feast tomorrow, I cooked another batch of mixed dal. It requires mung dal, tur dal and masur dal. And I discovered that I didn't have enough of either of the latter two. Clearly I had enough other sorts, so this batch was made with five different kinds of dal, but how could I run out without noticing?
| Saturday, 22 June 2013 | Dereel | Images for 22 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
A different power failure
|
Topic: general | Link here |
This morning at 9:34 we had Yet Another power failure. Nothing unusual in that: there's about one a week, and I even have a key combination to make it easier enter the URL into my diary. Called Powercor, spoke to Nadia, and she gave me the usual information about sending people out here as soon as possible.
Then I noticed something strange: the air conditioners were still trying to run. After a bit of investigation discovered that we still did have power, just not enough. The voltage was round 100 V, and it fluctuated wildly. In fact, the clock on the oven didn't fail at all. Still, nothing we could do about it, and after half an hour of waiting around for the power to finally come good, getting more and more frustrated, decided to go somewhere and wait for it to get fixed.
|
Going shopping
|
Topic: general | Link here |
But where? My travelling days are over, and I've seen much of what there is to see in the area. In the end Yvonne suggested we go into Ballarat and look at some fridges: for some reason she wants to buy a new fridge, and though it's probably animal-related, it would have the advantage of saving power when we retire the old fridges. It sounded better than any alternative I could think of, so set off and did a bit of shopping, finally ending up at a community market in Sebastopol, where we bumped into Julie Donaghy and a friend, Sharon I think, who once lived in Paynes Bridge Road before the bushfire. Spent some time talking to them: Julie is arranging a petition for a mobile phone tower, which I thought was no longer necessary after the assurances that Denis Napthine gave us after the bushfire. Still, she wants to come along and discuss the wording, so I'll soon find out more.
Then (finally) off to look at fridges. Why are they so expensive? The very cheapest cost over $500, and you can spend up to $3,000 on them. Came to the conclusion that we no longer wanted a separate freezer and fridge (we'll keep the ones we have), and gradually came to the conclusion that a side-by-side arrangement would be best. There are almost identical fridges from “Kelvinator” and “Westinghouse”, clearly made by the same company, with only minor differences in the controls. At Good Guys they're marked at $1,199 and $1,198 respectively—but after the discount, the prices change differently to $1,158 and $1,169. Why do they do that? In any case, the “Westinghouse” looks better equipped. In particular it has a more baskets for the deep freeze section (why are these going away?), which alone would justify the price difference. Now to research these things on the web.
|
Power failure, continued
|
Topic: general | Link here |
Back home round 13:45—the power was still jumping around! Called up Powercor and got Nadia again, who told me that the crew was on its way to our house, since we were the only people who had reported problems. That seemed plausible this time, since the symptoms were very different from normal failures. Turned off a few circuit breakers and discovered that yes, indeed, the voltage returned to normal. But what could use enough power to drop the voltage like that without tripping a circuit breaker? Heard the water pump trying to pump and sounding very sick about it. Turned it off and noted that it was very hot—had it failed and caused the problem? Or had the problem caused it to fail? Had it even failed? No way of telling until the Powercor people got here. And there was every chance that by the time we restored power, it would be too late to buy another pump. What a pain!
The Powercor people came only a couple of minutes later in two different vehicles, one with a cherry picker. The first bloke wasn't very interested in my circuits, though he confirmed my readings: just turning on one circuit, with nothing connected, caused the voltage measured at the switchboard to drop from 230 V to 209 V.
Instead he took a look at the feeder cable, on which he thought he saw some moss growing:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
They spent a lot of time investigating the connection to the transformer. There, it seems, the voltage was 234 V, so they re-tapped it (and presumably cut power to all the neighbours) for 240 V. Back to the switchboard: still 230 V! How can that be? But it was fairly clear that something was wrong with the feeder, so they replaced it:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
After that, tested two things at once by turning on the pump. It worked! So the whole thing was really the feeder cable, and the problem with the pump was just that the voltage was too low. All in all, for the Powercor people it was just another job, but it really stressed me.
|
Exposure against the sky
|
Topic: photography, opinion | Link here |
The Powercor incident gave me an opportunity for lots of photos, and also for proving how versatile the 18-180mm f/3.5-6.3 lens is. With any other I would have been continually changing lenses.
But how do you expose the photos of the cherry picker? Clearly the background would confuse the exposure meter in the camera unless I used spot measurement, and even then I wasn't sure. In the end I tried several different methods, with the result that the exposures, while acceptable, weren't particularly good:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
More experimentation needed, hopefully on other motives.
| Sunday, 23 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Where is localhost?
|
Topic: technology | Link here |
Sent off a relatively routine mail message today and got an unexpected response:
Huh? What does that mean? In particular, it's a resolution failure for an AAAA record, which relates to IPv6. And I don't use IPv6. More importantly, though, why now? Yes, I rebooted the machine yesterday after the power failure, but I didn't change anything. And both my postfix and DNS configuration haven't changed for a long time. So what happened? Some strangeness as the result of the reboot? Had any mail gone out at all since the reboot? It proved that it hadn't, but why should this happen anyway? And why IPv6? The configuration hadn't changed. And the system knows localhost, including the AAAA record:
=== grog@eureka (/dev/pts/7) ~/src 98 -> host localhost
But there was one difference: I need to run a tunnel to get mail out to the outside world, because Internode blocks outgoing SMTP. And I hadn't restarted it after the reboot. That's easy enough. I have a little script, the heart of which is:
When I tried starting it, it failed, spewing out log messages every 5 seconds. OK, take the heart out of it and run it from a shell. It worked. Does the script have trouble with the ssh keys? Pushed that on the tuit queue.
So: did that fix it? In the past it just hung around waiting for me to restart the tunnel. So it wasn't surprising that it didn't make any difference here. telnet localhost 2026 connected me to the remote MTA as expected, so there was no problem at the other end.
I really didn't want to look at the problem, so decided to try it on dereel, the remainder of my old installation. It's now a VM, and it certainly hasn't changed in any way in at least 6 months. What happened?
I suppose that's useful up to a point: at least it shows that whatever it was, it wasn't due to the reboot, and it tended to confirm a suspicion I had that the reference to the AAAA record was because a lookup for an A record had failed. But what was it? Peter Jeremy suggested that it wasn't looking for the relative name localhost but for the absolute name localhost.. And sure enough, that didn't resolve. But why should it? As the name suggests, localhost is local. Why supply an absolute address?
But where did the search for localhost come from in the first place? It's in /usr/local/etc/postfix/transport:
This tells postfix to send mail via the tunnel (port 2026 on localhost). And it seems that it was interpreting this to mean localhost.:2026. That can be changed:
And that worked.
But the problem isn't done. I've had the first form in transport for nearly 6 years. Why did it stop working now? And why does my tunnel script fail? As far as I can see these are two completely different problems that have bitten me at once. But I couldn't be bothered looking at them today. So far my hypothesis is that Internode, who forward DNS for me, have been supplying localhost. all this time, and today they stopped. But it'll be difficult to confirm that.
There's another issue that might have prevented this problem ever happening: resolving localhost. locally. Newer versions of FreeBSD do this, but my old configuration doesn't. Also a thing to try out, but it doesn't explain the problem either.
| Monday, 24 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More cold weather
|
Topic: general | Link here |
It's been particularly cold lately—every night for the past 5 days we've had a frost. Here's what it looked like this morning:
So yet another indoor day.
|
Following up on mail problems
|
Topic: technology | Link here |
Pulled my tuit item about the mail problems today. As I suspected, there were two different problems. The more obvious one was setting up the mail tunnel. The script is called mailtunnel:
=== grog@eureka (/dev/pts/10) ~ 17 -> which mailtunnel
I modified the script to include debug output, and none came. After quite some time it occurred to me that I had a function in my .bashrc, and that's what got executed. And at some juncture while setting up teevee, I modified its .bashrc, conveniently a symlink to /eureka/home/grog/.bashrc, and managed to save an older version:
It's not clear why I commented out the invocation, but then it's not clear why I have two copies.
And the localhost issue? While writing my diary, I tried again to resolve localhost.. It worked:
=== grog@eureka (/dev/pts/10) ~ 20 -> host localhost.
That's not from me. It must have been forwarded from Internode. This suggests that my suspicions yesterday were correct, but it's difficult to go back and check.
Edwin Groothuis also pointed me to Internode's firewall settings page, which includes an option to unblock outgoing SMTP. That seems reasonable, so tried that, but after well over the one hour processing time it was still blocked. Still, no hurry.
|
Networking, Apple style
|
Topic: technology | Link here |
While following up on my network problems, decided to compare what Apple does. Yes, like other BSD-based system they also have a localhost bound to the interface lo0. This seems to be a BSDism: the Linux interface is simply called lo. But somehow since last use the machine had changed its system name, from newyork to newyorkmmmmmmmmmmmmmmmmm. No idea how that happened, but it should be trivial to change it, right? hostname is your friend.
Well, it may be your friend, but Apple doesn't take it seriously. I'm told it works until you reboot, and then you're back to the old name again. The leet way to do it is via the GUI. With a click-click here and a click-click there you get...
Wait. The first click-click gives you the “System Preferences” screen:
|
|
||||||||
And then? Where do I click-click? “Network” seems obvious, but it doesn't offer the option. By a process of elimination discovered that it was “Sharing”. That seems to say more about the mindset of the people who developed the menu system than anything else.
When I try to enter the host name, it appends this stupid .local to it, and I can't change it:
|
|
||||||||
But there's this “global dynamic hostname”. Not what I want: I want a static hostname. But it looks as if I don't get much choice, so tried that. And what do I get?
|
|
||||||||
User name? Password? Clearly this has more to do with file sharing than it has to do with host names. Why have they mixed this stuff up so horribly? So I put in a name and a password that I have already forgotten (I don't even have an idea what kind of file sharing this is, but clearly it's not NFS), and it worked. Aren't GUIs simple?
Then, of course, Callum Gibson pointed me to the left hand version:
Looking at the man page, scutil seems to be a useful tool, one I should remember next time I have Apple pain. It even has a man page which claims it's part of the BSD System Manager's Manual, though the HISTORY section explains that it first appeared in Mac OS X. Maybe “BSD System” refers to Apple's left hand.
The only trouble is, it doesn't seem to see the same thing as the “Sharing” tab:
And does the right hand know about that? Of course not. It still has newyork. What a mess!
|
AEC again
|
Topic: general, opinion | Link here |
Another reply from the AEC today, once again not addressing my issues, but claiming both that parts of Dereel are in the Ballarat electorate, and that the maps, which show that no part of Dereel is in the Ballarat electorate, are correct. How do you deal with this?
| Tuesday, 25 June 2013 | Dereel | Images for 25 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Australian Liberal Party Spam
|
Topic: technology, opinion | Link here |
In my article about Facebook spam last week I couldn't find a really good example of a reputable organization spamming Facebook users. Today I got a perfect one. The “Liberal” Party of Australia, arguably a reputable organization, considers it appropriate to spam its potential voters:
|
|
||||||||
Never mind the polemic and inaccurate message (electricity prices have gone up, but not by 94%): would you vote for spammers?
|
Fast dog photos
|
Topic: animals, photography, opinion | Link here |
Zali from Enfield came along today with Bindy, her Borzoi bitch and Zhivago's daughter, to go for a walk. It was one of the first times that Bindy has been allowed loose in the forest. Went along to get some photos, rather complicated by the fact that Bindy wouldn't keep still:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Unfortunately, Zhivago wasn't very interested.
|
Investigating mail problems
|
Topic: technology, opinion | Link here |
So despite setting the firewall rules, I still wasn't able to connect to my remote MTA. But the reason was easy enough to find:
Clearly a configuration problem, so I sent a message to Internode Support asking them to fix it. Some hours later got a call from Stuart, telling me that they're not blocking anything. I pointed to the tcpdump output, and subsequently explained to him what it meant. He pointed out that ii.net is not an Internode domain. So what is it?
=== grog@eureka (/dev/pts/12) ~ 3 -> host lns20.syd6.on.ii.net
Clearly another configuration issue. But tcpdump -n showed the address to be 150.101.199.159, and traceroute showed that to be my next hop address:
=== grog@eureka (/dev/pts/12) ~ 4 -> traceroute w3
He then did some of his own testing, sending a ping to the machine in question to see whether it was alive. More explanations from me. After a while he declared his intention to go and talk to the code monkeys (his term) and put me on hold. Came back with questions like what operating system, what email address I was trying to reach (“none”; this was a simple telnet mx1 smtp), and so on. Explained again and voiced the opinion that there was probably an issue updating the firewall on the node in question, and that after checking it out and fixing he could send me an email.
That happened. They fixed the problem and sent me email, with a list of questions to answer in case it didn't work:
Most of these questions are meaningless, and they show that I wasn't successful in explaining networking basics. I doubt that anything would help them “to better understand where the issue is”.
|
Apple: four times as good
|
Topic: technology, opinion | Link here |
More investigation of Apple's host naming today. It's clear that there's a naming issue somewhere. Investigation with scutil revealed two different names:
Apart from these two, there's the “dynamic global hostname”, which is something else again. I can only find it in the GUI setup screen, but it does get saved:
|
|
||||||||
But wait! There's more! hostname has its own view of the world:
Four different names for one computer! Who else can give you that? And isn't it nice to have an intuitive system?
Later there was a claim from somebody that Apple was moving away from Unix. It's not clear if there was any foundation, but then Jordan Hubbard, Apple's Director of Software Engineering, announced (on Facebook) that he is leaving Apple after 12 years to go to a smaller company. He's not saying where, though I suspect it might be on the Maldives, and that he's really just pissed off because the latest release of Mac OS X is not named after a cat.
| Wednesday, 26 June 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Focal length calculations
|
Topic: photography, opinion | Link here |
Recently there was a thread on the German Olympus forum: somebody had compared the field of view of the Zuiko Digital 11-22 mm and the M.Zuiko 12-50 mm lens and discovered that in full wide-angle position the 12-50 mm lens had a wider field of view than the 11-22 mm lens. Lots of theories as to why that should be. One pointed out that both lenses had to be focused on ∞. I thought that that was unlikely to make enough difference, so adapted my fov program to take into account the difference in field of view when focused closely. As expected, that wasn't the explanation:
=== grog@eureka (/dev/pts/3) ~ 78 -> fov -d 0.28 11
=== grog@eureka (/dev/pts/3) ~ 79 -> fov 12
But what is the focal length of a lens? For a thin lens it's clear. But what about a complicated modern zoom lens with inevitable aberrations and distortion, especially if it's a wide-angle lens? If a rectilinear lens has barrel distortion, the effective focal length varies depending on the angle. And after correction things look different. But that doesn't change the physical characteristics of the lens.
Hugin has one interesting way of calculating the focal length: with 360° panoramas, it can calculate the exact angle of each image. And from that, given the sensor dimensions, you can work out the focal length. I've long noticed that it reports the focal length of my 9-18 mm lens as about 8.8 mm when set to the shortest focal length. And then there's the fact that processing software, whether in-camera or external raw format converters, return less data than the original image:
=== grog@eureka (/dev/pts/3) ~/Photos/20130622/orig 84 -> identify P6221055*
So what focal length would be reported if I processed a panorama with ufraw, which doesn't cut off the edges? That's straightforward enough, so I tried it out on one of last weekend's panoramas.
Surprise, surprise. The panorama didn't close. Further inspection shows that Hugin determined the focal length of the lens to be 13.953 mm, and thus didn't expect it to close. Even after I added some manual control points between the first and the last image, it didn't want to know: clearly it had decided that there was no way that this could be a 360° panorama. This has nothing to do with the issues at hand: Hugin reads the focal length from the EXIF data. But they appear to be correct. Presumably something in the data got corrupted and confused Hugin.
That's not a big issue: you can tell Hugin what the focal length is, so I did that, and the panorama closed. Here an overview:
| Converter | Focal | Avg | Max | |||
| length | error | error | ||||
| DxO | 8.881 mm | 0.9 | 2.0 | |||
| ufraw | 8.758 | 1.0 | 2.4 |
In fact, considering that the ufraw conversion didn't compensate for distortion or aberration, that's not at all bad. But I'm not convinced that it's plausible. Unsurprisingly, the resultant panorama didn't look as good, but some areas were Just Plain Bad. No prizes for guessing which of these two images is which. Move the mouse over each image to change it to the other::
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Even more alarming is the difference in the image quality, particularly noise:
|
|
|||||||||
|
|
|||||||||
But the real issue was focal length. I was expecting the ufraw version to show a shorter focal length than the DxO Optics “Pro” version, because it had a wider field of view, and it did. But my expectations were based on flawed logic: the field of view is wider because there's more of the sensor in the image. And the focal length is independent of the size of the sensor. So this is just a coincidence.
Or is it? I clearly need to look at the EXIF data. The only way to get that field of view on the original sensor area is with a shorter focal length lens. If Hugin is told that the resolution is 3084x4096 instead of 3024x4032, but is also told that the sensor dimensions are unchanged, it will calculate a shorter focal length. In this case, the calculation is 8.881 × 4032 ÷ 4096, or 8.742 mm. Given the accuracy of the fit, this seems plausible. Clearly there's more digging to do in the EXIF data.
|
Maultaschen: too much work
|
Topic: food and drink, opinion | Link here |
Yvonne bought some ravioli cutters at ALDI recently, with the intention of making Maultaschen, a kind of German ravioli-like pasta normally eaten in soup. Today we tried them out.
It's been a while since I made fresh pasta, so long that I had forgotten that I had a recipe online. And somehow it's more work than I remember.
| Thursday, 27 June 2013 | Dereel → Ballarat → Dereel | Images for 27 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
A new prime minister
|
Topic: opinion | Link here |
Woke up this morning to hear the news at 7:00: we have a new prime minister, reported as if it were the most usual thing in the world. But why?
We're working up to a Federal Election on 14 September 2013, and the opinion polls show that the incumbent Australian Labor Party are heading for a defeat, notably because of the poor approval ratings of the Prime Minister, Julia Gillard. She's even far behind Tony Abbott, the leader of the Liberal Party of Australia.
Again, why? I don't have any particular love for her, but under her leadership Australia has handled the Global Financial Crisis better than just about any other developed nation. And as far as Tony Abbott goes, I feel a deep disgust. He seems to have no idea what he's doing, he's continually attacking others, and in general he comes across as a bumbling idiot, the worst I've seen in an environment where people are used to idiots.
But Kevin Rudd, the new Prime Minister, was seen to have a better chance for Labor to win the next election. So last night, it seems, Julia Gillard called for a vote, and lost.
But again: why is Gillard so unpopular? Why is Abbott so popular? Are opinion polls skewing the political landscape? Certainly they reflect nothing like my own opinions, nor those of most people I know.
|
Towards pension age
|
Topic: general | Link here |
In to town today to see Peter O'Connell about my investments. Not a good time to talk about investments, especially in view of the change in Prime Minister. Instead more talking about my impending pension, which is interesting. From 28 September 2013 I'll be receiving a pension from Centrelink, the intuitively obvious name for the Australian Social Security department (only bettered by their choice of web site name humanservices.gov.au), and also a substantial pension from Germany, starting on 1 December 2013. The latter will diminish the payments I get from Centrelink—or will it? They have two ways of assessing alternative income, conveniently either not on their web site, or well-hidden. One is the “Assets Test”, which diminishes payments to people with lots of money (such as my superannuation), and the other is the “Income Test”, which diminishes payments to people with other income. Only one applies, the one that diminishes more.
In my case this means that if I have no other alternate income (superannuation doesn't count), I end up with approximately ⅔ of the full rate. If I have up to $26,000 per year in other income, this doesn't change, because the assets test still lowers the pension more. Only after $26,000 does the income lower the pension more. So effectively the first $26,000 are free.
The bottom line, though: it looks as if we will end up with lots more money starting at the end of September and increasing through to June 2014, when Yvonne gets her German pension.
While there, discussed the day's political events with Peter. We're in agreement. Tony Abbott can't possibly be as stupid as he looks, or he would never have become the leader of the Liberal Party of Australia. Presumably he considers it the right way to treat the electorate.
|
Constructive newsreading
|
Topic: opinion, multimedia | Link here |
Every morning at 7:00 I listen to the news on ABC Classic FM. Their newsreaders aren't always the best. At the moment it's Wendy Glamacek (if that's the spelling; I have no way of finding out, since the newsreaders aren't featured on the web site). She has never yet made it through the news without at least one mistake, and I've taken to counting them—the record is round 10. But today she came up with an explanation for the timing of the leadership change in the ALP:
Today is the last seduled parliamentary shitting day.
That's particularly amusing in view of Tony Abbott's behaviour, but it also means that the Opposition doesn't have enough time to organize a no-confidence vote.
| Friday, 28 June 2013 | Dereel | Images for 28 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More sausage making
|
Topic: food and drink | Link here |
Time to make more Bratwurst today. Preparations are a significant issue, especially as in this case we had forgotten where we had put the bucket for the sausage machine, and ended up searching the house. Yvonne finally found it in the pots and pans cupboard opposite the dish washer, where it belongs.
Today everything went according to plan, apart from the fact that my 1 kg of pork shoulder weighed 1.67 kg, and my 500 g of belly weighed 750 g. We'll have to keep an eye on the butcher next time. The proportions were still roughly OK, so I just processed the lot. And that's about as much as my tools, notably the Kenwood mixer, could handle, though it also nearly filled the bucket of the sausage machine. I had written the recipe for a total of 3 kg meat, but clearly that's too much. More modifications.
I'm continually worried that I'll end up with pockets of salt in the sausages, but so far that has worked well. But as a result I mix for some considerable time, and the total time to prepare the mixture must have been about an hour. By contrast, as the EXIF data of the camera shows, it took only 11 minutes to fill roughly 26 sausages:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
There are a number of minor issues in making sausages this way, the most important being that you don't know how much mixture is left in the bucket. Today we ended up with the last two sausages like this:
|
|||||||||||||||||||||||
And there's still the issue of the quantity of mixture left behind in the tube connecting the bucket to the filler:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
We made some “ćevapčići” of it again, but next time I think I'll just freeze it and fill it into the next batch of sausages.
And the taste? As before, consistent. Yvonne thought that there was too little salt, but I was happy. Rather too little than too much.
|
ALDI “Healthy Fryer”
|
Topic: food and drink, opinion | Link here |
We're continually buying things from ALDI to try out, knowing that we can return them with no questions asked. As a result we've accumulated a number of interesting devices that we wouldn't have bought otherwise, so I suppose it must be in ALDI's interests. But we also return a large proportion, maybe 25%.
We eat a lot of deep-fried food, probably too much, and I've been wondering how to cut down on the quantity. This week ALDI has a “healthy fryer” on special. At first I thought it was supposed to be a friteuse, but in fact they don't say that. It's also hardly an appropriate term for a device that uses one tablespoon of oil to fry food. But they said it was good for making chips, so we bought one to try out.
It's definitely a strange device. You need to mix the oil with the food before putting it in to the fryer. It then moves the food around in circles over a teflon-coated surface, crushing the chips in the process:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At the end (which took 20 minutes, an order of magnitude longer than a friteuse), the results were less than spectacular. The chips had shrunk considerably, they were unevenly browned, a number of them had been reduced to crumbs, and two got stuck in the central arm:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's clear to me that it will go back. But not to Yvonne: she prefers the results! I suspect it's just because they've been cooked significantly more than we normally do.
And what about the health impact? Clearly we're consuming half a “tablespoon” of oil each with the meal. Do we have that much in deep-fried food? I don't think so.
| Saturday, 29 June 2013 | Dereel | Images for 29 June 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More panorama stuff
|
Topic: photography, animals, opinion | Link here |
I've wound back my house photos quite a bit, at least for the winter, but today there was what passes for fog in Australia, so took a few more photos than usual:
|
|||||||||||||||||||||||
For some reason the cats also wanted to get in the picture, so spent some time replicating them:
|
|||||||||||||||||||||||
That one worked well, but another didn't:
|
|||||||||||||||||||||||
It's not clear what caused that problem. The areas in question were all present, but somehow they didn't get included in the image. More investigation needed; I suspect that it might have to do with the difference in exposure. Rearranging the masks gave me the results I wanted:
|
|||||||||||||||||||||||
|
A new fridge
|
Topic: general | Link here |
So we've finally bought a new fridge, the one we were looking at on Saturday, and it arrived today. We put it in place of the freezer, which in turn will replace one of the old fridges in the laundry. There's a separate freezer compartment in the new freezer, of course, and we'll use that for the more common things. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I decided to pay the $55 delivery charge rather than to pick it up myself, and I'm glad I did. The bloke who delivered it had enough difficulty getting it through the door, and I could imagine the problems I might have had if I had done it myself. It was enough work just installing the thing, which involved moving both the fridge to the right and the rickety bookcase to the left.
So now we have finally powered down our two old energy-guzzling fridges. Hopefully that will make itself felt in the electricity bill. But getting the contents in the new fridge was a challenge: the total volume of the two fridges was about 700 l, and the new one is only 357 l. As planned, I saved a lot of space which I had used for beer-related things, but even so I only just made it.
|
Australian Borzois
|
Topic: animals, opinion | Link here |
While cleaning things off the fridge, found an inappropriate use for one of the stickers:
|
|||||||||||||||||||||||
Yvonne had a better idea, though:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That one really was made in Australia.
|
Goodbye Evi Nemeth
|
Topic: technology, opinion | Link here |
Sad news went around a couple of days ago: Evi Nemeth was lost at sea at the beginning of the month. After Jim Gray that's the second person I know who went out to sea and never came back. And they were both well-known computer people, not exactly the kind of person you'd expect to end like that.
| Sunday, 30 June 2013 | Dereel | |
| Top of page | ||
| previous day |
|
w3 lives!
|
Topic: technology | Link here |
Yesterday was the day when RootBSD moved their server room, taking with it w3.lemis.com, which had an uptime of 4¾ years:
When I came into the office this morning, w3 was still up. Moved or not?
Clearly it hadn't been rebooted. But had it been moved? The answer is in the uptime: yes. Subtract the uptime from the times above and the first gives a boot time of 20:31 on 28 September 2008, while the second gives a boot time of 00:09 on 29 September 2008. That suggests a time offline of 1 hours, 38 minutes. That ties in well with what I see from my network connectivity information:
The change in date is also visible elsewhere. Here before, then after:
It's also interesting to note the CPU time. Looking at /var/log/messages, it seems that a lot of that happened:
This sort of message was typical for this release of FreeBSD, and I don't propose to investigate it. So now I can continue my quest for 5 years uptime.
But is that fair to call it uptime? Clearly the VM was frozen and then thawed again. But it wasn't rebooted. My web server is now a day younger, but it has still been running since the machine was booted:
And that's probably the best definition of uptime: that the system has not required a reboot. It has been off the net briefly from time to time, which is normal in this kind of situation. But many of the processes are still the same processes I started back in 2008.
|
Inundation
|
Topic: general | Link here |
Yvonne sleeps in a water bed, which has worked well for some years now. But today the inevitable happened: it sprung a leak. What a mess! And now we need to find a new one.
|
Reasoning with irrational people
|
Topic: general, opinion | Link here |
It takes all kinds of people to make up an IRC channel, and one person I've found hard to get on in the past has no real name that I know of. He calls himself ^X, and lives in Singapore. I think he's still going to school. In the past he has made himself objectionable with his unjustified hatred of all people Muslim—and this in Singapore!
This is a long diatribe. I've divided it into:
Today during a relatively normal discussion he objected to the word Dravidian:
It seems that he objected to the use of the term because, he says:
That's certainly more than OED knows:
Etymology: < Sanskrit drāvida pertaining to Draviḍa, name of a province of southern India.
But his real objection seems to be that he didn't want to believe in any difference between the people of North India (Aryans) and South India (Dravidians).
It's clear that the phenotypes are very different: Southern Indians tend to have very dark skin and invariably dark brown eyes, while the northerners have lighter skin and often green or blue eyes. There's also a difference in physique that I wouldn't try to summarize.
And then there's the language, again (Indo)-Aryan and Dravidian. It's well-known that Sanskrit, the ancient Indo-Aryan language, is closely related to Greek and Latin. Nobody disputes that. And southern languages are very different. But ^X has different views:
He then came up with a newspaper report on genetic disorders in Indian people, suggesting that the difference in genotypes is much less than previously thought. Certainly modern DNA analysis is coming up with lots of new information, and I need to read up on that. Indeed, this page states:
The supposed Aryan invasion of India 3,000–4,000 years before present therefore did not make a major splash in the Indian gene pool...
Certainly something to follow up on. But I can't see how that could blur the fact of continuity of language. So I left the DNA matters for later and asked:
If there's one thing that is beyond any reasonable doubt, it's that Sanskrit is an Indo-European language—that's why the term was coined. But not ^X. He posted a link to an email. The site appears to be no longer maintained, but was clearly intended to promote Hindu values. Nothing wrong with that, of course, though the attacks on Muslims are less pleasant, but it's not exactly where you would go to find unbiased research.
This particular message was clearly written by somebody with an axe to grind. But indeed it seems that the theory of an Aryan invasion of North India could be flawed. But as I see it, the emphasis is on the word invasion, not Aryan. But these statements seem simplistic:
Opponents of the Aryan invasion theory claim that it denies the Indian origin of India's predominant culture, and gives the credit for Indian culture to invaders from elsewhere.
As if any country in the world has escaped invasion! Where did the first people come from? The DNA analyses are finding that out.
it suggested that Indian culture was not a culture in its own right, but a synthesis of elements from other cultures
Of course. Like all cultures, with the barely possible exception of the Australian aboriginals. What's wrong with that?
In any case, no reason whatsoever to reject the theory of an Aryan migration. There is too much evidence to refute that.
But ^X went further. He claimed that there's no difference between southern and northern Indian languages, and that Sanskrit is unrelated to Greek or Latin. I pointed him to this page, from Wikipedia, because it's available for checking. But that's not reliable enough for ^X:
To prove his assertion, he came up with a number of statements that I don't understand:
But then he made more obviously wrong statements:
That, of course, shows his complete lack of understanding: firstly, in all the older languages the verb was at the end, but whether subject or object comes first is unimportant, even in modern languages such as German:
Ich sah, dass den Hund der Mann biss.
Ich sah, dass der Mann den Hund biss.
The words are very similar to English. You might read the first one as “I saw that the hound the man bit”, but in each case, this translates as “I saw that the man bit the dog”. Subject and object are identified by the declension of the article, not by the position, and the verb is at the end of the sentence.
So I quoted from Encyclopædia Britannica and my book on Sanskrit, which he chose to ignore. Instead—this seems a typical tactic for such people—he returned to attack Wikipedia again, though he did say:
It's difficult to answer that, since there is no dissenting view.
Clearly he hasn't been reading what we wrote, let alone what my Latin quote means. And his last statement contradicts what little argument he has. Peter, whose wife is Indian, states that Hindi still places the verb at the end of the sentence, which is confirmed here. But it seems that his main thrust is Malayalam, which he considers to be “80% Sanskrit”. At first I thought that his reference to “verb subject object” was in reference to Malayalam, but Wikipedia proves me wrong:
Malayalam has a canonical word order of SOV (subject–object–verb) as do other Dravidian languages.
So it's not clear what languages put the verb first, with the obvious exception of LISP.
And the vocabulary? This page shows the similarities in Indo-European languages. Here a short extract, with help of this Malayalam dictionary:
| English | Latin | Greek | Sanskrit | Malayalam | ||||
| Father | Pater | Piter | Pitar | Janayithavu | ||||
| Brother | Frater | Phrater | Bhratar | Sahodaran | ||||
| Me | Me | Me | Mam | Yenikku | ||||
| Six | Sex | Hex | Ṣáṣ | Aaru | ||||
| Seven | Septem | Heptá | Saptá | Eezhu | ||||
| Nine | Novem | Ennéa | Náva | Onpathu | ||||
| Ten | Decem | Déka | Dáśa | Pathu | ||||
| Knee | Genu | Gónu | Jā́nu | Muttukuthuka | ||||
| Nose | Naris | Nas | Mukku, Nasika | |||||
| Mouse | Mus | Mũs | mū́ṣ | Eli |
I've deliberately chosen words that haven't deviated too much in sound or meaning in the Indo-European languages, but I didn't translate into Malayalam until afterwards. It looks as if the alternative name nasika for nose might be one of the 80% of Malayalam words that is derived from Sanskrit.
Round about here I gave up for the evening. I suggested to ^X that he come up with some better references, and that I would write this article. Jashank Jeremy (also half-Indian) continued for a while, but basically there was little to do.
^X did indeed come up with some references, round midnight our time. A number of books: “Archaeology and Language: The Puzzle of Indo-European Origins”, “Foundations of Language”, “Language - Vol. 32, No. 1, M. B. Emeneau”, “Lucian's A True Story: An Intermediate Greek Reader: Greek Text with Running Vocabulary and Commentary”, “The Sanskrit Language By Thomas Burrow”, “Languages and Nations: The Dravidian Proof in Colonial Madras”, http://www.scribd.com/doc/137894386/Sanskrit-Introduction and http://www.youtube.com/watch?v=4jHsy4xeuoQ. I had asked for ISBNs, but he didn't even name the authors, let alone what passages in the books support his theories. Based on the titles, I'd guess most disprove them. The one text link confirms what the rest of us have been saying all along:
Language family: Indo-European → Indo-Iranian → Indo-Aryan → Sanskrit
Sanskrit (सं स् कृ तम् saṃskṛtam Sanskrit pronunciation: [sə̃sɹ̩t̪əm], originally सं स् कृ ता वाक् saṃskṛtā vāk, "refined speech"), is a historical Indo-Aryan language and the primary liturgical language of Hinduism and Buddhism.
I wonder if he read any of this before posting the titles. I can't even be bothered to look at the YouTube video.
This has been a long rant. Why did I do it? Three reasons, I suppose: firstly, I didn't want to give up without trying. Secondly I was interested to see if he could come up with any substantiation for his claims. And thirdly, it helps me understand things better too. I would never have compared Malayalam to Indo-European languages otherwise. But mainly I really wanted to try to convince him. Given the certainty of the material, that should be simple. But I failed.
Why? It seems to be because ^X has a religious fervour associated with his beliefs, and he's not prepared to think about it. In many ways he reminds me of Wendy McClelland. In each case they spout off on things they don't understand, ignore criticism they can't handle, and go back to pick on uncertain details which are no longer relevant. And, like Wendy, when he finally provides references, they disprove him. My best guess is that ^X has been inflamed with Indian extremist fervour, which is no better than Muslim or Christian extremists. Good luck to him.
He may have some extremist views, but they're not associated with an organized religion: he claims to be an atheist. I still have no idea what motivates him. By comparison Wendy McClelland is an open book.
|
Rye bread rolls
|
Topic: food and drink, opinion | Link here |
I've been experimenting with different compositions for bread rolls for some time, and lately those made of 50% each wheat flour and “Manhattan light rye” bread mix have tasted quite good. But I know that the Manhattan light rye, despite its colour, is only 12.5% rye. Today I tried baking with 50% wheat and 50% pure rye flour. I had tried this before with sourdough, which was not a success: the rolls were far too firm. It turns out that yeast doesn't help: they're still too firm. So it's back to the Manhattan Light Rye, though next time I might try with sourdough.
|
Direct delivery email problems
|
Topic: technology, opinion | Link here |
So a few days ago I unblocked port 25 and started delivering email directly to the destination MTA. And today I discovered:
=== grog@eureka (/dev/pts/12) ~ 35 -> mailq
The problem here is that the address that the remote MTA sees is a dynamic address from Internode's pool, and it might once really have been abused. So there's nothing for it: back to the mail tunnel.
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 |