|
| Greg's diary |
| January 2010 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Friday, 1 January 2010 | Dereel | Images for 1 January 2010 |
| Top of page | ||
| next day | ||
| last day |
|
Spamassassin: still in the 20th century
|
Topic: technology | Link here |
Yvonne in today to tell me that she had found legitimate mail in her SpamAssassin folder—false positives. Investigation showed:
I've long since learnt that it's futile to try to find the meaning of keywords in the SpamAssassin “documentation”, but the meaning appears clear: all mail with the year 2010 was being targeted. That could be a problem with my version of SpamAssassin, of course, but I had only installed it (with much pain) seven months ago. In any case, there's an update service, once I found it: sa-update. Ran that, and it finished immediately, so checked the return code:
=== root@dereel (/dev/ttyps) /usr/local/share/spamassassin 265 -> sa-update
=== root@dereel (/dev/ttyps) /usr/local/share/spamassassin 266 -> echo $?
What went wrong there? Tried running with a presumed -v (verbose) option, but of course it didn't have that:
=== root@dereel (/dev/ttyps) /usr/local/share/spamassassin 267 -> sa-update -v
OK, I can live with that, even if it violates POLA:
=== root@dereel (/dev/ttyps) /usr/local/share/spamassassin 268 -> sa-update -D
So, for some reason, SpamAssassin thinks that 1 is a good completion code to return on success. And it thinks I'm up to date. What a pain! Gave up and changed the entry in the scores table to make the value 0. But I have to wonder why I bother using this software; not only has it passed its use-by date, it's telling me as much.
Finally got round to trying to install Ubuntu 9.10 on my old Dell laptop. It's slow! And after a couple of hours copying, the installer reported an uncorrectable I/O error—and wouldn't tell me whether it was the DVD or the USB stick. I think I should do the installation on a disk and then copy the entire disk to the USB stick.
The installation set off an unexpected chain of events: the laptop was connected to the UPS, and in conjunction with dereel doing its monthly level 0 dump (and maxing out all 4 cores with compression), it used enough power to overload the UPS. Tried first switching dereel's display to a vty to lower the load on the display cards, and then turned off the monitors, all to no avail; I had to move the laptop to a non-protected power point, in itself not an issue.
Then back and turned on the first monitor, and switched back to X. It crashed: the other monitors weren't on, so it couldn't find its displays. What a pain!
| Topic: opinion, technology, general | Link here |
And another New Year, and a new decade too! The millennium is 10% over already.
Of course, many people claim that it just isn't so. The first century (and the first millennium) started in the year I and went to the year C, and so on, so this century didn't start until the year MMI, and the next decade doesn't start until MMXI. The problem here is that we've changed our way of counting since the first definitions of the current calendar (which didn't include terms such as “decade”, “century” or “millennium”). We have also moved from FORTRAN to C: consider the two statements:
The first gives you an array FOO (clearly) of 20 floating-point (implicitly) elements, numbered 1 to 20. The second gives an array foo of 20 elements, numbered 0 to 19. In general, people tend to think of this change as a Good Thing. But when it comes to dates, we're still stuck with a number system that didn't have a 0.
Some argue that measuring time periods in the modern manner would cause a conflict with older use. But what older use? The Oxford English Dictionary first mentions the term “century” in the calendar sense in 1638, and the term “decade” isn't used in the calendar until 1837 (in full, “decade of years”; like centuries, there are many other kinds of decade). And “millennium” doesn't appear at all in the calendar sense; it's a religious term.
Still, the OED shows that the controversy isn't new:
1846 Knight Pass. Working Life I. §i. 18 The learned had settled, after a vast deal of popular controversy, that the century had its beginning on the 1st of January, 1801, and not on the 1st of January, 1800.
In a time where people redefine terms such as “billion” to reduce their value by 99.9%, where they say things like “in the eighteen hundreds” to mean ”in the nineteenth century”, why do people insist on inconvenient calendar definitions?
Woke up in the morning to hear ABC radio news, with information about the New Year's Eve festivities yesterday: I was told that it was a Blue Moon, and that they had specially decorated the Sydney Harbour Bridge with a blue moon to celebrate the fact. The only problem is: there was no blue moon in December.
A blue moon is now defined as the second full moon in the month (here, too, there's an earlier, conflicting definition). We'll have one later this month. Conveniently, Emacs has a function phases-of-moon, which tells me:
So why the incorrect claims? Today's full moon was at 19:11 on 31 December 2009 UTC, but our time zone is UTC+11, so for us it's in January. That also applies for most countries in Asia, including India, where they published in the Deccan Herald:
On a full blue moon day on Thursday, skywatchers will have a bonanza when they find the lunar radiance, rather dim, as the country witnesses a partial lunar eclipse. ... The start of the partial eclipse will be at 00:22:43 am on January 1, 2010. It will end at 01:22:41 am.
That article contradicts itself. How can people make such mistakes? Checking on the ABC web site, I no longer found the claims that I had heard on the radio. Instead, one page states:
In keeping with this year's theme of 'Awaken the Spirit', there were plenty of blue sparks as well as a giant blue ring on the Harbour Bridge.
No idea what “Awaken the Spirit” means, but clearly they've changed their tune. The only reference I could find stated:
New Year's Eve also presented the world with a blue moon - the second full moon appearing in for the month — for only the second time in nearly two decades
That's doubly wrong if they're applying it to Australia. Blue moons aren't that seldom. The last one was four years ago, and we had not one but two in early 1999: two full moons in January, none in February, and another two in March. The interesting thing about them is that they tend to come in groups: the lunar month is only slightly shorter than the calendar month, so the date of the full moon tends to wander round the same time of the month.
ABC and the Deccan Herald are not the only ones to get it wrong. The Australian Fisheries Management Authority publishes a calendar which also claims two full moons in December 2009. And they explicitly state that the times are AEST (UTC+11 at this time of year). They also claimed a blue moon in December 2005, at least for UTC. Like this year, it could have been in January in Australia.
| Topic: gardening, photography | Link here |
As expected, the Crocosmia and Watsonia are both blooming:
|
|||||||||||||||||||||||
It's really difficult to get good photos of them. I need to work harder at it. Managed to get some photos of the caterpillar I saw yesterday, which had moved only marginally in that time:
|
|||||||||||||||||||||||
At least this time it's tolerably sharp, but I need to do something about the background. There must be an easier way to create better backgrounds.
Our cathedral continues to disintegrate:
|
|||||||||||||||||||||||
I fear it'll be completely gone in the not-too-distant future.
| Saturday, 2 January 2010 | Dereel | Images for 2 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
SpamAssassin: resolution
|
Topic: technology | Link here |
Other people are also gradually discovering the SpamAssassin bug—I suppose it's normal that we have this kind of thing first. Got a message from another list warning me about it, and went off again to get output to prove the issues I had had—but this time sa-update ran, and it replaced my workaround with a “solution”: now the rule matches dates from 2020 onwards. Much searching uncovered a bug report that the Bugzilla search engine couldn't find, dated after we discovered the problem. In the process, discovered that sa-update does return completion code 0 when it does an update; the 1 must have meant “no updates found”, which makes sense.
| Topic: photography | Link here |
Photo day today, and time to try out the new panorama arrangement with the camera on the focusing rail to eliminate parallax. Success! Hugin processed my difficult verandah panorama in automatic mode with almost no discontinuities, a far cry from last week (first photo), despite a significantly higher vertical angle of view:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I should have done this earlier, but I didn't think that it would be that simple.
| Topic: general | Link here |
The weather has been amazing lately. The last two days of last year were unpleasantly hot, with temperatures in the high 30s. Yesterday and today have been quite cool: the highest temperature we've had all year was 21.5°, and that was round midnight on New Year's Eve, as the last year cooled down. Today it barely went over 21°, and most of the time it's been in the mid-15s. So much for global warming!
| Sunday, 3 January 2010 | Dereel | Images for 3 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
HDR and time lapse software
|
Topic: photography, technology | Link here |
More playing around with processing software today. Daniel O'Connor gave me a link to some ideas on HDR software, so went and read that. He uses a program called align_image_stack, a particularly undocumented part of Hugin. Finally found some documentation of a sort and tried out his suggestions with five photos I had taken of the lagoon yesterday. The invocation:
=== grog@dereel (/dev/ttypa) ~/Photos/20100102 213 -> align_image_stack -a tmp lagoon*
=== grog@dereel (/dev/ttypa) ~/Photos/20100102 214 -> enfuse -o lagoon-HDR.jpeg tmp000*
The results are interesting. Here the middle of my group of 5 shots on the left, and the HDR version on the right:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Both of these photos were “natural”, in other words not optimized like the other five photos in the day's collection.
align_image_stack is supposed just to adjust individual images to the same shape and size, so it looks like it should be good for time lapse too. Tried that with one of my sequences of photos, with less than stellar results:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The second image had been adjusted to match the first (note the black areas top and right), but I have no idea how it messed up the next two. It went on like that, so it's clearly not the most straightforward thing to use. At a guess it's choosing the wrong control points, but how do I tell it otherwise?
| Topic: gardening | Link here |
Finally got round to mowing the lawn, the first time I've done it with a push mower for as long as I can remember. Things went fine until I ran out of petrol. After refilling (1:25 two stroke mix) it spluttered around for a while and then stopped. When I tried to restart it, the whip cord came off in my hand. Looks like a trip to Ballarat for that one.
After that, I should have prepared to finally tie up my hops, which have now grown several metres. Ran into more trouble, not to mention flies, and decided to put it off until tomorrow.
| Monday, 4 January 2010 | Dereel | Images for 4 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Focus stacking: how not to do it
|
Topic: photography, technology | Link here |
More playing around with multi-shot photo techniques today. I had read that the combination of align_image_stack and enfuse should also be able to create extended depth of field photos (“focus stacking”). Tried it with some photos of the Crocosmia with remarkable lack of success. The photos were taken on a tripod with 2.5 mm focus increments, giving images that were partially in focus, with a horizontal width of about 2.5 cm. The output photo (last image) had uniform focus: it was all out of focus:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly more experimentation is needed there.
|
Finding a ring flash
|
Topic: photography | Link here |
One thing that has been quite a success is my cheap ring flash attachment; it almost invariably gives better lighting than the normal flash or the studio flashes. The main downsides are the exposure and the unwieldiness of the beast, though it would be nice to have something with a 360° lighting angle: mine is missing about 60° at the top. There are more expensive dedicated units available on the market, notably the Mecablitz 15 MS-1 digital and the Olympus SRF-11. Spent about 5 minutes looking for the former on the web site; their search engine doesn't seem to understand “ring flash”, and they've hidden it under “slave flashes”, which it is not (it can happily run on a sync cable).
Years later I bought the 15 MS-1. It really is a slave flash: it doesn't have a sync cable, presumably because it needs the communication with the camera's flash protocol. But it's still silly not to categorize it as a ring flash.
Downloaded the documentation for the two and was greatly disappointed. The Mecablitz (or is that mecablitz?) attaches to the filter mount, and it weighs 190 g. The English version of the badly translated (and copy-protected) manual reads:
Recommendation: Only use internal focusing objectives! Note, with external focusing objectives you must perform the focusing manually. The risk of damage to the flash unit due to the weight of the objective or of the adjusting motor cannot be ruled out.
That's backwards, of course. Whoever translated that didn't understand it. It would be easy to think that the translator also didn't understand German very well, but there are some idioms there that look as if they were written by a German. In addition, the terms “internal focusing” and “external focusing” aren't very widespread; I couldn't find any useful web reference to the latter. The refer to whether the front of the lens (filter mount) moves when focusing (external) or whether the entire focusing movement is inside the lens (internal). The documentation should have mentioned that. A better translation of the original, along with clarifications, is:
Recommendation: Only use lenses with internal focusing, where the front of the lens does not move during focusing. If you use lenses with external focusing (where the lens moves), focus manually. If you use autofocus, the weight of the flash unit can damage the lens or the focusing motor.
Still, that's a problem. The main lens I want to use is the Zuiko Digital ED 50mm F2.0 Macro, which is external focusing. I can't imagine any macro lens that isn't: the focusing range is too big for internal focusing. So that kills the Mecablitz.
Revisiting this article in September 2017, I note that the current two Olympus macro lenses, the M.Zuiko Digital ED 60 mm f/2.8 Macro and the M.Zuiko Digital ED 30 mm f/3.5 Macro lens, do in fact have internal focusing.
The Olympus flash unit appears to connect to the lens hood bayonet, but it also has severe restrictions. The instructions (PDF) specify (on page 9) a very short list of lenses with which it can be used—only five. In particular, my Zuiko Digital ED 12-60mm F2.8-4.0 SWD is not listed. The 50 mm macro is, but the other macro lens, the Zuiko Digital 35mm F3.5 Macro is not. And for the 50 mm I need an FR-1 adapter ring, for which B&H ask the princely sum of $99.95—more than twice what I paid for my ring flash adapter. Somehow this doesn't make sense.
And what about the angle of flash, one of my concerns? It's not clear that the Olympus or Mecablitz do even as well as my present adapter; the Olympus appears to have four individual flash tubes. The Mecablitz doesn't clarify, but from the illustration on the front of the manual, it looks as if there are just two tubes, one left and one right, each with an angle of only 30°.
The instruction manual is (marginally) more help. It has adjustable angles of 0°, 10° and 20°, and also a diffuser, as I note in September 2017.
So until something better shows up, it looks like my current ring flash adapter is the best, regardless of price.
| Topic: gardening, brewing | Link here |
Another attempt to tie up my hops today, and we got further, attaching some fence mesh to some wood to hang in the gutter of the garage. But the wood, which I had also used last year, is rotting, the mesh made it very unwieldy, and it finally came undone again. Left it yet again to think of a better alternative, in which I was successful. There are three lengths of fencing mesh; we don't need to attach them all to the same piece of wood. By using individual pieces, it'll be a lot easier to lift, especially when the temperatures drop again (today they hit 32.9°).
| Topic: gardening | Link here |
One thing that looks on the road to recovery is our Staghorn fern. Six weeks ago it was looking as if it were dying, and I added some Sphagnum moss to keep the insides moister. That seems to have been the thing to do; the outside now feels moist, and new shoots are emerging:
|
|||||||||||||||||||||||
It's possible, of course, that the shoots would have come anyway, but the time frame and the moistness suggest that it was because of the moss.
| Tuesday, 5 January 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Satellite ISPs: which tariff?
|
Topic: technology, opinion | Link here |
My weather software has an interesting side effect: I'm transmitting more data, and last month I used up my quota. That's easy when you only have 2 GB/month between midday and midnight. Went to the web server, (now Aussie broadband, not Wideband) to check the tariffs. That looked good enough; the prices have dropped further, and I can now get an “Aussie 14” contract for $74.95, only $10 more than I'm paying now. But my current plan (“Aussie 12”) was marked at $49.95, $15 less than I'm paying now.
This isn't the first time I've discovered by chance that they've dropped the prices and not told me. Rang up and spoke to Jess, who told me that they had sent an email message to everybody concerned—just like last time. But I keep all mail from my ISP, even irrelevant stuff, and I hadn't received it. Asked to speak to a supervisor, who wasn't available, but who promised to call back.
Shortly later got a call back from Kylie, who told me that they hadn't sent email, but that it was on the web server, and it was my duty to check the web server. She couldn't tell me where it was on the web server, however, and suggested that it might no longer be there. I told her that I didn't find that acceptable, and that even if they had the right to do so, they wouldn't be acting in the interests of their customers. Asked her to inform management of my dissatisfaction.
A bit later, a call from Ray McClusky, the sales manager. He told me they had sent out paper mail before Christmas informing people of the changes, since people don't read their email. That's possible, and the fact that it—and no other mail I know of—didn't arrive (yet) may be due to industrial action that the postal workers started before Christmas, but it should still have got here by now. It's also strange that Jess and Kylie knew nothing of this action (13,000 letters!). He promised to send me a PDF copy of the letter in a couple of days' time.
All this is somewhat irrelevant, unfortunately. All the new tariffs count both upstream and downstream data, while my current one only counts downstream traffic. My upstream traffic is nearly as high as my downstream, so this would mean that I would have to go to a much more expensive tariff to get any improvement in my effective traffic limit. Hopefully I'll find an alternative before much longer; it seems that WiMAX is looming on the horizon, and maybe I'll find something there. At the very least I'll find whether I'm in line of sight of a tower (probably Mount Buninyong).
|
Focus stacking: what increments?
|
Topic: photography, technology | Link here |
My experiments the other day with focus stacking were less than successful, and there are a number of things I need to investigate. There's a newer version of Hugin out there, and thus newer versions of align_image_stack and enfuse. I'll install them tomorrow morning when I have bandwidth.
The other issue is: how far apart should I make the partial images? That depends on the depth of field, of course. And what's that?
There are plenty of online calculators for this, but they're inconvenient to use in the field. And it's straightforward enough to calculate, so decided to write my own, based on the formulae in the Wikipedia page.
That went easily enough. But any depth of field calculation makes assumptions, and I needed to know whether mine came up with the same results as the other calculators. And, of course, it didn't—not with any of them. On the other hand, the online calculators don't agree with each other either. So much depends on the circle of confusion and other considerations (in Cambridge in Colour I read: “An acceptably sharp circle of confusion is loosely defined as one which would go unnoticed when enlarged to a standard 8x10 inch print, and observed from a standard viewing distance of about 1 foot.”). Spent some time trying to calculate what this really meant, and in the end settled on a circle of confusion of 15 µm, which seems to work. For my 50 mm lens at f/11 and 1 metre distance, I get:
| Calculator | Hyperfocal distance | Close limit | Far limit | |||
| Mine | 15.152 m | 0.938 m | 1.071 m | |||
| Wrotniak | 0.941 m | 1.070 m | ||||
| Cambridge in Colour (4/3) | 0.937 m | 1.072 m | ||||
| Cambridge in Colour (APS-C) | 0.923 m | 1.091 m | ||||
| DOFMaster | 14.800 m | 0.940 m | 1.070 m | |||
| Photo96 | 14.731 m | 0.939 m | 1.069 m | |||
| Punit Sinha | 15.201 m | 0.941 m | 1.067 m | |||
| OnlineConversion (APS) | 9.091 m | 0.905 m | 1.117 m |
That's relatively consistent, but only relatively. Why are there such big differences? In particular, the hyperfocal distances and depths of field don't seem to quite match. Still, so far things seem to work. Now I need to do a separate series of calculations for macros.
| Topic: gardening, brewing | Link here |
Finally put up two sections of wire mesh for the hops. They've all grown together, of course, and lifting them up to hook them into the mesh was a pain. We'll see what happens. Hopefully things will be better next year.
| Topic: food and drink | Link here |
Huevos rancheros for breakfast today, and once again had problems with my tortillas. And by chance we also had pollo en adobo for dinner, so had to make more tortillas.
I've come to the conclusion that tortillas made with masa harina are a very different beast from the ones you buy in restaurants: the masa is crumbly, and because there's no gluten, there's no way to join the stuff together. The result is that you need to oil the surfaces to handle the tortillas at all, and the result still doesn't taste the same. In addition, the ratio of water to masa is very critical, and in my experience it seems to vary. Initially I tried 1.33 parts water to one part masa and ended up with a dough that wouldn't hold together. I went all the way to 1.7 parts water before I got a satisfactory result. But this morning that was far too watery, and I had to add more masa to get the right proportions. In the evening, I tried 1.5 parts, and that worked.
Part of the problem might be my scales; I suspect that the firmware of modern digital scales make assumptions about errors and attempts to recover from them. In particular, if you weigh something on digital scales and then remove the bowl, the scales should return to 0. On both of my scales, what I see is that they return to near zero, stay there for a while, then go to exactly 0. I'm sure that this is some kind of programmed correction, but it seems to apply to creep in general, and I'm not convinced that the device will weigh correctly if you add things in small quantities. So I've taken to removing the bowl from time to time to confirm the final weight (which, in fact, seemed correct).
| Wednesday, 6 January 2010 | Dereel | Images for 6 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Updating hugin
|
Topic: photography, technology | Link here |
As planned, upgraded my installation of hugin today. My concerns were justified: I couldn't even build the file until I reconfigured; it seems that there are two versions of a dependent port, called autopano-sift and Autopano-sift-C, and I had to choose one. Tried the first and waited an eternity for the hugin package to build, producing nice, pretty and illegible output:
|
|
||||||||
That's white on beige; if my background had been pure white, it would have been completely illegible.
When it was finally done, started up, and it fell in to a heap on the floor. After selecting the files, I got the message:
Reading the Makefile for hugin showed that the preferred autopano was autopano-sift-c, so installed that, rebuilt hugin and reconfigured. Then I got:
|
|
||||||||
I suppose this is an incompatibility between the versions of autopano-sift-c and hugin, specifically the command line parameters for autopano-sift-c. But they're both from the current version of the FreeBSD ports tree.
Still, there was one other option: hugin now offers further alternative control point detectors, one of which is align_image_stack, the main reason I installed the new version. Tried that instead, and it ran. But the results it created were terrible: it wasn't able to process either of the panoramas I fed it, the ones that had no problems with the old version of hugin. At least it answered the question about using align_image_stack.
Spent at least a hour recovering the old versions from the backups. There should be a better way to do this. An obvious way would be to look at the contents of the /var/db/pkg/<portname> file and back them up from there. Maybe there's a way; I should investigate.
| Topic: general | Link here |
Another power failure in the middle of the night. That's an average of one every 3 days so far this year.
| Topic: animals | Link here |
Yvonne off to the vet in Bannockburn to have some horses pregnancy tested, and she took Tony, our remaining whippet, with her: he had some lumps on his skin which needed looking at. The vet confirmed: cancer, inoperable, and it could burst out at any time. So Yvonne left him there; he had a good life. He was nearly 13, and he outlived his litter sister Tina by 10%. But it's sad, and strange not to have a dog around the house.
| Thursday, 7 January 2010 | Dereel | Images for 7 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Ports collection: saving old ports
|
Topic: technology | Link here |
As I suspected yesterday, yes, there is a simple way to back up old versions of FreeBSD ports based on the package description file in /var/db/pkg/:
=== root@dereel (/dev/ttyp7) ~ 70 -> pkg_create -b /var/db/pkg/hugin-0.7.0_2/
Cleverly, you can either give the directory name or the port name (which is the basename of the directory). Thanks to Callum Gibson for the footwork on that one; I should have found out before I installed the new version.
David Griffiths had an old ThinkPad 380Z to give away; he lives in Daylesford, and he's more often in Ballarat than Melbourne, so he gave it to me. He even brought it here. It's not exactly the latest model: 10 years old, 225 MHz processor, 96 MB memory and 6 GB disk. But it's exactly what I need for a kitchen computer, the battery is OK, and it looks like it's hardly been used. I have an old Dell Inspiron 7500 which is slightly younger and has better specs, but it's falling apart.
| Topic: general | Link here |
Into town in the afternoon to take the lawn mower in for repair; it won't be ready until tomorrow, but they had repaired Chris' lawn mower, so brought that back with me. While I was in town, down to Midland Irrigation to find out about fire fighting equipment, and back with hoses and connections for the irrigation. There's more work to be done, but at least I've got started. Not too soon: they're predicting weather into the 40s over the weekend, and experience shows that it can be even hotter here.
| Friday, 8 January 2010 | Dereel | Images for 8 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Web breakage on the increase
|
Topic: technology | Link here |
I'm more or less accepting the conclusion that the average intelligence of computer users is diminishing. On the face of it, that's to be expected: people who have been in the game for decades entered when they did because of a specific interest in new technologies. Once the technology hits the mainstream, people look at it as a resource, as a tool, and to some extent I do too. People who have no particular interest in computers use them, and they're often shocked by the experience. So computer companies do what they can to make it more comfortable (not necessarily easier) for these people.
The result? Otto Normalverbraucher, the average consumer (the English link to Average Joe doesn't quite mean the same thing). Instead of advancing computer technology, big companies are maximizing their profits by catering to a very limited view of computing. Otto has a Microsoft “Windows” machine with a display of 1024x768 (oh, sorry, these numbers scare people. Let's call it XGA so that people know exactly what we're talking about) and various other components that don't interest me right now.
Why should I care? Mass market stuff is cheap. Special equipment isn't. I'd like to have a monitor which can display my photos in reasonable resolution. But there's almost nothing on the market beyond 1920x1200^WWUXGA, with 2.2 megapixels, enough to display a 19% rectangular portion of a photo at a time. And even this resolution—less than what I had 10 years ago—is disappearing in favour of 1920x1080 (which doesn't seem to have an *XGA buzzword).
I've come to accept most of this. But more and more webmasters (which, I predict, will someday become a derogatory term) have chosen to pessimize their web pages so that they will only render correctly on displays with such resolution. Today I discovered that O'Reilly have joined the club:
|
||||||||||||||||||||||
Examination shows that the box on the left overflows to the left if you set a minimum font pixel size of 15 pixels or more. Many web sites commit this kind of obscenity, but it's a shame to see that O'Reilly is now one of them.
One of the others is the Australian Bureau of Meteorology, who like so many others knows better than you do what layout you should have. Thus we get weather forecasts like the one below, where they have pulled in the edges and limited the width of the central display to a certain number of columns. The result is not overflow—that's something—but a gratuitous wrap of the most important information. Once again allowing fonts with 14 or less pixels solves the breakage and gives you flyspeck instead.
|
||||||||||||||||||||||
But I suppose people are used to that. Why?
I entered feedback about this particular error and got a message telling me that I would receive email from them, and if not, it would be because I entered the wrong email address. The things they know! I didn't get the feedback, of course: SpamAssassin did, because their email message came from the fictive DNS name sydwww-g.bom.gov.au:
Finally finished downloading FreeBSD 8.0-RELEASE and installed it on the Thinkpad. It recognized the wireless card (a Lucent Orinoco card about as old as the laptop), but wasn't able to bring it up. Investigation showed that they've changed the way of configuring these cards under release 8, and the installation software doesn't seem to know about it. It seems that the command needed now is something like:
But that didn't help much. After reboot the system died during boot with the message:
That's fatal, of course. Possibly it's related to the age of the laptop. But given the pain of messing around with this stuff, decided to install 7.2-RELEASE instead. That worked, and it recognized the wireless card. But I couldn't start X—a problem I've seen before.
That makes three serious problems: the third isn't FreeBSD's fault, but if nobody produces web browser plugins for FreeBSD, it might as well be. Started downloading a CD with the latest version of Ubuntu. I have a DVD, but the ThinkPad only has a CD-ROM drive, so I'll do from the CD. But that'll have to be tomorrow.
|
Kangaroos: a solution?
|
Topic: animals, gardening | Link here |
A few months back Diane Saunders visited us and brought her dogs. Roxy crowned herself in glory by eating Di's medication, but she also had fun chasing kangaroos—not a thing to recommend: kangaroos can easily kill dogs, though normally they only defend themselves. But nothing happened to Roxy, and we didn't see another kangaroo on the property for months.
Roxy's not here any more, of course—neither is Tony. But every time I've seen kangaroos in the paddocks, I've chased them back into the lagoon, which takes a bit of doing: they're not timid. Still, since then we haven't had another kangaroo in the garden, and we're gradually removing the mesh we put around a number of our plants.
|
Surviving the heat
|
Topic: gardening | Link here |
It's hot again, and it's clear that some of the plants aren't handling it well, in particular the Fuchsias. Took down the one in the hanging pot before the thing shriveled up altogether. I'll leave it down until the coming hot weather is over: today we had a maximum of 39.1°, compared to only 32.8° in Ballarat. That seems to be a typical temperature difference. They're forecasting 39° in Ballarat for Monday. If that's really the case, the temperatures here might rival the 46° days we had round Black Saturday.
There's another Fuchsia planted in the garden; since the demolition of the Cathedral it's in full sun. It doesn't look as bad as the other one, but I decided to put some shade cloth around it, wrapped round one of the mesh protections we had made against kangaroos. Hopefully that'll make it happier.
The “succulent garden” that Yvonne made last month is doing well, with the exception of the Ledebouria Here last month and today:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yvonne was pleased enough with the progress to make another one for Chris:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Topic: photography | Link here |
After taking the photos of the succulent garden, decided to do some more macro photography:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Once again, exposure seems to be a matter of luck. I've noticed that photos taken with smaller apertures (in the case in point, f/11) and TTL flash are frequently underexposed. It would be logical to expect that the flash just isn't strong enough to produce enough light under these conditions. Tried it with manual flash and confirmed that this was by no means the case: the test photo with full power and f/22 almost completely burnt out the image. So it seems that there are other, seemingly random factors here.
In passing, it's irritating that the EXIF data contains so little information about flash exposure. It would help to understand what's going on, but all I get is information on bounce (dubious), fill-in or not (what's the difference?), exposure compensation. Decades ago I kept a notebook of my exposures; maybe I should do that again.
| Saturday, 9 January 2010 | Dereel | Images for 9 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Web breakage: a solution?
|
Topic: technology | Link here |
A while back I installed a firefox add-on called NoSquint which can modify the size of displayed text. If I set the text size to the size that the O'Reilly webmasters expect, I can really display the text as the webmaster intended:
|
||||||||||||||||||||||
In this case, apart from the (presumably intentional) extreme width of the page, all is OK. Unfortunately, it's not always that simple, and the fonts for some pages look particularly ragged—presumably that's NoSquint trying to fake a non-existent font size. Still, the approach is correct: there's no earthly reason to expect a fixed ratio of pixel size to point size. Hopefully browser designers will come up with an idea of “virtual pixels” to work around HTML deficiencies.
| Topic: general | Link here |
We've been here 2½ years already! And I still haven't done a lot of the things I had planned on moving in, such as putting up shelves. The half year makes itself felt: it was stinking hot, over 40°, and we spent most of the day inside.
| Topic: photography | Link here |
A bit more experimentation with the weekly photos today. It's becoming clear that the north and north-east views get more even light in the afternoon, so took the photos then. The exposure is now much more even, particularly when it's sunny. Here an example (today is the second one):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I've been concerned about exposure accuracy for some time, and I sometimes idly wondered if it was a problem with Olympus cameras. Today I revisited some photos taken by Yana or with her camera (a Canon EOS 30D), and found the same exposure issues, particularly with flash. Particularly given the low dynamic range of current digital sensors, you'd hope that you could get better exposures.
Also did some playing around with another panorama, the one in the south, where the shadow of the birch trees makes exposure difficult. Tried turning off the manual exposure settings I normally use for this kind of photo. The results were not encouraging (first manual exposure, then automatic):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, the details in the birches are better, but only at the expense of the rest of the image. Maybe I can do better with more software intervention, but currently I'll stick with manual exposure.
| Topic: general | Link here |
Chris Yeardley along in the evening for dinner. For as long as I've known here, she's had waist-length hair, usually in a ponytail. Now she's had it cut short:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Sunday, 10 January 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Software installation: the pain
|
Topic: technology | Link here |
Got round to installing Ubuntu 9.10 from the “alternate” CD image on my ThinkPad today. In contrast to FreeBSD 8.0, it recognized the wireless card immediately and configured it via DHCP without as much as a by-your-leave. It then proceeded to spend fully 3½ hours installing its choice of software, including OpenOffice, but not Emacs. Rebooted, which took forever, with just an emblem instead of boot messages. Finally found out how to display the messages (hit ESC), and found it was trying to configure the wireless card via DHCP, and failing badly:
I couldn't stop it with ^C (you can with FreeBSD), but it finally timed out. And yes, the device was there, but dead in the water. It looks as if the installer forgot the firmware.
This is supposed to be a quick install for a machine for the kitchen, which will mainly run a web browser. Why is it so difficult? I had started downloading an OpenBSD 4.6 ISO, but the network connectivity ground to a halt, and I hadn't finished it by midday. After midday the tariff switches, the volume is much more limited, and I tend to overrun it, so I put off the rest until tomorrow.
Instead found a FreeBSD 6.3 install disk and installed that—30 minutes. And it recognized the network card, installed it, and X works. In principle, everything's fine. But 6.3 is ancient, and I anticipate problems with the ports. Why are more recent versions of FreeBSD so broken?
The weather has been very hot lately, and the forecast for Ballarat today was even hotter than yesterday (35° today, 34° yesterday). It didn't happen like that here: yesterday we had a maximum of 38.7°, and today it was “only” 36.3°. But it reminded me that I wanted a graphic display of comparisons. Last time I did this, it was like pulling teeth. Today made a more general version and, well, it was like pulling teeth. gnuplot remains an enigma. After finishing the plot (left), there was only one problem, the date under the plot. That's the default, and I could live with that. But I'd rather have it gone: the time is sufficient. The info page for gnuplot tells me that %H and %M give me the hour and time:
Tried that and got the second graph. Why? This software makes no sense at all.
| Topic: general | Link here |
Another power failure today. That's an average of one every three days this year. Hopefully it won't stay that way.
| Monday, 11 January 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Laptop installation: OpenBSD
|
Topic: technology | Link here |
As planned, finished downloading the OpenBSD installation CD this morning and installed it on my long-suffering ThinkPad—the 5th OS installation in four days.
It's been years since I last installed OpenBSD, but the installation program doesn't seem to have changed much in the intervening time. Peter Hansteen wrote on this subject recently, even in relation to installation on a ThinkPad (though it's 11 years newer than mine).
Peter is used to the installer, but he still found it useful to describe the whole process in that link. But he was happy with the default disk partitioning. I wasn't on this machine, and I wouldn't have been on his either. I've had this argument with many people before, including OpenBSD people, but basically the time is past where you need more than one file system per disk. I've written a HOWTO describing my scheme, effectively as published in The Complete FreeBSD.
So: I had to partition manually. The program didn't help much; it didn't even use curses, and all values were shown only in units sectors of 512 bytes. But it wasn't that difficult to work out. Choosing what to install is quite a different matter (here Peter Hansteen's photo of the screen):
What is this stuff? In some cases I can guess, but in others I have absolutely no idea. What's bsd.rd? bsd.mp? I really don't know, though it now occurs to me that bsd.mp might be related to multiprocessor support, which would explain why it wasn't selected (the only one that wasn't). But the installation CD was only 230 MB in size, so even on this old ThinkPad there was no trouble installing everything.
Then: packages. I wanted Emacs and firefox, but I didn't want to have to compile everything on this machine. Checked pkg_add, which has been rewritten since the FreeBSD version and now has incompatible options. But after looking more carefully, it's actually quite clever: FreeBSD pkg_add takes either the file name of a package, or with the -r option the name of a package to be found somewhere the net. OpenBSD pkg_add has a different meaning for -r, but by default it accepts either a file name or a URL. That proved to be quite helpful. For it to work correctly, it needs a path, in my case:
That searches the local directory first and then the two ftp sites. It doesn't seem to have any difficulty with the double usage of “:”.
That took its time, of course—though less than the base installation of Ubuntu—but it worked. And after a bit of messing around with my configuration files I had a functional (if somewhat slow) system. The only thing that I haven't been able to do is to remap the keyboard to put the Ctrl key where God intended; it insists on generating a CapsLock in any case. I assume this is hard wired, and, like the silly nipple mouse, I'll have to put up with it.
|
All-time-high temperature
|
Topic: general | Link here |
As threatened, the weather was even hotter today. The Bureau of Meteorology increased its forecast maximum temperature for Ballarat from 39° to 40°, and here it hit 44.9°, the highest since I've been keeping detailed records. Somehow I'm becoming less tolerant of these temperatures. cvr2, my CVR, also hung again. I must get a new motherboard for the machine. Also had a number of satellite dropouts, which Aussie Broadband claim are due to extreme heat. I have continual dropouts, of course, even when it's cold, but it's barely possible that this time they're right.
The Bureau of Meteorology has a difficult time making forecasts, of course, but sometimes I wonder how much care they put into preparing them. Today we had:
So the highest temperature on Wednesday is 1° less than the lowest for Tuesday. Clearly this implies that both temperatures occur round midnight Tuesday, and that there will be a step change in temperature of -1° at that time.
| Topic: animals | Link here |
Another potential casualty of the heat is Lilac, our 13 year old Burmese cat. She looks in the best of health, and it's no surprise that she's moving slowly in this weather, but she seems to be having difficulty jumping recently. We have always fed the cats on top of the washing machine, where she has to jump about 60 cm, but lately she can't make it. I hope that's not a sign that she's fading away.
| Tuesday, 12 January 2010 | Dereel | Images for 12 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Old laptops: sending a boy on a man's job?
|
Topic: general, technology | Link here |
So now we have a computer in the kitchen again. Tried using it in earnest today, with less than spectacular success. firefox 3.5 may be faster on normal computers, but on the ThinkPad 380X it's glacial. Even things like paging down eat up several seconds of CPU time.
It's possible that OpenBSD is part of the problem: the system has a 225 MHz CPU, but OpenBSD reports 118 MHz. I suspect this is an issue with APM, but it's what the boot messages report, and I can't change it. I can check all sorts of things, but it looks as if I've chosen too weak a machine. So, once again, back to the drawing board.
Mail from Peter Jeremy and Michael Hughes explaining what I did wrong with the strange plots the other day: it seems that the variable timefmt describes the input format of the data. The display format on the graph is set with set format x. Then it works, as the graph below shows.
But how was I to know that? The documentation for timefmt doesn't say what it's for, and it doesn't refer to set format x or anything else except for set xdata. Only at the very bottom, in an example, does it mention that it relates to reading the data. And there was no error message, and it's still not clear how it could have interpreted numerical timestamps in that manner.
| Topic: general, gardening | Link here |
The weather's been really strange lately, and the 5 day comparison graph I included a couple of days ago has proved very interesting. The Bureau of Meteorology changed its forecast high for Ballarat from 30° to 35°, which would correspond to about 38° here. In fact, the highest temperature we got was 32.2°, and that only just beat the temperature of 32.1° at 01:45. In fact, the temperature started rising shortly before midnight and continued through the night, dropping when the sun came up. And then it hung round the same temperature for most of the morning before dropping steeply during the afternoon (cyan line):
Still, we got a fair amount of rain (6.5 mm if you believe the old-fashioned rain gauge, or 10.8 mm if you believe the weather station), and it looks as if the temperatures will be better in the coming days. Not too early: a number of plants have suffered.
| Wednesday, 13 January 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Databases: Simplicity or corner-cutting?
|
Topic: technology, opinion | Link here |
Strange information from my weather station this morning: lowest outside temperature -0.1°. A quick check of the database showed this to be true. The following display is truncated to fit on the page:
All the other values were correct, including a number of intermediate values, so this can't be a problem with communications between the station and the computer. In any case, I read every page at least twice, until I get values that agree in every detail. So it seems to be something flaky in the station itself.
With SQL it's easy enough to find these errors (these were the only ones of this nature), and the follow-up command update observations set outside_temp = NULL where outside_temp < 0 is just as easy. What's it like with wview? I don't know how to do it. The data storage format is essentially a black box. Others have also suggested something “simpler” than a fully-fledged database for storing the results. But the lack of flexibility soon puts paid to any “simplicity”.
| Topic: technology, general | Link here |
Time to renew my domain (lemis.com) with Gandi, whose web site still drives me to distraction. And I had to authenticate the transaction with Verified by Visa, with which in the past I have nothing but trouble. Today: success! For the first time ever, it remembered my password and let me authenticate. That's still not much of a record.
| Topic: general | Link here |
Into town today to visit Peter O'Connell and discuss investments. He's been very cautious for two years now, probably with good reason—we've weathered the financial crisis pretty well—but he's still just as cautious now.
While in town, looked for a new SIM card for my mobile phone: I have routing problems with the current one—I'm not reachable from the Telstra network—and since it need recharging, it should be just as cheap to replace the SIM. That proved not to be the case, unfortunately. So who do I get to fix the routing problem? How do I find somebody to understand the issue?
Also down to David Chestnut's to pick up the lawnmower. Chris had tried yesterday, but they hadn't been able to find it. When I arrived, they had no difficulty. I asked about yesterday, and the bloke recognized her. Apparently she had asked for the name “Lee” instead of “Greg”. But nobody leaves their surnames at these places.
They wanted $22 for the repair, including $7 for a new spark plug—on a machine that had just been completely overhauled. They backed off on that, but not on the starter cord, so I still needed to pay $15. Hopefully that's the last it will cost me.
| Topic: food and drink, opinion | Link here |
Also in town, found a whole shop full of marked-down books. Browsed through and found a number of cookbooks, so tried to find one which mentioned the word “wombok”. I wasn't successful, of course, but in the process, I found yet another word for Chinese cabbage in two books from Hermes House (apparently a specialty supplier of sale books): “Chinese leaves”. When will this nonsense stop? They're clearly a cabbage, not leaves, and one recipe doesn't show any picture, so the recipe is useless to anybody who doesn't know what Chinese leaves are. I'm sure that applies to almost any greengrocer or supermarket in Australia. And the other book explains (my recollection):
Many people think that Chinese leaves are a kind of lettuce, but in fact it belongs to the Brassica family, and in America it is called Chinese cabbage.
Well, my understanding is that most people in America call it Napa cabbage. The Free Dictionary knows them well enough to redirect to the article on Chinese cabbage, where there is no further mention of the term “Chinese leaves”. Wikibooks probably sums it up best:
Chinese Leaves is an ambiguous term that can mean bok choy or nappa cabbage.
| Thursday, 14 January 2010 | Dereel | Images for 14 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Google reader: a magazine you design
|
Topic: technology, opinion | Link here |
Somebody pointed me at Google reader today: it seems that a number of people have subscribed to my RSS feed from there. Went to take a look. One of the popup ads said something like “Google Reader is like a magazine you design”. OK, that sounds promising. It came up with the entries upside down, of course, and with times in the 12 hour clock. Still, most aggregation services are configurable enough to fix that, so the “magazine [that] I design” should do so too, right? Well, yes, it should. But if it does, I don't know how to set it; the settings don't offer anything so esoteric. Still, given my general distaste of this sort of thing, I should just move on.
|
Browsers for wimpy machines
|
Topic: technology | Link here |
It's clear that cojones, our kitchen computer (don't ask, but the name doesn't mean what you think) is too weak to run firefox 3.5, but that doesn't mean it can't run any web browser. Things like flash are almost certainly too much for it, but at least part of the issue is probably the now-minuscule RAM memory, only 96 MB (6 times the “theoretical maximum, never attainable” I had learnt about in conjunction with the IBM System/360 mainframes). Went looking in the OpenBSD packages collection for something lighter, and came up with three candidates, dillo, kazehakase and midori. Installed two of them before the midday cutoff.
dillo is really pretty bare-bones. To be fair, this is version 2.0p1, while the latest version is 2.1.1. It doesn't have a direct configuration method; you modify the file ~/.dillo/dillorc instead. Tried that out to set the home page, but it was ignored. Still, I could enter the URL manually and access the files. What came then was a surprise:
This proved to be from PHP. The line in question is:
A comparison with other browsers shows that dillo isn't providing this information; possibly there's more. I can live with that, maybe. Rendering photos was also strange. Here my photos for 26 December 2009:
|
||||||||||||||||||||||
It also doesn't seem to support Javascript: the exposure information for my photos didn't appear. After playing around for a while, dillo had a memory footprint of about 14 MB—a far cry from firefox.
midori looks a lot more like mozilla or firefox. It's configurable and straightforward enough to use. It does support Javascript, and it looks more like a normal browser. After much the same playing around as with dillo, it had a memory footprint of 23 MB. In view of the ease of use, I'd certainly prefer it over dillo.
I didn't get kazehakase installed before the midday cutoff; more tomorrow.
| Topic: gardening | Link here |
Over to Chris' place early in the morning to take back her lawn mower, and came back with a somewhat overgrown pot of what I think were Echeveria. Spent some time cleaning it out. Chris got the original pot back, still pretty full, and I have lots of things spread over a number of pots, including one where I had planted the ill-fated Viola tricolor, and also in the ground:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Viola tricolor weren't the only things to feel the heat last week. The Fuchsias are still looking very unhappy, and the Petunias could be better too. I've moved things around in the hope of getting better results; at least the miniature lemon tree is now looking happier, and has more fruit on it than can possibly ripen.
| Topic: photography | Link here |
One of the potential problems I have with my way of processing photos is that I can load them on the computer and then forget to do anything with them. Went looking for them today and found a total of 98 directories with no corresponding web directory. Some are clearly experimental, others are left over from the last big change, and some are photos—mainly of horses—that I had taken and Yvonne had processed on her pages. But even after all of those, I found a number that I had scanned in some time in the past, and which I hadn't processed. Did a couple of them, but this horrible Canon scanner really produced poor results. I suppose I need to do another action with the Epson.
| Friday, 15 January 2010 | Dereel | Images for 15 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Weather station: software configuration
|
Topic: technology | Link here |
Another query for my weather station software today. I could have given him what I have, but there was still this obsolete shared memory stuff there, and in addition the configuration was still all in the C source—including my passwords! Decided to fix that, and ran into a classical chicken-and-egg problem: some of the configuration data specifies how to access the database, including where it is and what user and password to use. Decided to put that on the command line, and have defaults for everything, even the station name (which isn't really that important until you have more than one). That seems to work.
More experimentation with browsers for cojones today, and finally installed kazehakase. Interestingly, the version I installed was 0.5.6, but the web site is clearly out of date; it says that the latest version is 0.5.4. Initial experiments didn't show much difference from midori. I suppose that's the issue: I'd guess that none of them can do flash, but flash is almost certainly too resource-hungry for this machine anyway. And either kazehakase or midori seem to do the job I want. Now if only the machine had a touch pad and a remappable keyboard!
One of the things that I've had continual trouble with is ERC, my IRC browser. It runs under Emacs, and the last version that I have been able to find runs only under Emacs version 21 (current is 22). I've had various problems with it hanging, and had wished to find an updated version. Today with cojones I discovered that it's now part of base Emacs—thus the lack of a more recent version. Unfortunately, it seems that by installing the old version, I've overwritten the new version. To be investigated.
|
Police caught in own speed trap
|
Topic: opinion | Link here |
An amusing item in the news today: Ken Lay, apparently Victoria's top traffic policeman, got caught in a speed trap. He had all sorts of things to say about it, including “Perhaps I was a little fatigued,...”—from driving too slowly, doubtless. But what really happened? He got caught doing 80 km/h in a 70 km/h limit, in Tooborac. I don't know the area, but 70 km/h limits are relatively uncommon. There's one in Dereel along the main road, which is almost deserted. On the other hand, round where we live, only a few metres from several houses, the limit is 100 km/h. I strongly suspect that Ken Lay's speed wasn't in the slightest dangerous. Instead of talking about reducing the “road toll”, they should be thinking about why the limits are what they are, and what the real danger is of driving 10 km/h over the limit.
| Saturday, 16 January 2010 | Dereel | Images for 16 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
HDR photo processing with Hugin
|
Topic: technology, photography | Link here |
Photo day again today. I have one more problem photo: the “lagoon”. It's a classical case for HDR, and two weeks ago I tried one approach, using align_image_stack. Further experiments with align_image_stack showed that it was markedly inferior to other programs in finding control points, so much so as to make it useless for panoramas.
So today I tried other alternatives for the HDR images. Hugin has various menus that offer HDR processing. The first steps in merging HDR images are the same as for building panoramas: identify the relationships between the images by defining control points that represent the same place on each image. In panoramas the control points are in different parts of each image: if you have two images side by side, the control points will be to the right on the left-hand image and the left on the right-hand image. With HDR they should effectively be in the same place on each image. So you should be able to use the same “Assistant” function in Hugin to identify the control points. That works, too. The control points are marked on this comparison of two of the images:
|
||||||||||||||||||||||
But merging the images didn't. It took forever—over an hour—during which time nona grew to about 1.5 GB address space and wrote copious amounts of data to /tmp, causing the root file system to overflow. The results were not what I expected—in fact, nothing at all worth looking at. I'll have to go back and try to find some documentation, then try again, if only to compare the results with align_image_stack. Today did the latter and got some reasonably good results. Here the standard exposure, then the HDR version:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Topic: technology | Link here |
The overflowing root file system was a nuisance. The correct thing to do would be to put the temporary files in /var/tmp, which is on /home and thus has much more space. But how do you do that? Hugin is particularly lacking in configurability, and I didn't find any way of doing so. Move /tmp elsewhere? Not a good idea, given that it's needed in single-user mode. Further investigation showed that there was a lot of junk in there that I could probably remove. And after all, you're supposed to be able to remove files in /tmp. So I removed everything more than 30 days old. Not a good idea. dereel has been up for 50 days, and I removed things that were still needed:
ssh-agent creates directories there for its socket. Remove them and you no longer have any stored keys. In conjunction with the bug that hangs X when I try to open an encrypted connection without a key, this meant that I crashed X.
There's also a socket /tmp/mysql.sock for talking with mysqld. It wasn't immediately apparent that this was an issue, since it stays open for a while, and it took me several hours to discover: I had disappeared from the Wundermap, and further investigation showed that the report program had lost communication with the database several hours earlier.
Restarting the server wasn't that straightforward either: the canonical way to shut down the server is with mysqladmin, which communicates via /tmp/mysql.sock. In the end, killed it, and restarted it. Also had to restart the report program, which then did what I had intended and sent all the intervening reports. That's satisfying; most software simply omits the reports if it or the network is down for any length of time.
| Topic: general, technology | Link here |
I've been concerned about the rainfall reporting of the weather station, but today it showed that there seems to be something wrong with the reporting hardware itself: it didn't rain at all, but the station reported over 7 mm. It's the station and not the software: it appears on the display as well. Cleaned up the database and resent the reports, which Wunderground accepted and ignored. It now claims that I had 7.6 mm rainfall, interesting in itself because it's not a multiple of 0.3 mm; I suspect something has got lost in the translation to and from old units.
Apart from that, the 5 day graph continues to fascinate me. Until I bought the weather station, I had a rather naive impression that the temperature dropped in the night and rose in the day in something approaching a sinusoidal pattern. In fact, it's very different. When the sun's shining, the temperature rises very quickly about an hour after sunrise (probably when the sun rises above the trees on the edge of the lagoon), but on many days the pattern appears to be random. Today the temperature rose particularly quickly, from 13° at 7:15 to 30° at 8:45, but then it dropped a little and stayed there, oscillating between 23° and 28° all day long:
|
Road safety and trees
|
Topic: opinion | Link here |
Somehow road safety has been in the news lately. One of the things that has always puzzled me is why so many people seem to drive into trees. In a recent ABC news broadcast (ABC loves reporting individual accidents) we heard of a woman who was killed when her car hit a tree in the middle of the road. At the time we joked about whether the tree had been planted there, or whether it jumped out in front of her. But today I was told of a news report, now dead [2016] that showed, if not a tree, then a power pole on the road. Maybe it's a variation of what they do in Germany, where they put obstacles in the road to make it more dangerous and thus make people slow down.
| Sunday, 17 January 2010 | Dereel | Images for 17 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Updating Hugin: another attempt
|
Topic: technology, photography | Link here |
My recent attempts to install a newer version of Hugin weren't very successful, but I see that the port has been updated in the meantime, so tried again. I had to replace another port, but after that once again I ran into the same problems with Autopano-sift-C (here the window from last time: it looks almost identical):
|
|
|
||||||||
This time tried to run the program manually. It failed, of course, but it also produced an error message:
=== grog@dereel (/dev/ttyp9) ~/Photos/bloody-stupid-hugin 53 -> autopano-noop.sh --maxmatches 20 --projection 0,87.749934 /tmp/ap_resMqvLRS "/home/grog/Photos/bloody-stupid-hugin/00.jpg" "/home/grog/Photos/bloody-stupid-hugin/01.jpg" "/home/grog/Photos/bloody-stupid-hugin/02.jpg" "/home/grog/Photos/bloody-stupid-hugin/03.jpg" "/home/grog/Photos/bloody-stupid-hugin/04.jpg" "/home/grog/Photos/bloody-stupid-hugin/05.jpg" "/home/grog/Photos/bloody-stupid-hugin/06.jpg" "/home/grog/Photos/bloody-stupid-hugin/07.jpg" "/home/grog/Photos/bloody-stupid-hugin/08.jpg"
That's something, I suppose. But what JPEG library? And why does it want an older version? Looks like a job for ktrace, but I can't be bothered. But the newer versions of hugin offer a range of control point detectors, 5 of them in fact. I've already established that autopano-sift is incompatible with autopano-sift-C, and so it wasn't installed. But apart from Autopano-SIFT-C (every time I see it, they capitalize it differently), Hugin offers Panomatic, Match-n-shift and align_image_stack. Tried them all, with interesting results:
Panomatic does a very good job, arguably better than AutoPaNo-sIft. Here the default output of the difficult verandah panorama, first with autopano-sift (done with the old version of Hugin), then with Panomatic:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The first one has a discontinuity in the rafter in the top middle of the image, but it took a lot of searching to find any errors at all in the stitching of the second one. The most obvious is a discontinuity in the other side of the rim of the round table, and a smaller one in the boards between the two tables, so small that it's almost impossible to see until the image is expanded to full size. Given that, I think I'll stick with that one until proof of the contrary.
align_image_stack was, if anything, even worse than the last time. It wasn't able to find any control points in any of the images I gave it.
In passing, it's interesting to note that Hugin does allow you to configure the temporary directory, and by default it's /var/tmp. It seems that nona just doesn't respect it.
| Topic: gardening | Link here |
Our potatoes are ready for harvest, at least some of them. Pulled these up from just under the surface; I'm sure there are plenty more:
|
|||||||||||||||||||||||
These are Kipfler, which are rather strange in shape, and the one at the back is pretty much exactly the way it should look. But why are most of them so small? Maybe the rest will be better. I seem to remember planting one potato that was very small; maybe this is the one.
Also found that the stem of our Hardenbergia has been growing far beyond what we expected:
|
|||||||||||||||||||||||
Originally it twined around the wire, but it appears that it then straightened out and twisted the wire instead. I think I've found it (and removed the wire) in time. It can hold itself now.
| Topic: food and drink, opinion | Link here |
Watched another programme on TV today, marginally related to cooking: “Gourmet farmer”. It's about a Sydney restaurant critic who retires to Tasmania and becomes a farmer intent on producing good food. It's clearly at least marginally a true story, but I'm left wondering just how much they have changed things in the expectation of making it more interesting.
From my point of view, if the story were completely true, I'd be horrified. The subject of the story is Matthew Evans, and this instalment was about pigs and pork. He bought two piglets with apparently no idea how to look after them, and also a ham to cure, but he also collected a package of offal from the abattoirs—and didn't know what to do with it! It seems that he has very little experience with liver, almost none with heart, and no idea about the other components. So he took it along—without the blood—to somebody to cook it for him. It seems that English cooking culture has forgotten the days when people killed their animals at home and made sausages of the offal—that's the background of black pudding, but this background, the counterpart to the German Schlachtfest, appears to have been forgotten.
As if that wasn't enough, he called his (female) piglets Prosciutto (male) and Cassoulet (because, he says, you can use the entire pig in cassoulet). Clearly he had forgotten the offal already, but it makes you wonder what he thinks the main ingredients are.
In passing, it's interesting to read the comments posted on the web site. You'd think that people were no longer capable of assembling sentences of more than 8 words.
| Monday, 18 January 2010 | Dereel | Images for 18 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Manufacturer's web sites: help or hindrance?
|
Topic: technology | Link here |
We're planning to go to Melbourne tomorrow, and one of the things I need to do is buy replacement parts for the computer that Powercor blew in November. But what? Took a look at the MSY catalogue, which as usual gave almost no information, and spent far too much time looking for information on the web. By far the worst must be LG, who directed me to their German web site and produced me with a flash animation of what they wanted me to see. I couldn't find any way to stop it and get any useful information. What kind of idiot designs this web site? Clearly they think I'm in Germany, but it seems that they do the same in Australia as well.
The other sites (Gigabyte and ASUS) weren't nearly as bad, but it's still not easy to find specifications on the web. It looks, though, as if I can replace the motherboard for $66, the CPU for $65 (the cheapest I could find, but still dual processor) and the memory for $50—a total cost of $181. Things are getting cheaper all the time.
Also decided to buy some more memory for dereel, the machine I bought in October 2008. It has 2 GB of memory, but photo software is extremely memory-hungry, and another 2 GB wouldn't go amiss. But what hardware? I forgot to write it down, I couldn't find the documentation, and I didn't want to take the machine apart to find out. But IJK came to the rescue—the order details are still there, so I now know that I bought:
| Products | ||
| 1 x | WideTech WT-655 Power Supply:550w,12cm fan,SATA connector | $36.00 |
| 2 x | Kingston KVR800D2N6/1G 1GB N6DDRII 800MHz DESKTOP RAM | $50.00 |
| 2 x | Galaxy GXY-9500GT-E2 ( DDR2 ) GeForce 9500GT PCI-E 512MB DDR2 128-bit, NVIDIA SLI-Ready, dual-link DVI, 2560x1600 Resolution | $158.00 |
| 1 x | Gigabyte GA-MA790X-DS4 AM2+, AMD790X, FSB4000, DDR2, 2xPCIE2.0, SATA2, GbLAN, 1394a, ATX | $140.00 |
| 1 x | AMD HD9550WCGHBOX 9550Phenom 9550 AM2+ Quad-Core Processor 2.2Ghz 4MB Cache 95W | $195.00 |
| 1 x | Samsung HD103UJ SATA 1TB 7200rpm 32M Cache 3.0Gbps, | $158.00 |
| 1 x | Samsung HD501LJ SATA 500GB SATA2 300Mbps 7200RPM 16MB Cache | $79.00 |
| 2 x | IJK TR-29 DVI-I_m->VGA Adapter: DVI-I24+5 Pin Dual Link (M) to VGA (F) (for DVI output to VGA display) | $16.00 |
And no, it's not silly to put this kind of thing in my diary. It helps me find things in the future. The only problem is that I copied their HTML, and it's probably buggy. But I was also able to find the manual for the motherboard online. I should save this kind of information more often.
I originally had the URL for the manual here, but it has atrophied in various ways. I can't be bothered checking if they still have it somewhere.
|
Weather stations: hardware problems
|
Topic: technology, general | Link here |
Mail from Steve Woodford today referring to some of the interesting things I've been experiencing with my Fine Offset WH-1081 weather station. He has had trouble with sunshine affecting the temperature readings, which seems like it could also be my issue; but I've measured the same temperatures elsewhere as well. He also mentions that wind and birds can cause the bucket of the rain gauge to tip multiple times. That's a possibility (birds, not wind), but I'll have to look at it, and I wonder if it can have caused the spurious readings the other day when there was no rain at all.
On the other end of the scale, spiders can cause the rain gauge bucket to jam. Again a thing I'll need to keep an eye on.
| Topic: gardening | Link here |
Harvested more potatoes today, the “Dutch Cream”. And as I expected, they looked normal:
|
|||||||||||||||||||||||
Hopefully the other Kipflers will be OK as well.
|
More flash exposure woes
|
Topic: photography | Link here |
I'm still having problems with flash exposures. The photos of the potatoes I mentioned in yesterday's entry are all unsatisfactory. I used the built-in flash and automatic exposure for the first one, but the camera selected 1/60s and f/3.6, with the result that the depth of field was inadequate:
|
|||||||||||||||||||||||
That's not an issue with the strength of the flash unit: it has a guide number of about 11, so at the distance of 73 cm it should have been able to select up to f/16. I then tried with the external Mecablitz and aperture-priority mode, so I was able to force f/11. But then it was underexposed:
|
|||||||||||||||||||||||
That can't possibly be a flash strength issue; the unit has a guide number of about 50 at that focal length. Why is it underexposing? The vignetting at the bottom is presumably because of the close distance, and I could fix that by repositioning the head, but by this time I couldn't be bothered any more. You'd think that modern cameras could guess exposure better.
| Tuesday, 19 January 2010 | Dereel → Melbourne → Dereel | Images for 19 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
ABC: Don't use
Internet Explorer
|
Topic: technology | Link here |
Computer-related matters are increasingly in the news nowadays. This morning in the ABC radio news they reported about the latest bug in Microsoft's “Internet Explorer” web browser. I had understood that the problem is fixed in the latest version, but that's not what the report said. Instead, they recommended “use another browser” (this is the only reference I found on the ABC web site, and it's not quite the same as what I heard).
| Topic: general | Link here |
Off to Melbourne this morning for shopping. First to the Queen Victoria Market. It's becoming increasingly difficult to find the things that we use to cook with; it's been some time since we found any lean cooking ham (somewhat incorrectly called Schinkenspeck in Australia); this is the second time we haven't found any. When (if?) we do, we'll have to stockpile a lot of it.
Also bought some meat. In the butcher's hall they have a number of criers who, in almost mediæval fashion, call out the virtues of their wares, much to Yvonne's disgust. Today they were very subdued. I asked one whether they had run out of steam, and he told me that today was only Tuesday, and that he was keeping his energy for the end of the week.
| Topic: brewing | Link here |
Then down to Grain and Grape, with some wheat malt that had a bad case of the flies I've noticed before:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, the photos are terrible; they're taken through the container, which I don't intend to open anywhere near the house. The ones taken 15 months ago are of reasonable quality. Nik tells me that they're Indian grain flies, and that they're a real pest; it seems they've had their own fights with them. It's interesting that I've only seen them with Powells malt, though Nik isn't certain that it couldn't have happened at Grain and Grape.
| Topic: technology | Link here |
Then on to MSY to buy my computer stuff. It makes sense not to make too detailed lists, though it probably makes even more sense to order in advance. They didn't have my $19 power supplies, and I had to pay $25 for the next model up. And that claims to deliver 680W, which I consider unlikely. They also didn't have the Hitachi 1 TB disk drives for $99, and I had to pay $101 for a Seagate. And it turned out that the $66 motherboards are a micro ATX form factor, and the cheapest suitable ATX motherboard I could get cost $108—that's a significant difference. But I have a number of ATX cases, and buying a new case would have been even more expensive.
Also bought an LG W2243T-PF monitor for Yvonne (I found the link thanks to Google; LG's own web site is completely useless). It's 1920x1080, and the first LCD monitor she's had.
| Topic: general, gardening | Link here |
Then off back home, earlier than usual, so decided to stop off in Werribee to visit the park. Finding it was really confusing: on leaving the freeway, there was a sign pointing straight ahead, just before a junction where you had to turn left or right. After that, nothing obvious, but it proved that a left turn just to the right of that junction did in fact lead there.
I didn't know what to expect, but we ended up in the Victoria State Rose Garden. It's internationally acclaimed, and it has 5,000 rose varieties, but at this time of the year it looks just plain boring, and also badly maintained:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
We were somewhat disappointed. We picked up a brochure at the second information stand (the first one was empty) and got a list of the cultivars—much fewer than 5,000: I assume they mean 5,000 individual bushes (though the web site says “roses”; surely they can't mean the flowers). Decided to look for the “Gruss an Aachen” that we have planted, and which hasn't flowered properly yet. It's there—somewhere—but we didn't find it.
Somewhat disappointed, decided not to visit the rest of the park, and into Werribee to get some lunch, ultimately ending up at a Vietnamese restaurant who offered a lunch special that left us wondering why we bothered. A pie or pasty would have been better. We really shouldn't go to restaurants any more.
| Topic: technology | Link here |
Back home and set up the new monitor for Yvonne. Not the easiest thing in the world: in principle I should just have had to plug it in, but of course it has a different aspect ratio, and a different resolution. Playing around with xorg.conf got a “frequency out of range” message with a horizontal frequency of only 83 kHz. Went looking for the frequency parameters, which I couldn't find anywhere in the excuse for printed documentation they supplied. There's a CD, but how can you look at that without a monitor?
After a bit of log file reading (aren't Xorg.0.log files fun?) discovered that the real issue was that the version of X I was using was too old to know “high definition” formats like 1920x1080 (it did know 1920x1440). Put in a mode line from another system and it worked fine. But it's yet another documentation issue, and most people wouldn't have had the background to debug it.
| Topic: photography, technology | Link here |
Processed the photos I took in Werribee, including the panorama above. The combination of the lower resolution of the Nikon “Coolpix” L1 and the new Panomatic control point detector made panorama creation an order of magnitude faster; possibly it'll be faster for my normal panoramas once I've installed the additional memory in dereel.
| Wednesday, 20 January 2010 | Dereel | Images for 20 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Hardware reconfiguration at a snail's pace
|
Topic: technology | Link here |
So now I have more hardware. What do I do with it? The power supplies were the easiest, of course: just replace them. But then there's this 1 TB disk, and another that's been hanging around unused for nearly a year. Where do I put them? Decided to use the old 1 TB drive for photo backups, and started transferring data across. There were 270 GB of data, so it ran all day. And that's all I did. I'll need to turn dereel off to add a disk and memory, so I need to plan that carefully.
| Topic: general | Link here |
What is it about postmen that makes them want to fold things? I've asked them repeatedly not to, but there seems to be some innate urge to do so. Today it was c't, which frequently includes DVDs: Fortunately, there was none in this one. But what's the point? It's easier to insert it flat. It's also interesting to note that somebody—almost certainly the postman—has bent the right-hand of the cover over the letter slot. It's pretty solid, and it must have taken a fair whack to bend it. And who else comes so close? It obviously happened some time between 17 November 2009, when I took the first photo, and today (second photo):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Farbmalz
|
Topic: brewing | Link here |
I'm finally getting around to my next batch of beer, a Dunkelweizen. I'm planning to use a recipe from Hubert Hanghofer. It calls for Farbmalz, which translates as “coloured malt”. I confirmed yesterday that they didn't have it at Grain and Grape, but in his book Hubert describes how to make it yourself: just roast pale malt in the oven at temperatures between 200° and 230°. There's still the question, of course, how long to roast for. Erred on the side of caution.
|
|||||||||||||||||||||||
It seems that Farbmalz is a “Reinheitsgebot”-conformant replacement for roast barley (which, in fact, is perfectly conformant with the Reinheitsgebot, much more so than wheat malt), and that I could probably have used roast barley instead.
| Thursday, 21 January 2010 | Dereel | Images for 21 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Hardware upgrade
|
Topic: technology | Link here |
As planned, finally did a bit of serious hardware work today: Yvonne got one of the old test boxes, which proved to be double the speed of her current machine. In the past she hasn't needed anything fast, but since she's started doing more photo processing, more speed would be really helpful.
Then the part I dread: adding disk and memory to dereel. In fact, the rebuild went without incident, but only half my new memory was of any use. This is an AMD Phenom machine running in 32 bit mode, and it already had 2 GB memory. You can't use the full 4 GB memory space, because memory-mapped devices take up some of it; but I hadn't been prepared for a total of 1 GB of mapped memory. As a result, I can now only use 3 GB main memory.
Creating a file system on the new disk was also fun. Recent versions of FreeBSD seem to have difficulties with on-line disk reconfiguration; after creating the partitions, devfs couldn't find them until I rebooted. Ended up rebooting three times before all was done.
Then: what to do with the disk? I currently have three disks in the machine, all almost touching—not a good thing for reliability. So I'll move the contents of /home to the new disk and later use it as the system disk. Spent the rest of the day copying the data across with rsync, and only at the end did I discover that I had used the wrong options, and it had ignored links: as a result, the old disk had 377 GB, and the new one had 450 GB. Probably I can start all over again. That in itself isn't such a bad idea; I saw plenty of crud there that doesn't really belong.
| Topic: technology, general | Link here |
Getting the weather software working again was also a bit of a challenge. For the first time, I had a “hole” in the readings, and the program that reports to Wunderground ended up with a SIGSEGV on a null pointer. That was easy enough to fix, but the sources had moved on since then, and I had a fair amount of work to do to bring it up to date. And when I did, a bug occurred that I can't fix. According to the documentation, you can send messages with no real data at all:
- NOT all fields need to be set, the _required_ elements are:
- action
- ID
- PASSWORD
- dateutc
So that's what I did:
It was accepted, but then I discovered an entry (for 14:12) on my station page showing a dew point temperature of -73.3°. I've seen this value before: Wunderground still use the old Fahrenheit system internally, and this value corresponds to -100 °F. I can only assume that its appearance is a bug.
| Topic: general | Link here |
Also finally got around to connecting up the fire hose. Yes, the bore pump is strong enough. Now we just need to find a way to run it when the power fails. Also took advantage of the hose to spray down the outside of the house, which looks pretty dusty—until I found an area of paint that wasn't in the best of condition, and the hose washed it off too, down to the undercoat.
| Topic: general, food and drink | Link here |
Chris along for dinner in the evening, lapin aux pruneaux. Things are changing for her: she's been admitted to university to study IT, so from the month after next horses will take the back seat.
| Friday, 22 January 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Microsoft's revenge
|
Topic: technology | Link here |
One of the places that this diary lands is as an RSS feed to ACM Queue. But for the past few days, no new article of mine has appeared on that site.
Where's the problem? I established that the aggregator has been picking up the file faithfully every hour, and nobody else has reported problems. Looking at the file with the feedvalidator.org RSS validator showed no problem. Finally it occurred to me that the ACM feed is a little special: it has only computer-related topics, and I've introduced a kludge to create a heading that normally wouldn't appear there. So I tried checking the exact URL with the topic limitation. Bingo! One of the strangest messages I've ever seen:
line 2450, column 6: Undefined named entity:
ldquo(2 occurrences)“Internet Explorer”</title>
Who would have thought that you can't use that kind of character in titles? And of course it's because I insist on putting quotes around common words that Microsoft insists on trying to turn into brand names. This time they appear to have got their own back.
Yesterday I had identified a large number of files that needed removing—in my old MySQL directories alone there were nearly 700,000 files, most of which I don't need. First thing this morning I decided to remove some of them. It killed the system.
I should have remembered that; at least in the version of FreeBSD that I run, there's a bug in soft updates that freezes the system if you delete too many files at once. In addition, the situation can occur again during fsck: you need to turn off soft updates first. That took nearly an hour, by which time lost+found had overflowed, and it appeared to contain about 80 GB of data, far more than the 11 GB in the MySQL directories.
Spent the rest of the day cleaning up. I don't know where those files in lost+found came from (well, some were clearly from the MySQL directories), but nothing seemed to be missing (fortunately, apart from the backup, I had a complete copy already on the new disk). Copied the files again, this time using tar, and got better results. But it took up the entire day, and once again I had to postpone my brewing.
| Saturday, 23 January 2010 | Dereel | Images for 23 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
C++: how to obfuscate code
|
Topic: photography, technology | Link here |
Photo day today, and in view of the upcoming experiments with panomatic, took a few more panoramas than normal. As expected, processing was an order of magnitude faster, and the results were at least as good as last time. I only found one very small discontinuity in the verandah panorama, even at full magnification, though admittedly I didn't look as hard as last week.
Clearly it's time to document some of this stuff. Went looking for the documentation for panomatic, and found none. The best documentation I can find is the Panotools wiki, and it's pretty bare-bones. In this case, it just refers to the panomatic link above, and there's no documentation there.
Went looking in the source code. There's something about the way people program in C++ that goes against the grain, but even when I ignore my own prejudices about correct source code layout, it seems that this is sub-optimal. Is the following code really preferable to a loop, and indeed does it need to exist at all? And does it really have to be indented like this (with secondary indentation that differs from one line to the next)?
And why does the spacing around brackets differ from one line to the next?
That last line hides the clue to what the command line options may mean. To me, it's symptomatic of the problems created by excessive use of classes. The arguments refer to the program: there seems to be no reason to have to understand the internals of cmd, a variable which doesn't show up at all in any etags tags. It's created by this constructor:
And that's defined in tclap/CmdLine.h, along with a static parse function. It's written by a different author, as the style clearly indicates: the ^M characters at the end of the lines are missing, and the indentation is marginally different.
After 30 minutes looking at this code, I still don't know what it's doing. The idea of object-oriented programming is supposed to make things easier. What I see here is unbelievably complicated and obfuscated. Why?
| Topic: gardening | Link here |
Lisa Graber along today with a friend from Western Australia, and was quite impressed by the garden. She took a lot of cuttings, including Buddlejas and birches; she's promised to bring us some other Buddlejas.
| Topic: photography, animals | Link here |
The birds are using the bird bath a lot more lately, and today I saw a large number on it at one time. By the time I had set up my camera (in the bathroom), most of them had gone, of course. Only got a few photos:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And, of course, the subjects were far too small for the lens (300 mm, corresponding to 600 mm on a full-frame camera, and with the subject only 9 m away). This is an excellent reason for a longer telephoto lens—if only something useful were available at a reasonable price.
| Topic: animals, general | Link here |
Yvonne is training Dacio, a new horse. She went out for a ride in the forest this afternoon, and the next thing I heard was Helen from across the road calling down the hall to tell me that Dacio had returned without a rider. Off looking for Yvonne—she always tells me where she's going when she's alone—and found her having nearly made it back home. It seems that a bird flew up in front of Dacio, who first jumped sideways, bringing Yvonne out of her equilibrium, and then jumped forward while she was trying to catch up. Fortunately she's not seriously hurt, but she has a number of bruises, including probably her ribs, and I suspect she'll be sore for quite a few days.
| Sunday, 24 January 2010 | Dereel | Images for 24 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Panomatic help
|
Topic: technology | Link here |
More investigation of panomatic, and found that there is a help of sorts: if you start it with no arguments, it prints out a usage:
It seems that there are lots of command line options, most of them long and confusing. Clearly things would be better if I understood the algorithm—but where can I find out about that?
| Topic: brewing | Link here |
Finally got round to brewing my Dunkelweizen today, using the malt that I bought on Tuesday. I had had it crushed there, and it was much coarser than I grind it myself. The results were corresponding: I had expected 50 litres of wort with a gravity of 12.5% (about 6.25 kg sugar), and what I got was 40 litres at 10% (4 kg sugar). After mashing, the grain looked like this:
|
|||||||||||||||||||||||
The grey glutinous areas are unconverted starch. But how can this happen? They know how to crush grain at Grain and Grape. My best bet is that the crush they use works for most heavily modified malts. There have been discussions in the past about poor yields from Powells malts, and investigations at the time showed that there was less Amylase in their malts than in other brands. On the other hand, I had never had any problems with it until today. I had attributed this to the German style multi-step mashes that I did, but that's what I did today as well (and extended the times to be on the safe side). So it looks as if the coarseness of the crush is one aspect of the problem.
|
Flash exposure: more puzzlement
|
Topic: photography | Link here |
How do you take photos of mashed malt? Clearly flash is one possibility. But how? I had a number of options available: no flash, built-in flash, studio flash with or without the tent, and flash on the flash shoe (Mecablitz 58 AF-1 O digital) with or without the ring flash attachment. Tried all the options with the very surprising result that there was almost no difference (details on today's photo page). This was all done with manual exposure (1/250 s at f/11), so this means that the illumination on the subject really was the same. But the studio flash units are completely manual. My best bet is that “manual” on the camera doesn't influence the TTL regulation of the internal or external flash, and that the studio flashes weren't bright enough to overexpose the image, and that in these cases the internal flash compensated for them. Next time I should use a cable to trigger them instead of the internal flash. The images also showed a differing difference between the raw image and JPEG.
| Monday, 25 January 2010 | Dereel | Images for 25 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Australian Internet blackout
|
Topic: technology | Link here |
On the sidelines of the Chinese invasion of Google, it's interesting to note that Google retaliated by no longer supporting China's Internet censorship. Hillary Clinton called for an end of censorship, apparently likening it “to the rise of communist Europe, warning that a new "information curtain" threatened to descend on the world unless action to protect internet freedoms was taken”.
And yes, for pedantics, that last quoted quote is correctly nested. My quotes are “ and ”, while the ones from the Guardian are the standard ASCII ".
But that's what the Australian Government appears still to be intending. Admittedly, Hillary Clinton was referring to the “other side”, including China, Vietnam and Iran. Yes, China is censoring information that the West considers common knowledge, and the Australian censorship purports to be about child pornography. But it's all in secret, and without controls we can't be sure. This matter has been on-going, and last year I commented about it and the reasons why it's stupid. Round the same time somebody leaked the blacklist, and it seemed fairly clear that it was anything but well-sorted. But despite all the evidence that it won't work, and is subject to abuse, they're pressing ahead. This week a number of people, including Electronic Frontiers Australia, launched an Internet Blackout campaign to draw people's attention to the problem. I don't expect many people will pay much attention, and indeed I don't think that the censorship will work, even if the bill is passed, but I'm annoyed by the stupidity of people who introduce bills like this. Nobody expects a cabinet minister to be an expert in all parts of his portfolio, and indeed Stephen Conroy demonstrates that he can hold on to his portfolio with almost no understanding. Even his party colleague Kate Lundy disagrees. And Conroy shows his competence with statements both agreeing and disagreeing with Hillary Clinton's stance against censorship.
So I'm going along with the blackout campaign, for the rest of the week. The first time you access my site, you'll get a popup drawing attention to the problem. It should happen only once.
|
|
||||||||
The popup offers you instructions on how to modify your web pages. Unfortunately, they're broken:
|
|
||||||||
Once again assumptions about text size have truncated the page, this time exactly at the top of the code. It is:
For my own reference only, I've put it in the following files:
| Tuesday, 26 January 2010 | Dereel | Images for 26 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Old computer cases
|
Topic: technology | Link here |
I still haven't assembled the motherboard that I bought last week, though I did discover that a MicroATX motherboard will fit in an ATX case—so I blew $42 for not knowing that. Searched the shed for a case, and finally found one that had suffered somewhat from the storage conditions:
|
|||||||||||||||||||||||
I've also decided to give away my PDP-11, which I've had for nearly 13 years now and never powered on, to somebody whose name is apparently Ted Bullpit, to believe /whois on IRC. As the quote continues, “No! Not the VAX!”, which appears appropriate. While in the shed, finally dug it out of the mess:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
One of the issues were the dimensions of the cabinets. They're 54 cm wide, 76 cm deep and 1.06 m high.
|
Temperature control problems
|
Topic: brewing, technology | Link here |
So now my Dunkelweizen is fermenting happily—or it was. Checked the display on the screen—nice, just on 24°. Only one thing was wrong: date and time. The program had hung about 18 hours ago, while it was cooling. So it carried on cooling, and instead of 24° I really had 11.2°.
That's potentially disastrous, but in this case it wasn't. At that temperature the yeast shuts down, but it doesn't die. Other strains of yeast ferment happily at temperatures down to 7°. All I needed to do was to warm the wort up again, and it continued fermenting. But that would have been a very different story if it had hung while heating. 24° is pretty much the upper limit for an acceptable fermentation, and if it had gone to 30° it would have had a very negative effect on the beer. Only a little warmer would have killed the yeast. How did it happen?
Spent the day observing, and yes, it hung again. I also got one of the old problems with bad contact for the temperature sensors. My best bet is that the serial line somehow hung as a result, something that never happened in all the years with the old machine. I suppose I should put in some watchdog timer function.
| Topic: animals, photography | Link here |
We've established that Lilac, our Burmese cat (13 years old today if I recall correctly) is getting older, and Yvonne has taken to keeping her inside with Piccola overnight. I don't know if that's part of the reason, but I've been noticing increasing bird activity in the bird bath. This morning there were three Crimson Rosellas at one point, followed by what appears to be a Western Ringneck. We've seen one here from time to time, and only one, so maybe he escaped somewhere. Of course, by the time I had got my camera set up, they were gone again, and all I got were the usual suspects. On one occasion a blue male Superb fairy-wren showed up, and I got only one photo—overexposed! It'll take a while before I get some useful photos.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Wednesday, 27 January 2010 | Dereel | Images for 27 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Web censorship is Internet censorship
|
Topic: technology | Link here |
More reactions to the Internet censorship issue today. A cartoon, now unfortunately lost, draws attention to the common misconception that “Internet” means “World-Wide Web”. We've seen this before in the previous government's anti-smut bill; it didn't cover pornographic spam, for example. And another comment seems accurate to me:
Clinton's speech would indeed have caused some uncomfortable moments for the Rudd Government, and has probably done more to cause some second thoughts about the filter plan within the Rudd Cabinet than any of the ineffective campaigning from the net community.
He's probably right about the last part, too; all we can really hope to do with “campaigning” (including a petition to be sent to Canberra) is to draw the eyes of the general public to the nonsense.
Another hang in the temperature control application today. As suspected, it was in reading the temperature probes:
Put calls to alarm around it. I'm not sure that that's enough; that delivers a SIGALRM (note the missing second A) to the process, which then stops. And it doesn't get a chance to reset the relays (can't write from a signal context), so unless something resets it, stopping the program wouldn't help. I've put the whole thing in a loop, so if it stops, it restarts. But is that enough? I'll experiment once this batch of beer is out of the fridge.
Also put together the new machine from the components that I bought last week, and installed Ubuntu 9.10 on it. That wasn't without issues: the motherboard wants an 8 pin CPU voltage connector, and the power supply only has a 4 pin connector. Nothing in the documentation to suggest that it works (it does). Also, it seems that the “power” switch on the front of the cabinet is defective. And that's the standard way to turn the machine on. I was able to start the machine by shorting the motherboard contacts with a screwdriver, thus confirming what the problem was, and configured the BIOS to come back on when power was restored, so I can turn the machine on and off from the switch on the power supply. That doesn't give the soft shutdown, of course, but I never use that.
| Thursday, 28 January 2010 | Dereel | Images for 28 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Digital appliances: “trust me”
|
Topic: general, technology, opinion | Link here |
Into town for a regular health checkup today, during which the doctor decided to measure blood pressure. Cheap digital blood pressure measurement devices (which, Wikipedia tells me, are called a Sphygmomanometers) have been available for decades, but for some reason, all doctors I know do this with an old-fashioned manual device and stethoscope. I asked him why, and he pointed to the digital blood pressure measuring device next to him; he didn't use it because he doesn't trust it. He suspects that it reads the systolic pressure too high, but he doesn't know how it works, and he has done dozens of experiments measuring his own blood pressure. Each one turned out differently, and appeared to be dependent on the way he held his arm.
I didn't know how the things work either, and the Wikipedia page isn't very helpful, though it does state that the readings are derived—only not how. But there's a more profound consideration here: we're continually relying on digital devices without knowing how they work, how accurate they are, nor what factors affect their accuracy.
Another example: conventional scales (weighing machines) measure either as a balance, which is difficult to get wrong, or as a function of a spring, which can weaken with age and thus show excessive weights. Digital scales work with strain gauges. They need temperature compensation, and they can show hysteresis. Very often I weigh something, then remove the pan, and the scale doesn't return to 0 immediately; first it goes to a small (possibly negative) value, then it returns to exactly 0. There's no prize for guessing that this is a reading error which is being compensated for by the firmware.
I've also seen cases where adding ingredients (flour, for example) slowly seems to have no effect. I suspect that this is a variant of the “0 error” code: with less than a certain change, the firmware assumes that the strain gauge is drifting, and “compensates”. The result: it ends up showing less than the correct weight.
Yet another case is my continual fight with photographic flash exposure. How does it work? What causes the exposure discrepancies that I see? I'm still not much closer to understanding that.
Despite all that, digital devices are very useful, and I wouldn't stop using them. But it's worrying how little we understand what we're letting ourselves in for.
| Topic: gardening | Link here |
While in town, dropped in at Formosa Gardens looking for indoor plants. Found a number, though many looked too big—Ficus and Radermachera sinica, both of which in the wild are enormous trees—and also a few others, notably the “Flamingo flower” that caught Yvonne's attention (first image).
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's been nearly 10 months since we bought some succulent cuttings in Melbourne:
|
|||||||||||||||||||||||
From top left to bottom right, they were: Crassula “Springtime”, Haworthia, a dark Sempervivum, another Haworthia (presumably a Haworthia fasciata, though the label didn't say so), a Sedum americanum f Cristata, and another Sempervivum. Their fates have been very different. I went looking for the Crassula, but couldn't find it; the closest I could find was this, but it looks more like an Echeveria:
|
|||||||||||||||||||||||
So maybe the Crassula just died on me. Clearly I should keep better records.
In fact, quite the contrary. On 7 February 2010 I found it: I had planted it in the Japanese garden, and it's looking quite happy.
The first Haworthia hasn't grown much:
|
|||||||||||||||||||||||
Neither has the first Sempervivum (the larger of the two when we bought them), though it's looking quite happy and has produced a number of child florets:
|
|||||||||||||||||||||||
The Sedum never even got started: over a period of about 6 months, it just faded away. The second (smaller) Sempervivum grew like fury and is now about 40 cm high:
|
|||||||||||||||||||||||
We planted the second Haworthia (fasciata) in the Japanese garden, where it gets full sun and little water. It's hardly any larger than when we got it. If it hadn't flowered, I would have thought it was dying. Today I found one on sale and read the instructions: part to full shade, keep moist. Under those circumstances, it's done pretty well where we put it. Pulled it out and planted in a pot:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Hopefully it'll do better in its new environment.
| Topic: food and drink, photography | Link here |
Baked Camembert cheese (well, in fact deep-fried) for dinner today, the first time we've done it this way. I somewhat overdid the cooking:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That was 5 minutes at 175°, roughly what the recipes on the web wanted. Next time I'll try it with 2 minutes.
In the process, yet another exposure problem with flash; the top of the first photo (taken with the ring flash) was darker than the rest. I thought this might be due to the distance, but I wasn't able to get rid of it. The second photo (without ring flash) was much more evenly lit. I wonder if this is due to the gap in the ring at the top of my adaptor.
| Friday, 29 January 2010 | Dereel | Images for 29 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Verified by VISA: Security for morons
|
Topic: technology | Link here |
Years ago I grumbled about the unbelievable lack of security and reliability of the Verified by VISA online credit card authentication scheme. Now, finally, somebody with a bit of influence (well, more than me, anyway) has presented a paper on the subject, mentioning most of my objections (not the lost passwords) and then some. I wonder if it will be enough to draw attention to the incredible gap in understanding between the banks and the crackers.
Phone call from somebody in Sydney today asking if I could repair a circuit board for his Fujitsu air conditioner. It seems that somebody has posted my name on a forum somewhere stating that I'm capable of performing such repairs (far from the truth). I suppose that's what you get when you post photos like these:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Relcating house plants
|
Topic: gardening | Link here |
More garden work. Now that we don't have a dog any more, we literally have a gap in the kitchen where his basket used to be. Yesterday I looked for some plants, and today it was Yvonne's turn. She came back with two of the planters that I saw yesterday, a Philodendron of unknown species, an Anthurium andreanum (for some reason the Wikipedia page has been deleted) and a variegated Schefflera arboricola cultivar “Jacqueline”. Problems: there's not much light where we want to put these plants, and none of these are really low-light plants. In addition, the Philodendron wants different watering from the others. Ended up planting the Philodendron separately where the Spathiphyllum had successfully recovered, and putting the Spathiphyllum in one of the planters along with the Kalanchoe and an unidentified plant which Chris gave us:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The photo was taken with flash and gives a misleading impression of the brightness. I'm quite concerned that the right-hand planter won't get nearly enough light. But it's easy enough to keep an eye on them and move them elsewhere if necessary.
Yvonne also bought a Plumeria (Frangipani), not really an appropriate plant for this climate—they were everywhere in Malaysia—but it seems that they grow well in pots. Hopefully it will do so: replanting showed that it had almost no roots, and anywhere sunny round here is also windy, so ended up tying it up to a couple of stakes:
|
|||||||||||||||||||||||
Lately I've done a lot of thinking about the amount of light that plants need. I established yesterday that my Haworthia was surviving in far too bright and dry an environment. Or was it? Went looking again, and found firstly that a surprising number of people are up in arms about Haworthia fasciata (which is what I assumed my plant is):
Be warned! There are lots of imposters out there. The majority are forms of H. attenuata.
What's in a name? Maybe this is Haworthia attenuata. But that same page claims that this plant requires full sun to light shade, and that it's drought tolerant, something very different from the instructions I saw yesterday. Still, it didn't look very happy where it was, so it was probably the right thing to bring it to a more hospitable environment.
Last spring we went to St Erth and took out a membership. I got a number of plants at the time, and some free seeds afterwards, but I was unable to order anything: their web site is so horribly convoluted that I couldn't even correlate the list from the catalogue with what was on-line. A mail to them specifying what I want wasn't answered for two weeks, and the answer to my question “how do I tell you my credit card number” received the answer “Write it in the email. This is a secure site”. When I explained that that had nothing to do with email, they offered to call me up so I could tell them the number over the phone!. Clearly not a company I can do business with. In addition, so many of the plants I bought didn't do well. I can expect one or two things, but almost nothing thrived. So I haven't renewed my membership.
Another mail order company is Tesselaar, whose domain name tesselaar.net.au suggests that they consider themselves an ISP. There is a tesselaar.com.au, and it's related to gardening, but it seems not to have anything to do with the Tesselaar I'm talking about. Family squabble?
They specialize in bulbs and have been bombarding me with catalogues. The catalogues are nice, and the prices are such that I don't feel much temptation to order anything. But last year I renewed my subscription to Gardening Australia magazine, I was offered a present of 5 “Hugs n kisses” tulips. We planted them in a pot and put it on the verandah. They didn't do well; only four of them germinated, and only 2 flowered. Today, while planting the other plants, I cleaned out the pot with the tulips, and another with mixed tulips and daffodils. There was almost nothing left of the Tesselaar tulips (on the left):
|
|||||||||||||||||||||||
Not a good advertisement.
| Topic: food and drink | Link here |
Wiener Schnitzel for dinner tonight. How do you cook them? There's a surprising amount of misunderstanding (for example, many Germans think they're made from pork). In Vienna they apparently deep fry them, so went looking for details, in the process coming across one of the most irritating web pages I've seen in a long time. It produced a popup that I couldn't get rid of:
|
|
||||||||
In the end I had to look at the HTML source, all 94 kB of it to find the instructions: “The oil is hot enough if it causes bubbles when you put a wooden skewer into it”.
Went looking for details in the instruction manual of our deep fryer. It has four preset temperatures: 175°, 175°, 175° and 190°, though you can set others from 100° to 185° in 5° increments. The accompanying cookbook suggested 180° for most of the recipes. Another case of left hand, right hand?
Decided on 175° and deep fried the Schnitzels. They were acceptable, but I think I prefer the previous method, frying in lard in a frying pan.
| Saturday, 30 January 2010 | Dereel | Images for 30 January 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Coming to terms with Ubuntu
|
Topic: technology | Link here |
More playing around with zaphod, a system name I have recycled for the dual processor machine currently running Ubuntu 9.10. I never cease to be amazed how Microsoft-like people want to make Linux. You'd think that Vive la différence! would be a better motto. From my point of view, it meant that many things I take for granted were missing, notably NFS and ssh. No Emacs either, of course (it does include OpenOffice). I've seen this before, and I found out what to do by grepping through last year's diary files, but clearly I needed to do something better, so updated my Linux HOWTO to include the names of packages that I need to install.
Played around with that for a while, mainly to compare it to FreeBSD. One thing's clear: the serious performance problems I've seen with firefox on FreeBSD are not limited to FreeBSD or my version of firefox. firefox 3.5 renders big images just as abysmally slowly on zaphod (2.8 GHz Athlon II). But maybe it does better with other multimedia, which is why I'm thinking of letting Yvonne try it out. There's certainly a disadvantage that xv isn't available, presumably because of the shareware licensing conditions.
There are other strange things, too. It seems that X wants some extension I've never heard of. Every time I open a window on a different system, I get something like this:
That wouldn't be bad enough—it seems that a lot of modern GUI programs mutter to themselves where they think that nobody will hear them, somewhere behind the root window—but you see it if you start them from an xterm. firefox and gimp are very good at this. But now Emacs starts up with ugly non-standard fonts and an enormous window size:
Now, I'd just love a display with a resolution of 7320×20306, but on my current 1920x1080 display it's rather useless, and I can't even get to the resize button to change things. Where does this come from? If it can't get the screen dimensions (which should be trivial, even without “Generic Event Extension”), it should at least use a sensible default.
| Topic: gardening | Link here |
We're still having problems with our citrus trees. We have five of them: two grapefruit, two limes and a miniature lemon tree. We've only ever harvested two fruit, one from each of the grapefruit trees—and they were on the trees when we bought them. The big problem seems to be the wind.
Our lemon tree is trying, but not getting very far. Today one of the fruits fell off:
|
|||||||||||||||||||||||
I suspect that's a combination of the wind and the hot weather. Hopefully things will get better come autumn.
Chris gave us an indoor plant a few weeks back, and it found its way into the planters that I planted yesterday. But I still don't know what it is, and neither does Chris. For now it's mystery plant 21:
|
|||||||||||||||||||||||
|
Twice in a blue moon
|
Topic: opinion, photography | Link here |
At the beginning of the month, ABC news, amongst others, told me that there was a blue moon last month. Today I got the news again: today there will be a full moon, and it will be much brighter than normal, but not blue. The report must have left many people thinking that the brightness and the “blue” were related.
This time they're right; there was a blue moon today, and for the fun of it I took some photos:
|
|||||||||||||||||||||||
That was 60 seconds exposure at f/4, and also at least as long in postprocessing. And still the picture quality is only barely acceptable.
In passing, it's interesting to note how completely useless their web site search function is. I can't find any reference to this news item with it; the top hit for the search term “blue moon” was a week-old report on the Haiti earthquake, which seemed to have neither term. I don't know why they bother; Google found a number of more relevant references on their site, though it appears that their news broadcasts aren't archived. The best I could find, now expired, was pretty much on the periphery. At least that way they have fewer contradictions on the site. The report I heard was probably related to the one in the Sydney Morning Herald. But they still come up with improbable claims:
Mr Jacob said this was an unusual year because there would be two blue moons, this month and in March, an occurrence that would not happen for another 19 years.
If you've had enough of full moons, February will be without one, a one-in-30 years occurrence.
Clearly if you have a blue moon in January and March, you don't have a full moon in February. And that's what happened 10 years ago—twice in 29 years.
|
Exposure problems: no end in sight
|
Topic: photography | Link here |
One thing about the photos above: in each case, the first attempt was badly exposed. The lemon was overexposed by 6 stops, the mystery plant underexposed by 2 stops, and the verandah photos underexposed by at least 3 stops:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why does this happen? In the case of the lemon, it could be a combination of the stupid insistence of the automatic exposure program to use flash as little as possible, resulting in an aperture of f/2 (!), which is also a very poor choice from the perspective of depth-of-field. I took the final photo in aperture priority mode at f/16.
The photo of the mystery plant was taken shortly afterwards, and was also at f/16. Is this beyond the range of the built-in flash? Possibly. This one could be my fault. And the third one is somewhat unusual because of the extremely low light levels. But you'd still expect it to do better than that.
| Topic: general, food and drink | Link here |
Chris around for dinner, pork fillet in spinach and puff pastry.
| Sunday, 31 January 2010 | Dereel | |
| Top of page | ||
| previous day |
|
Neighbours move out
|
Topic: general, gardening | Link here |
Quiet day. The weather was once again quite hot, hitting 38.7°, and it was very windy, so we didn't feel like doing much. I'm paying more attention to keeping our plants alive during such weather, hopefully with better results: I think I'll need to take down a lot of the hanging baskets and put them in the shade and out of the wind.
The people across the road are moving out—again. This time they were the owners, but they sold the house a while back and they'll be gone in a couple of weeks. They're the third people to live in the house since we moved in. From my point of view, I'm surprised how much money they got for it, which might bode well for our plans to subdivide the property and sell this part. Over today for the remains of the garage sale they had yesterday—the advertisements were at the end of the road, so we hadn't seen them—and came back with some fluorescent lighting.
|
Bushfire info: other opinions
|
Topic: opinion | Link here |
I've ranted on numerous occasions over the last 12 months about the dangerously poor standard of information on bushfires available from the DSE. Now somebody else has noticed: found an article on ABC radio which comes to many of the same conclusions, but also suggests that fire shelters (basements?) would be a cost-effective alternative to the infrastructure they're still trying to get in place. It seems that the current planning allows for “Places of last refuge” where people can flee if their homes are in danger. Where's our nearest one? I don't know: the Golden Plains Shire has none at all. I wonder what it will take for the authorities to get their act together.
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 |
