|
| Greg's diary |
| January 2013 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Tuesday, 1 January 2013 | Dereel | Images for 1 January 2013 |
| Top of page | ||
| next day | ||
| last day |
|
Another bloody year
|
Topic: general, opinion | Link here |
Yet another year has gone by! It's been 50 years since I started keeping a diary. It's depressing how quickly time goes by, even if you're having fun.
|
Towards easier sourdough bread
|
Topic: food and drink | Link here |
A couple of weeks ago Callum Gibson expressed an interest in baking bread with sourdough. The problem: he has a large family and needs to bake about every other day. It currently takes me 2 days to prepare and bake a loaf of bread. I've been doing some thinking about that, and yesterday I tried a different method, tailored to the smaller “non-stick” pan that I bought recently:
Take 100 g of starter (50 g flour, 50 g water) and add exactly 200 g wheat flour and 250 ml water. Stir well, leave 24 hours.
Remove 112 g of the resultant mixture as next starter (50 g flour, 62 g water). We're left with 200 g flour and 238 g water.
Add 400 g rye flour and 206 g water, to give 600 g total flour and 444 g water (74%). Also 10 g salt and 10 g caraway. Knead in a mixer, pour into pan and allow to rise for about 3 hours in an oven at 40°.
Bake for 15 minutes at 220°, then 60 minutes at 190°, spraying with water every 15 minutes.
I was a little concerned about whether the bread would rise sufficiently. That's 1 part starter to 2 parts rye flour, where most recipes want 1:1. But it went just fine. First results suggest that the bread is a little overdone, but that's related to the pan, of course. In any case, a lot less work than previously. I'll experiment further.
|
Braised oxtail
|
Topic: food and drink | Link here |
While in an experimentative mood, tried braising some oxtail that we bought 6 months ago. There are various recipes, but the one that looked the most interesting was on in Stephanie Alexander's Cook's Companion. It includes black olives, which I don't think fit, but it looks good enough if you leave them out. Part of the recipe requires leaving it overnight to be able to remove the fat, so we'll see how that goes.
|
Photographic multishot techniques: review
|
Topic: photography, opinion | Link here |
One of my perks as an O'Reilly author is free access to some books on Safari Books Online. I've grumbled about the interface in the past: in particular, it was very slow, and I didn't have access to the books I wanted to read. Since then two things have changed: I have a network connection with lower latency, and my access conditions have changed. I now have access to Photographic Multishot Techniques: High Dynamic Range, Super-Resolution, Extended Depth of Field, Stitching, the book I was trying to access 2 years ago.
What hasn't changed is the emetic interface. It's a good thing I get access for free: I don't think I'd want to pay money for it. Why don't they do something about it? There's almost no overview at all.
Started reading the book with the overly long title. It's a topic that greatly interests me, and I started a detailed review, which I didn't get finished. But the book's disappointing: once again it is too Photoshop-centric, and the techniques I have read so far are disappointing. Super resolution appears to be a technique to extract the maximum information from multiple copies of a normal resolution images, using techniques such as dithering. I still haven't found out whether the output image actually has a higher resolution, which doesn't say much for the book. But all this would be easier with stitching more detailed images together.
But the “Stitching” chapter is the most disappointing of all:
In this book, we will limit ourselves to describing how to construct simple panoramas containing no more than six source photos, and to other situations in which stitching can be a useful tool.
What use is that? That's like a book on swimming telling you that you won't go beyond the wading pool. Maybe this explains:
While Photoshop is capable of creating 360° cylinder panoramas, it is still not capable of ensuring a smooth transition between the first and last image of a 360° stack.
Apart from the fact that they then decide that newer versions of Photoshop (which they presumably don't have) can handle 360° panoramas, why limit yourself to a suboptimal choice of software? I have been using Hugin since before this book was published, and it doesn't have that problem. And the authors mention Hugin, but don't use it.
There's much more to read, but so far the book is quite disappointing.
|
More macro photography
|
Topic: photography, opinion | Link here |
Took the photos of yesterday's bulbils today. Macros up to 1:2 (i.e. an area of about 35×43 mm) are easy enough with the Zuiko Digital ED 50mm F2.0 Macro:
Beyond that, I have different possibilities. I can use a 10 dioptre close-up lens to come considerably closer with no loss of light:
|
|||||||||||||||||||||||
Or I can use my Pentax SMC Macro-Takumar 50 mm f/4. Although it also only goes to 1:2, I can fit extension tubes, which gives me a total extension of 87 mm (61 with the tubes, 26 from the lens). That gives me a magnification of 2.74:1, or an area of about 7.9 × 6.3 mm. It also shifts the effective aperture from f/22 to v/60, nearly 3 stops. The results are that I can barely use it. Here an exposure at the same settings as above, and the same with the sensitivity set to 3200/36° ISO:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly I'm not getting enough light. I typically use the studio flash with umbrellas:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Removing the umbrellas and turning the flashes around made a big difference:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But somehow the results are still not as sharp as I'd like. I've never managed a really sharp superscalar macro photo. Maybe that's the best you can do with conventional macro photography, but somehow it's not good enough.
| Wednesday, 2 January 2013 | Dereel | Images for 2 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
CFA web site improved?
|
Topic: general, opinion | Link here |
Fire danger season is on us again, and it looks like this year could be a bad one. Took a look at the current bushfire information on the web. Things have changed since my rant a few years back. I don't know why. It's still a game: “Guess whether you have a problem”. The default map shows not only Victoria, but also most of New South Wales and the inhabited areas of South Australia. And there are bigger, less obvious symbols:
|
||||||||||||||||||||||
Fortunately there's a legend below to explain what the particularly funny looking symbols mean:
|
|
||||||||
Well, some of them. One of the more common symbols,
, is
clearly so obvious that it doesn't need an explanation.
The same goes for the use of the map itself. It no longer uses mouseovers; running the cursor over the map does nothing. Clicking on a symbol does do something: it sends off a request to the web server, which even now doesn't respond very quickly. When it does, it gives you more partial information:
|
|
||||||||
I particularly like the scroll bar. It uses JavaScript, and it works (slowly) with Chromium, but not with my firefox. Once you follow one of these links, you can't get back: the back arrow returns you to the same page again. Instead, you have to find the link “Back to Warnings and Incidents”, which brings you back to the view of half Australia.
Somehow they've been working on this vitally important site for years without making any progress. About the only thing I've found is that the distinction between private and public bushfires is gone. Instead, there are just at least two ways to access the site: via the CFA and via the DSE, with slightly different markup.
|
Too hot again for gardening
|
Topic: gardening | Link here |
The weather was warm and windstill, so went around with the weed spray again. It's amazing how much survived the last spraying. Gave them a little more this time, and then inside to escape the flies. In the evening started removing the sweet peas from the garden arch to the north of the house. They've overgrown everything, notably the Mandevilla laxa that we bought in Lambley 2 years ago, and which still hasn't flowered. I had already removed some from the adjacent roses, which weren't looking very happy either—since then, they're coming up with lots of new buds, so I can only assume that the sweet peas were to blame.
Things looked amazingly bare after that. I suppose we'll get used to it, and hopefully the Mandevilla and Solanum laxum will fill things out. But what do I do with the cuttings? I can't put them on the compost heap like that: there must be hundreds of seed pods. I suppose I'll have to pull them all off and boil them.
|
Did you use your credit card?
|
Topic: general, opinion, technology | Link here |
Renewed the registration for lemis.com today, and paid by credit card. Within minutes I had a phone call: “I'm Cindy from ANZ. Did you use your credit card to pay Gandi a while ago?”.
Should I answer? She didn't ask for any confidential details, and the fact that she knew at all suggested that she must have been well informed. But it's this kind of thing that lowers people's security thresholds. The banks should really have a way of authenticating themselves beyond knowledge of transactions (after all, somebody at Gandi could have done it too). But I was curious: why did she want to know? It seems that, for some reason, Gandi is a popular test site for stolen credit cards. The thief registers a domain with them, tries to pay with the stolen card. Whether it works or not, he has tested it at no risk to himself.
|
That's what we need a dog for!
|
Topic: animals, opinion | Link here |
For some time now we've had at least one rabbit in the garden, and I suspect it of much of the plant damage we've had recently. How do you get rid of them? I wanted to let Zhivago chase them, but Yvonne was afraid that he might become uncontrollable, like Nemo was. Today, though, he got the better of her: he found one in the north garden, and by the time Yvonne could do anything, he had killed it and was eating it, from back to front, leaving the ears till last:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Now there's a use for a dog! And we didn't have to feed him, either.
|
Oxtail, part 2
|
Topic: food and drink, opinion | Link here |
So today we ate yesterday's braised oxtail. I didn't need to remove the fat: there was almost none there. I suppose it depends on the condition of the tail you buy. But it really wasn't worth the trouble. Nice and gelatinous—the whole thing had turned to jelly in the fridge—but of course not much meat. Yvonne liked it so little that it looks as if I will have to finish the leftovers alone.
|
First power failure of the new year
|
Topic: general | Link here |
Another brief power failure this evening at 22:35.
| Thursday, 3 January 2013 | Dereel | Images for 3 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Summer with a vengeance
|
Topic: general | Link here |
The last few days have been relatively mild, but today things changed:
A top temperature of 40.9°, 13° higher than the previous days. Stayed in house all day.
|
CFA web site: criminal negligence?
|
Topic: general, opinion | Link here |
The hot weather is set to remain high for several days. Tomorrow's a particular concern, because it should be (slightly) higher than today, but accompanied by strong winds. Looked round the CFA web site again. Nothing!
|
||||||||||||||||||||||
I thought it might be a problem with their silly JavaScript loading, which doesn't give any indication that there's more to come. But I waited for some time, and nothing happened. And things get worse. I wondered which areas would have a total fire ban tomorrow, and of course the page I looked at just listed which districts. “Central West and East Gippsland”. What does that mean? Central “total” fire ban district or Central Gippsland? Typical sloppy formulations. Never mind, on the left of that page there's a link to a Total Fire Ban page:
|
||||||||||||||||||||||
“Sorry, we couldn't find the page you were looking for”. An internal link! To one of the most important pages on the web site! Clicked on their “Feedback” tab, which has improved to the extent that I can now send a message to them rather than to Facebook, and got two out-of-office replies: one from somebody who will be back on 7 January, the other on 14 January. I'm flabbergasted.
|
Curry tree infestation
|
Topic: gardening | Link here |
While watering the plants this morning, noted a number of ants scurrying surprisingly quickly up and down the Murraya koenigii (curry tree):
|
|||||||||||||||||||||||
Gave them some Pyrethrum to think about and took a leaf frond for closer investigation.
|
|||||||||||||||||||||||
The little bumps are clearly scale insects, which explains the ants. They're also clearly visible in the first photo, once you know what to look for. Oil treatment required, once things cool down a bit.
|
Macro experiments, continued
|
Topic: photography | Link here |
The Murraya koenigii leaves were somewhat flatter than Tuesday's bulbils, so took the opportunity to see if I could get better quality macro photos of them. Not really. The following were taken with the Zuiko Digital ED 50mm F2.0 Macro with and without 10 dioptre close-up lens, and with the SMC Macro-Takumar 50 mm f/4 and 61 mm extension tubes:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The real close-up images are not better. Cropping the same section from the previous photos, we have:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Coming even closer, we have:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
The first image is only half the size of the “thumbnail” view, so it's enlarged and fuzzier than it might be. But the second is exactly the thumbnail size, and the third is double the size (“small”). Clearly the third looks worse, even at the same magnification. So it's fairly clear that the Olympus lens with the close-up lens does better than the Pentax lens at twice the magnification. So there's not really much point keeping it. And I'm still no closer to a good extreme macro solution.
| Friday, 4 January 2013 | Dereel | Images for 4 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More hot weather
|
Topic: general | Link here |
As promised, today was even hotter than yesterday—a high of 43.2°, one of only 3 days since I've been keeping records that the temperature has exceeded 43°.
Revisiting this entry 10 years later, I'm puzzled. Of course I hadn't forgotten the 45° or even 46° on Black Saturday (7 February 2009). My guess is that by “keeping records” that I'm referring to the weather station software that I started in September 2009.
Interestingly, the other two dates are also at very much this time of year: 2 January 2012 (45.4°) and 11 January 2010 (44.9°). Interestingly, the maximum temperature for 2011 was “only” 41.9° (on 30 January 2011). So we spent another day mainly inside pondering the meaning of “Activate your bushfire plan”.
|
CFA web site: even more catastrophic
|
Topic: general, opinion | Link here |
As a result of the weather, spent considerable time monitoring the CFA web site. Things were even worse: it just didn't refresh. It seems that the site was completely overloaded, although (as it turned out) there wasn't too much in the way of bushfire activity. Fortunately the via the DSE site didn't suffer from that problem.
|
A computer for Microsoft
|
Topic: technology, opinion | Link here |
I had intended to change over dereel (test computer) and cvr2 (TV receiver) and then using the ex-cvr2 to run Microsoft for DxO Optics “Pro”. That idea failed for no better reason than that the tuners wouldn't both fit in the motherboard. But then it occurred to me: this would require me to buy a new version of Microsoft (“Windows” 7 or 8—people recommend 7 because of the draconian licenses of 8) would cost me about $100. For that price I can buy a used computer with Microsoft on it—can't I? Did some enquiries and found that yes, I can. The cheapest machines on the market seem to come with Intel Core 2, 2 GB RAM and “Windows Vista”, and they start round $80. Did some enquiry, including discussion on IRC, and discovered:
Everybody hates “Vista”, but then I hate all Microsoft. Nobody could give me a good reason why I should use “Windows” 7 instead of “Vista”.
None of the machines on the market had 64 bit versions of “Vista”, though they did have the “Business” or “Professional” versions, which I need to support Remote Desktop.
They all use DDR-2 RAM, of which I have plenty lying around.
The CPUs were things like 2.66GHz T8300, which CPU Benchmark gives a rating of 1485 at 2.4 GHz, or a 2.33 GHz E6550, with a rating of 1465. By contrast, my current “big” machine, with an AMD Phenom 9550, gets 2455. So per CPU they're actually faster. More to the point, though, that should be good enough for what I want to do.
All 32 bit versions of “Windows” have the same process size limits: 2 GB, expandable to 3 GB with the IMAGE_FILE_LARGE_ADDRESS_AWARE linker option. I wonder if DxO does that. In any case, it seems that DxO runs two processes happily in 2 GB of memory, so with 4 GB that should be enough. If not, presumably an upgrade to “Windows” 7 would be cheaper than a new version. And if all else fails, I can still run FreeBSD on it.
So I bought a machine for $99, plus $29 for postage. It's just down the road in Clayton, and I could pick it up. But that's just more procrastination, and I'm already a master procrastinator, so I'll have it sent.
|
A triplet of scales
|
Topic: food and drink, general, opinion | Link here |
I'm thoroughly sick of changing batteries—all 6 of them—in my digital kitchen scales. It has a jack for a power supply, presumably 9 V (no indication on the scales themselves), so I can probably find one lying around and try it out. But what happens if I fry the electronics? Not a serious loss: it's old and clunky. But I need scales so often that it would be good to have a replacement on hand, just in case.
And as if answering an unspoken prayer, ALDI had kitchen scales on special this week, so I asked Yvonne to bring me one back when she went shopping. She came back with not one, but three.
The scales I asked for have a great advantage over the last ones I tried from ALDI: they show the correct weight. They're certainly a lot slicker than my current scales:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But they have the same other problem as the ones we bought last September. They're too small. Put a bowl on top and all you can see are the completely unnecessary guesswork information about calories and things:
|
|||||||||||||||||||||||
Add to that the apparent lack of a hold function, and a far-too-short timeout before switching off, and the things are particularly irritating to use. So I suppose they'll go back.
For some reason Yvonne wanted new bathroom scales. She claims that the old ones (now about 25 years old, in fact) show different weights every time, something that I can't confirm. So she brought a new set:
|
|||||||||||||||||||||||
That black mark in the bottom middle reads “Slippery when wet”, presumably the reason our
old scales have webbing that is not slippery when wet. So why do they supply the
scales with a slippery surface? It also wants to make guesses about my state of health.
The instructions refer to some buttons somewhere (it doesn't say where) that you press to
configure the thing. Where? Right in front of your nosetoes, conveniently so
close in colour to the background that you can barely see them:
|
|||||||||||||||||||||||
In addition, the thing turns on whenever I walk past it, and like the scales we bought 2 years ago, it shows my weight as 2 kg more than the old scales do. That really suggests that our old scales are wrong, but Yvonne finds that depressing enough that she wants to return them. I'm not sure yet.
The third set of scales she brought back was astounding:
|
|||||||||||||||||||||||
Clearly scales for people who'd much rather use spoons. But it made sense to at least try them: the small measuring container I use is clearly not dishwasher proof:
|
|||||||||||||||||||||||
Yvonne had bought a “replacement” some time ago, but it's so baroque that I don't want to use it:
|
|||||||||||||||||||||||
Looking more carefully at the spoon scales, it has three different attachments, two like a spoon and the third of more conventional form:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Of course, it wouldn't do to show a scale in millilitres. The bigger spoon (shown above) is marked, from bottom to top, “1 2 1 1½”, with explanatory notes to the right: Tsp and Tbsp, which the instructions conveniently explain are teaspoons and tablespoons. And the other container is marked from 1 to 7 Tab, which isn't explained:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Experiments with the latter showed that they're British tablespoons, 15 ml and thus 25% smaller than the Australian “metric” tablespoons of 20 ml. They seem to be quite accurate: I measured out 7 “Tab” of water and ended up with a weight of 104.3g, as close as I could expect to 105 g.
So: is it worth it? I already have some scales for small weights, 50 g in increments of
0.01 g. These are 300 g in increments of 0.1 g. But the cylindrical container proves to be
quite useful, and for the first time I was able to confirm the specific gravity of our
cooking oil:
60 ml 4 “tab” of oil weigh 52 g, a specific gravity of 0.87. So, despite the
strange appearance, this one looks like it's worth it.
|
Poor ABC TV reception
|
Topic: multimedia, opinion | Link here |
I've been keeping my page on TV reception quality for over 18 months now, and it has given me some useful insights. First, it's fairly clear now that reception problems are not related to the cabling. And they tend to be associated with particular channels and particular times: as a result, there have been weekly series that have had consistently poor reception quality. Potentially that's an indication of interference at those times. And more general problems come and go, suggesting that the technicians have repaired some problem. Based on experience reporting faults, I thus tend to just wait it out.
But I've been having trouble with ABC TV for months now, so finally I got round to calling up on 1300 13 9994. No, there's nothing wrong with the transmitter: that particular transmitter services 300,000 viewers. Is there something wrong with your equipment? Try re-scanning the frequencies. Yes, I suppose I have to accept that, so I went and rescanned, of course with no change. Called back, got the same person (Elaine), who said that if nobody else complained, there wasn't much that she could do. She said, though, that we live in a rural reception area (“red area”) with marginal coverage, and that I might be eligible for VAST, a new digital satellite connection. She pointed me to the Digital Switchover Taskforce on 1 800 201 013, where I spoke to Matthew, who pointed me further at http://mysattv.com.au and http://digitalready.gov.au, also giving the phone number 1300 993 376. Went looking at the second URL, where I finally found the relevant page, but it told me to find a mySwitch, whatever that is, without linking to it. Finally found it, entered my address and discovered that we have Good Coverage. Clearly there's a disconnect here. But, as I feared, lots of talking and no results. It'll be interesting to see if things change after my call to ABC.
About the only thing of interest was the details of the transmitter: they don't give the name, but it's clearly the one to the west of Ballarat, which I think is called Lookout Hill, 73.6 km away with a bearing of 323.5°, and they even show a map to give an idea where to point the antenna, which appears to match the direction it's pointing.
| Saturday, 5 January 2013 | Dereel | Images for 5 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden changing
|
Topic: general, gardening, photography | Link here |
The weather was cooler today, and there was little wind: good conditions for the weekly house photos. The first Saturday of the months is also when I do some extra photos, including that of the cathedral. But that's changing so fast that I've had to find a new viewpoint, about 8 m further west. Here last month and this month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's not so different, though, when you consider what this looked like 5 years ago. The following view corresponds roughly to this crop from the right of today's panorama above, overlapping the edge:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Still, the birches are growing well, some already 3 m high. One day it will look nice again.
|
Irrigation hell
|
Topic: gardening | Link here |
After the recent hot weather, a couple of plants are looking quite unhappy. Went around looking at the dripper lines, with disappointing results: a large number of the drippers are completely clogged, and even the holes I made in the dripper line last week had already clogged up. And now the Hebes alongside, which I planted in July were looking very unhappy: their drippers had clogged up too, as had those in the Japanese garden (second image). The leaves are looking dull and sick:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Strangely, the only ones that were looking happy were the ones that I planted last, in August, and which don't have any drippers at all. They're flowering already:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But I suppose that they'll need something soon too.
|
CFA website backlash
|
Topic: general, opinion | Link here |
Surprise, surprise! Other people have been complaining about the CFA web site too. Not about the content, just the poor performance, for which there's a workaround. And still, this thing is really bad. I heard on the news this morning that Victoria had been spared any major bushfire, unlike Tasmania, where there were serious bushfires around the Hobart area, involving much loss of property and evacuation by sea. I wouldn't have expected that down there.
In Victoria, there was just one in the “extreme south-west”. So I went looking on the CFA/DSE map:
|
||||||||||||||||||||||
What, was the fire out already? No, it was still reported in the list below:
|
||||||||||||||||||||||
So where was it? Searched for Nelson in Google Maps, and discovered where it was: really in the extreme south-west of the state. And enlarging the map shows:
|
||||||||||||||||||||||
|
||||||||||||||||||||||
So if you enlarge the map sufficiently, it will deign to show the fire, with an icon smaller than any of the others! What use is that? All these “advice” symbols were covering it. And what good are they, anyway?
|
|
||||||||
OK, then “Read full message”. Problem: I had already clicked on it some time prior, and I left it to do its thing. Nothing happened. Wouldn't you think that they'd put more useful information on the popup than the date and two copies of the address? Hopefully the public backlash will get people to consider what information to provide, and not just how fast.
| Sunday, 6 January 2013 | Dereel | Images for 6 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Thoughts on garden flowers
|
Topic: gardening | Link here |
Now that things are a little cooler again—only 30° today—did a bit more looking round the garden. I've decided that Gladiolus are not the easiest flowers to handle. They don't bloom for very long: the ones in the centre of the garden, despite being planted at different times, all flowered and withered within 3 weeks. And the ones further south, though still flowering, show that they really need some kind of support to prevent them from falling over:
|
|||||||||||||||||||||||
On the other hand, the Hesperaloe parviflora is doing very nicely and standing nearly 2 m high. The flowers are more intricate than I had expected:
|
|||||||||||||||||||||||
|
More work on system upgrades
|
Topic: technology | Link here |
Once again I've been dragging my heels on my system upgrade method. In principle amd64-stable now contains all the ports I asked for and a relatively recent version of FreeBSD 9-STABLE, but I still need to customize it, and then I'll be in a position for the first upgrade. Spent some time customizing /etc/group and /etc/master.passwd; the latter contains a lot of history, user IDs and passwords from people who must at at least one time in the last 20 years have accessed the machine. Should I remove them? It's somewhat nostalgic to have the IDs there, and they won't really do much harm.
But doing those updates was more complicated than I thought: for some reason, running Emacs on a remote X server caused the server to hang. Tried on two different machines, so it's not the server. That used to work. I wonder what went wrong there. I'll have to check that after the upgrade. For the time being, did things in terminal mode.
|
eBay conflict resolution
|
Topic: opinion, technology | Link here |
Last month I bought some horse accessories for Yvonne on eBay. They came from the USA, and the shipping cost an arm and a leg: $46.00. But that's (pretty much) what USPS Priority Mail International costs, so grudgingly I coughed up.
The packet took a while to arrive. When it did, the reason was obvious: it was sent with First Class Mail, which only costs half the cost. OK, no worries. Contacted the seller and asked for a refund of the difference. But he (she?) came up with some cock-and-bull story about having to pay people to package the goods and take them to the post office, 20 miles away. That's clearly incorrect, since the shipping inside the USA didn't include that kind of markup. But he (she?) was only prepared to refund $12, rather than the $23.31 real difference. So I opened a “case” with eBay, and promptly managed to put in the wrong sum that I wanted back. Never mind, there must be some way to alter that. The mail I got from eBay was helpful:
eBay sent this message to you. Your registered name is included to show this message originated from eBay. Learn more. Your case is open in the Resolution Center
...
Remember, you can always check the status of your case in the Resolution Center.
Conveniently, they didn't include a URL for the Resolution Center. So I went looking under Customer Support. Nothing obvious. Tried “Help”:
|
|
||||||||
What a loser! Of course, Google knew the answer. But why can't eBay make it more accessible?
The good news was that the result was immediate: the seller refunded the full amount immediately with no further comment, though PayPal kept $0.11 without any explanation. So I closed the case, showing a race condition in eBay's site:
|
|
||||||||
Like Schrödinger's cat, the case was both alive and dead.
| Monday, 7 January 2013 | Dereel | Images for 7 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Network problems recur
|
Topic: technology, opinion | Link here |
It's been nearly a week since my last network interruption, a relatively long time for my recent experience. And I didn't get disconnected today either—I just fell back to GPRS and stayed there. Discovered that I can force a reconnect without stopping the PPP connection by disconnecting the antenna, and sure enough, it returned to UMTS mode—for a few seconds:
By observation, the 3-digit IDs at the end of those reports are for GPRS cells, while the 4 and 6 digit ones are for UTMS/HSPA-capable cells. This looks very much like the connection couldn't be sustained: the lines without any cell are exactly that, no association.
Restarted the ppp process. No change. Popped the modem and restarted, and I was connected with HSPA again. But why? I monitored the signal strengths for a while and noted they were the lowest I've seen for a long time, as low as -107 dBm (reported at 2). Normally it's round the -97 to -101 dBm, so maybe the signals were just too weak to maintain the connection. Why? The heat? Today was “only” 39.3°, and we had much higher at the end of last week. Interference? How can you tell?
|
Catastrophic TV reception
|
Topic: multimedia, opinion, technology | Link here |
In mid-afternoon I record Al Jazeera news and then watch it. Today there was a problem: the normal recording is 2.1 GB. Today I got 48 MB. And of course there was no content. Previously I have had trouble on ABC, but this was SBS, normally not a problem. It also wasn't a problem of SBS getting as bad as ABC: ABC was so bad that I had no reception whatever.
Did some further investigation and discovered that I could barely receive anything at all. Rebooted the machine, which of course didn't help, but in the process discovered that the tuners were running very hot. They're in a cupboard, and there's normally not much ventilation. So I turned it off for an hour or so and left the cupboard doors open, after which things looked better, but still not good. I wonder if the “masthead” antenna in the ceiling is failing. Or is it related to the Internet connection? Is there some source of strong interference around here?
|
Virtual Machine pain
|
Topic: technology | Link here |
Continued with my installation of a reference “new” machine under VirtualBox. It kept hanging, and after a while I got panics related to disk space allocation. What went wrong there? I see I'm now using journaled soft updates. Is there some issue there where the file system is left in a broken state after recovery?
Went back to an earlier snapshot—much earlier, as it turned out: it didn't have any ports installed. That ran, so I decided to bring userland and kernel up to date, after which—hopefully—I'll be able to reinstall them on the newer snapshot and continue from there. That took all day, of course.
|
Investigating DxO performance
|
Topic: photography, technology, opinion | Link here |
My “new” Microsoft machine should be here shortly, so spent a bit of time investigating why DxO Optics “Pro” runs so slowly under VirtualBox, discussing with others on IRC. I've recently been noting the processing time estimates that DxO makes, and comparing them with actual elapsed time:
| Camera | Estimate | Actual | Ratio | |||
| Canon | 5:33 | 28:39 | 5.16 | |||
| Nikon | 1:16 | 3:39 | 2.07 | |||
| Olympus | 1:46 | 1:45 | 1.00 | |||
This was done without restarting DxO. It seems that it adapts its prediction based on prior experience. The last entry, with the Olympus, was a single photo. So I'd be more inclined to believe the timing for the Canon.
One of the things that is really horribly slow is just reading the working directory to find files. At least, that's the way it is under SMB. So fired up Wireshark to see what was going over there.
I didn't actually get round to looking at reading in the directories, just writing the result files (I think). What I saw blew my mind:
|
||||||||||||||||||||||
Screen after screen of writing a single byte to a file and then asking for information about the file! No wonder it's so slow. But I'm told that one-byte writes are common in the Microsoft environment. I'm horrified. But it confirms my suspicion that the inefficiency of DxO is at last partially bad programming rather than the complexity of the algorithms.
| Tuesday, 8 January 2013 | Dereel | Images for 8 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More build machine issues
|
Topic: technology | Link here |
Into the office this morning to find my world and kernel build complete, so shut down, restarted the crashed version of the VM, installed and booted the new kernel, and tried to install the new world. No go:
I've seen that a couple of times before, always related to this kind of not-quite-normal installation. The last time I did a little research, which suggested that it only got run if sys/param.h was newer than osreldate.h. Nope:
Investigated further and found the rule in /usr/src/include/Makefile:
So it's true that the file depends on ../sys/sys/param.h, but in this case that wasn't the cause. It was newvers.sh itself:
Still, why did that go wrong? Looking in newvers.sh I find:
Why should that fail? No idea. But I found a workaround: touch osreldate.h. And then, of course, things succeeded. I really should go back and find out what's wrong, but I have other things to do.
So: did the new kernel cure the crashes? Started building yet another world with it. No go. It hung again. I suspect it might be unrecognized file system damage. So I backed up the entire file system, most importantly for the ports, fired up the old system and restored the file system. Then another buildworld, this time successfully. So it really looks as if there's some damage in the file that is causing the problem. Do I want to look? No, I don't think so.
|
Public Records Office Victoria wiki
|
Topic: general, technology, opinion | Link here |
One of the mailing lists that I am on receives regular informative mail messages from the Public Records Office Victoria. They continually release sensitive documents whose protection has expired, sometimes after as long as 99 years. My family has been in Victoria for 120 years, so I thought it might be interesting to take a look. To my surprise, they're running a wiki. It looked as if I had to sign up to have any access (misassumption on my part), so I tried that:
|
||||||||||||||||||||||
Bloody Captchas! But what's the error message? It appears to be trying to tell me “We don't like your password because of the content” (“Don't spam me”). Despite the indications, it refused to let me use that password. With another one I was successful:
|
||||||||||||||||||||||
Wiki users will recognize a problem there: Main Page is pink, meaning that it doesn't exist. Following the link brings you to an amazing discovery:
|
||||||||||||||||||||||
The page was deleted due to vandalism! And yes, the main page is no longer called Main Page, it's called PROV_Wiki_-_Home. What an ugly name, and what a violation of POLA!
But things in there are interesting. Found some official documents relating to Eureka Stockade, including this report by Charles Hotham, Lieutenant Governor of Victoria, about a “serious riot and collision at the Ballaarat Gold Field”, dated 20 December 1854. That's some of the most reliable information on the incident yet—and it had a typo! Not “Ballaraat”; that was the spelling at the time. But the title originally contained “report on a a serious riot”. Time to fix things. And I haven't even read the report yet.
|
More drippers, still not enough
|
Topic: gardening | Link here |
The Hebes in the north garden are looking unhappier and unhappier. Today the weather was a little cooler, so went in and added more drippers. It doesn't seem to be worth removing the old ones, since they're clogged, and the whole dripper line has been covered over by leaf mulch and such. I still haven't decided what kind of dripper I should be using, but it's becoming clear that they should be on short 10 to 15 cm lines, so that I can at least find them when I go looking. But what a pain these things are. And there are many more to do and to check—the birches in the cathedral are also showing signs of stress.
|
Free Adobe Photoshop
|
Topic: photography, technology, opinion | Link here |
I've heard from many different places that Adobe has released Photoshop CS2 for free. Just in time to try it with my new Microsoft box. Followed the link, but for some reason I couldn't access it. Then I found a statement from an Adobe employee:
On behalf of Adobe Systems Incorporated ...
You have heard wrong! Adobe is absolutely not providing free copies of CS2!
What is true is that Adobe is terminating the activation servers for CS2 and that for existing licensed users of CS2 who need to reinstall their software, copies of CS2 that don't require activation but do require valid serial numbers are available. (Special serial numbers are provided on the page for each product download.) See http://forums.adobe.com/thread/1114930.
You are only legally entitled to download and install with that serial number if you have a valid license to the product!
Well, what is there to make of that? Adobe puts up a download page with no caveats, no restrictions, and (presumably valid) serial numbers. If the forum post is correct, wouldn't there be something on those pages? This looks to me like the poster on the forums is just plain wrong. And you can't exactly see a statement on a forum as official Adobe policy.
In any case, when I returned the page was active again, so I started downloading. I didn't realize how big this thing is. A couple of gigabytes of data in several files; hopefully it'll be worth the effort.
|
New extension cards
|
Topic: technology | Link here |
I established some time ago that the new Ethernet card I bought for dereel needed to have the on-board (and defunct) chipset disabled before it would work. And then I discovered that I couldn't use the motherboard for cvr2 because of component placement. But I forgot to leave the Ethernet board in dereel. In the meantime I have also received a USB 3.0 adapter, so put them both in dereel to see what would happen.
Not a success. The Ethernet board still doesn't work under FreeBSD. It can now transmit, but it still doesn't receive. And the system doesn't want to know about the USB card at all:
This is FreeBSD 8.2, so it's quite possible that the latest and greatest (9.1) knows it, but first I need to finish my system update procedures.
| Wednesday, 9 January 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New Microsoft box
|
Topic: technology, opinion | Link here |
My “new” Microsoft box arrived today, the very first machine I have ever bought to run Microsoft “Windows”. Nice looking little Lenovo “ThinkCentre”, much smaller than I had expected, but showing signs of less-than-careful treatment in the deep scratches on the lid and the faint smell of tobacco from inside.
Setting up Microsoft boxen has always been a pain, but today just about everything Just Worked. The machine comes apart nicely without tools: took the disk out to back it up, then added another 2 GB of memory I have lying around. I could put in even more, but now it has 4 GB, and the 32 bit system can't handle more than that. Setting up remote desktop was no problem, and even network shares worked—without a password! That looks like an error in my SMB configuration, but in this environment it doesn't make much difference. Installing DxO went almost without a hitch—it wanted to connect to DxO to confirm the (still) trial installation, but once I enabled that it worked.
About the only problem I had was that occasionally the display settings got mixed up, and I ended up with a signal out of range. It seems that this is at least a known problem, as pages like this one explain. But why? Vista isn't that old. It should read the EDID information. It seems to happen at random: on first boot it didn't happen, and it happened on a KVM after being switched away from and then back again.
That's irritating, but not overly serious, since I intend to use it only with remote desktop. Still, it's nice to know how to shut the machine down remotely, and Andy Farkas gave me some significant help on IRC. It seems that there's a command called shutdown. These three invocations have the same meaning:
| BSD | shutdown -p 5 | |
| Linux | shutdown -P 5 | |
| Microsoft | shutdown -s -t 5 |
Yes, Microsoft says you should say shutdown /s /t 5, but it still understands UNIX conventions. Somehow things are less painful than they have been.
|
First DxO run on new machine
|
Topic: technology, photography | Link here |
So, the big question: how fast is the new machine? I've already established three problem areas with DxO Optics “Pro”: running under a virtual machine, running with SMB shares, and running on a non-physical display. Time to time each of them. Started a long conversion run using the photos I took last Saturday, all 188 of them, with rdesktop and the photos on an SMB-mounted file system.
Observing the conversion showed a number of things:
The extreme file system behaviour I saw a couple of days ago is not typical. There wasn't much network traffic at all.
DxO claim that 2 GB of memory is marginal, which is one of the reasons why I increased memory to 4 GB. But although there were two conversion processes, the Task Manager showed total physical memory use between 1.5 and 1.8 GB. It seems that DxO have never tested their assertions.
The display issues were evident. In particular, the display refresh that frequently failed under a VM was still there, and probably worse. This image was taken after iconifying the task manager, so that DxO was the highest window:
|
||||||||||||||||||||||
It seems not to refresh the main part of the “Customize” window. Apart from the remains of the task manager, there are also remains of the “Process” window immediately above. There must be a lot of sloppy programming in that area.
It's still glacially slow! It estimated a conversion time of 24:05, which would be one image every 7.7 seconds. In fact it took 2:50:46, one every 55 seconds, more than 7 times as long as the estimate, and a time comparable with the VM approach.
So: why is it still so slow? This is a slower processor, of course, but DxO must have had some basis for its estimates. Tomorrow I'll try local display and local file system.
| Thursday, 10 January 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Completing the Microsoft install
|
Topic: technology | Link here |
In principle everything works on my new Microsoft box, which in a break with tradition I have called dxo.lemis.com (previous machine names were decidedly unflattering). In practice, though, I still have this issue with the display. I had these instructions to go by. They asked for safe mode, presumably because you couldn't see things otherwise. I expect they should work in normal mode too if you have a display.
I had a display, but it was via rdesktop, and it only showed me details of the simulated display. So I had to reboot after all. What happened was rather unexpected: the machine came up with a GRUB boot menu—for “Windows” 7 only! And whatever I tried, merrily pounding the F8 key, it came up in normal mode.
More discussion on IRC. Is there a file boot.ini? Not on this machine. Finally Andy Farkas pointed me at msconfig, something like nextboot. But we weren't sure about the options: one of them was “Base video”, and we weren't sure whether I should set it or not. My concern, of course, was that I'd end up with the machine in safe mode with a scrambled display, and have no way to get out of the situation.
Prepared to shut down anyway, and turned the KVM back to dxo—and the display was OK! So I was able to enter the configuration menus after all—and there was nothing there to modify! The only thing I could set was the refresh rate, and that only to the values 59 Hz (currently set) and 60 Hz, which sounded more normal. Did that, and didn't have any more problems in the course of the day. But clearly something's very funny here, and I don't think I'm out of the woods yet. Possibly the fact that it only shows “generic monitor” is relevant: is there something wrong with the display adapter or the driver that stops interpreting the EDID data?
|
DxO processing speed: further investigations
|
Topic: photography, technology, opinion | Link here |
As planned, continued my experiments processing photos with DxO Optics “Pro”. There were a number of issues that might affect the processing speed. Clearly DxO is very inefficient in a number of areas, and the display in use (local or remote) and the file system in use (local or remote) were the most obvious ones. Yesterday I had tried with both remote, and today I tried the other combinations. Here the results:
| File system | Display | Time | ||
| remote | remote | 2:50:46 | ||
| local | local | 2:15:18 | ||
| remote | local | 2:13:49 | ||
| local | remote | 2:06:15 | ||
| remote | remote | 2:10:42 | ||
The first run was the one I did yesterday, and it seems I (or something Microsoft) was also doing something else: the time is much longer, but not repeatable. In fact, given the fact that I was doing various things with the machine, there's no obvious difference in the processing speed between local and remote. The fact is that DxO really is glacially slow. If I had known that I might have bought a faster machine, but then again, maybe I wouldn't have. In any case, I paid $99 for this machine with an operating system. My free trial of “Windows” 8 expires on Tuesday, so at the very least I had to buy a copy of Microsoft. This way I get a computer with it. And it is faster than the VM: 45 seconds per image, compared to up to 2 minutes for the VM.
There are other improvements as well. One of the things that drove me to distraction was that it could take minutes for DxO to access the source directory when I changed the contents. That is much faster now, and so is the general responsiveness of the UI. But the remote desktop has one serious disadvantage: display refresh doesn't really work, and I'm continually left with bits of other windows in it. Clearly a bug, but how do I report it? I'll just be told it's my fault.
|
More trees go
|
Topic: gardening | Link here |
CJ along today with his chainsaw, and removed an old, unhappy Acacia baileyana and trimmed one of the Pittosporums, both in the driveway. Things look very different now. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
There's more to go—a whole clump of lilacs—but it was too hot today.
|
More information about panorama brackets
|
Topic: photography | Link here |
Over the course of the years I've looked at a number of panorama brackets and commented on their poor design or poor presentation. On 25 October 2011 I compared three that I was thinking of buying. Today received mail from Ed Horka, telling me that the third one is in fact complete. There's a review on Flickr. It does, indeed, come with the third rail, so it has everything that you might need except for flexible angles on the rotator:
That wasn't obvious in the images that the seller provided, which also have the camera pointing in the wrong direction:
Given my experience with the Sunway rotator, which still has not been replaced, this might even have been an option.
| Friday, 11 January 2013 | Dereel | Images for 11 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Microsoft update hell
|
Topic: technology, opinion | Link here |
I suppose I'm in a pretty unique position, being a veteran of the computer industry but having almost no understanding of Microsoft. In the last few days, that has changed, of course, and now I'm learning things that even beginners take for granted.
Today spent some time bringing my new machine up to date. First I had to wake it from “sleep”, which I thought was suspend to RAM, but based on the time and disk activity it took to come back, it seems to have been at least partial suspend to disk. But the “run” light stayed on and blinked the whole time, which wouldn't have been necessary for that. And when it came back, it was at a resolution of 800×600! That's better than being off the scale, of course, but still irritating.
111 updates to install! So I installed them, which for some reason required several reboots. Another 15 updates! More reboots! Another 8 updates! And so on. I'm running low on my network traffic allowance this month, so I stopped after that—only (it claims) 3 important updates, the newest of which is dated 12 October 2010. It seems that this is “normal” behaviour, or at least to be expected, but why? Is it running down some kind of dependency tree and only discovering that it needs new updates after the ones that depend on them have been installed? In any case, the machine's only barely on the net, so I don't think I'm overly vulnerable. I'll continue with the new months' traffic allowance.
The good news is that the problems with the display resolution seem to have gone away. I think this is basically because it's finally accepting my settings, not that the underlying problem is fixed.
|
DxO on “Windows” 8
|
Topic: photography, technology, opinion | Link here |
So my experiments with running DxO Optics “Pro” in various configurations have shown that there's not much difference beyond what has proved to be DxO's complete inability to display the “Customize” window on a remote desktop. It's still glacially slow. But that's a 32 bit system, and they claim it's faster on 64 bits (why?). I still have a few days before the “Windows” 8 preview expires, so installed that on a disk and tried it out. It came up with a 640×480 display resolution, and I wasn't able to set a 16:10 aspect ratio, so ended up running at 1600×1200. The more I look at this, the more it looks as if there's something wrong with the EDID data transfer.
The first thing of interest was that it couldn't find the network drives. I've seen this before. I suspect that it's one of many race conditions inside the system, like the display refresh problems I've already mentioned.
So I copied the files to the local disk and let it run. The first thing I noticed is that the display was considerably sluggisher than under “Vista”—that's one of the things that I had been complaining about. Maybe it's related to “Windows” 8. In any case, it converted my files in only a fraction over 2 hours. That's faster than under 32 bit “Vista”, but not faster enough to be statistically relevant. Certainly it hasn't transformed a snail into a ball of fire.
|
More network disconnect problems
|
Topic: technology, opinion | Link here |
I've had several network disconnects lately, more than usual. Are they due to the hot weather? Certainly my signal strength seems less than usual, and dropping back to GPRS must be some kind of emergency action. I have adapted the /usr/ports/net/e169-stats/ port to log various events, so today I spent some time extending it to report RSSI before and after a cell switch, and also to report low RSSI (< 3). Bingo!
The numbers in brackets on cell change are the last reported RSSI. So it does seem that this cell switching is due to low RSSI, and in some cases it may cause a fallback to GPRS. What to do about it? I could buy a longer antenna, I suppose, but it's easier to pray that VCAT will finally hear Wendy McClelland's case and they'll build the Radiation Tower.
|
DSE web site: worse than ever
|
Topic: general, technology, opinion | Link here |
Another bad bushfire danger day today, and spent some time monitoring the DSE web site. It works better than the corresponding CFA site, but it's still terminally broken. To access any real information, you need to click on a link, which launches a Javascript function such as loadSummaryPageFromParent():
|
|
||||||||
And this function appears to be broken, or at least unreliable. It's not a browser-specific problem, though I wouldn't put it past them to write stuff that only displays on “Internet Explorer”. But I tried it with “Internet Explorer”, and it didn't work there either. To its credit, though, it reported “link error”.
It's not the page itself that's the problem—that works just fine. They've replaced a perfectly normal link with a flaky Javascript function that doesn't allow you to display the page in a different tab or window, and which brings you back to the default view when you're done. Bravo!
On the positive side, Callum Gibson came up with alternative views, designed for mobile devices, like http://osom.cfa.vic.gov.au/public/osom/websites_headline_mobile.html. This appears to contain the information without the stupid Javascript. It's not very helpful, though—the towns named seem to be very small:
20KM NNE LICOLA12/01/13 08:53:00 AM
Incident Information: Department of Sustainability and Environment advise that there is fire activity in the area 20KM NNE LICOLA, one kilometre from the McMillans Walking Track and chalet. ... Read more
Where the hell is LICOLA? Looking at it on the map, it seems to consist of about 10 houses in the middle of nowhere. And most of the reports are like that. How can anybody guess where the incident is? Geographical coordinates would help a lot, but even their map is hopeless. And then there's an incident summary page, conveniently with subpages ordered by “township” names. So not only do I not know where LICOLA is, I have to go looking for it ordered by the first letter of its name. Why can't they order by region?
| Saturday, 12 January 2013 | Dereel | Images for 12 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Panorama exposure
|
Topic: photography | Link here |
House photo day again today. The weather was suitable, so took another 360° panorama of the garden centre. That's a particularly difficult one to do, because the proximity of the leaves, the difficulty of establishing control points, and the differences in brightness.
One of the rules in panoramic photography is that you should expose all the component images the same. But even today, without sun, this scene had a difference of nearly 7 EV between the darkest and the lightest parts. How do you handle that? Even with tone mapping I'd need to take at least 5 images of each view, and my camera can't handle that many automatically. And that would be completely catastrophic for an image like this one:
|
|||||||||||||||||||||||
Those leaves come within 15 cm of the lens. To get them sharp, I need a small aperture. This image was taken at 1/60s (EV 11.9). The slowest would be 1/15s, and the slightest movement would mess up both the individual image and the tone-mapped composite.
I've already established that it makes more sense to expose the images “correctly” and let enblend work it out. That works up to a point, like in the verandah centre panorama, which is made from tone-mapped images made from 3 images bracketed 2 EV apart. Today the total range was from 7.7 EV to 14.0 EV, a range of 6.3 EV. And that worked.
But today I overdid the shots into the sky, and that was wrong. Without tone-mapping, I had a range of 9.7 to 16.6 EV, and the sky shots were completely underexposed. Even after applying as much compensation as I could with DxO Optics “Pro”, the results were terrible:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And I couldn't find enough similarities to stitch. That's the one that got away.
For some reason the verandah centre panorama wasn't as successful either. I think I forgot to level the tripod before taking the images, and the result was slanting verticals:
|
|||||||||||||||||||||||
There are things called horizontal and vertical control points, which describe lines that should be horizontal or vertical. I've only had limited success with them so far. Time to try again.
|
Flash! Bang!
|
Topic: photography | Link here |
Taking some photos with my Mecablitz 58 AF-1 flash this afternoon, and suddenly there was a loud bang! from the flash unit. Dead.
What caused that? The electronics seem to be OK, but it no longer flashes. My best bet is that it's the flash tube itself. Looked around and discovered replacement tubes for exactly my flash on eBay, for quite a reasonable price. But is it the tube? And how do you replace it? Asked on the German Olympus forum, but didn't get much of use beyond “You gets what you pays for”.
|
No dinner tonight
|
Topic: general, opinion | Link here |
We had invited all the Yeardleys to dinner tonight, but it didn't happen. For reasons I don't understand, David blew up at Chris because he wanted her to do something with him when she had planned to do some training with Yvonne. What right does he have? They've been divorced for 10 years or so, and he's on his third wife since then. In any case, Chris was particularly cut up about it, so we decided to put it off until tomorrow.
I can see happy times ahead. David's behaviour is one problem, but he's leaving at the end of the month. But Tuyết and Min Chau will stay, and what will they do with themselves? I fear there's more trouble brewing.
|
Another bloody power failure
|
Topic: general | Link here |
Another power failure this evening, just a short one. But that's 2 so far this year, well above the average.
| Sunday, 13 January 2013 | Dereel | Images for 13 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Still more blocked drippers
|
Topic: gardening | Link here |
Took a look in the veggie patch this afternoon. Parts of it are doing well, but in other places plants are dying back: a combination of heat, lack of rainfall and... blocked drippers and drip pipe! The areas that were well irrigated were in the range of a mini-sprinkler. This is becoming a serious problem. A while back I had come to the conclusion that drippers were better than sprays because they watered only the things I wanted watered, and they didn't wet the foliage, but I'm reconsidering now. In any case, put some sprinklers into the dripper line. It'll do for now; looks like I'll have to buy more sprinklers.
|
Yeardleys to dinner
|
Topic: general | Link here |
Yeardleys over in force today: Chris, David, Tuyết and Minh Chau, along with Tuyết's mother Ngu:
|
|||||||||||||||||||||||
David is getting to be more and more of a problem. He can't talk about anything except himself, and even then you've got to be really careful. We were talking about birthdays—it seems that Minh Chau was born on 13 October 2000—and he mentioned his son, who was born on 26 October 1976. I forget what he was saying about him, but he made it sound as if they had a good relationship, when it was my understanding that they hadn't seen each other in years. So I said “I thought that was the bloke who didn't come back”. Bang! One exploding David, who gathered the rest of the people and left. Yvonne was really upset, and doesn't want to have him here again. I can understand that.
| Monday, 14 January 2013 | Dereel | Images for 14 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Repairing flash units
|
Topic: photography | Link here |
Started looking for somebody to repair my Mecablitz 58 AF-1 today. It seems that the Australian distributor for Metz is C.R. Kennedy in Melbourne. Called them up, got transferred to their repair department, and it timed out. Left a message, but they didn't call back.
In the meantime, went looking for local repairers. Found Thornton Richards Camera House, “Photographic Equipment & Supplies--Retail & Repairs”, so called them up and was told that “Repairs” means that they find out who repairs them and send the thing to them for repair. So much for that.
Called Kennedy's back again and was told what I expect—send it in, we'll look at it and give you a quote. Asked what sort of money I was looking at: if it's only the flash tube, $180. It could be more, of course. Add to that $20 transport in each direction, and it's a minimum of $220. Is it worth it? It's nearly 4 years old, and I can buy a new Mecablitz 58 AF-2 digital, the successor, for $345 US. But wouldn't it be nice to find a way to replace the tube myself? Looking at the head, the diffuser has two cutouts at top left and right, suggesting that it's designed to be removed forward, exposing the tube (and the reflector):
|
|||||||||||||||||||||||
But how? So far I haven't been able to work it out.
Spent some time looking at online offerings for flash units. When I researched this unit over 4 years ago there were only two makes of flash unit that understood the Olympus flash protocols: Olympus and Mecablitz. Now there are lots of them, starting as low as $50. Some make obviously incorrect claims: “TTL” flash mode with only one contact. The best you can do there is “automatic” flash, without camera involvement. Much research needed.
|
Your account has been disabled
|
Topic: general, opinion, technology | Link here |
Logged into the ANZ web banking application today—or I tried to:
|
||||||||||||||||||||||
“This Customer Registration Number has been disabled”. Why that? I have a link to on a private page, but of course they don't allow you to save passwords, so it's possible I made a paste-o. Tried again on another browser, and sure enough, it worked.
But why did I get the message? One thing's clear: it's imprecise. “This” CRN. Why doesn't the page repeat the number? That way it would be clear if you've just made a mistake in the number. But if that's the case, why “disabled”? Most likely it would be an invalid number, in which case the message should say so.
Of course, maybe there was something more sinister going on. It certainly made sense to check. Called up ANZ, who confirmed that the account had not been disabled and asked me to log on Yet Again, and confirmed that all was well—apart from stupid messages. They could fix that.
|
More panorama refinement
|
Topic: photography | Link here |
Saturday's verandah centre panorama was less than perfect: I think I must have not leveled the panorama head properly, and the verticals were slanting: The following views are rotated roughly 180° from my preferred view to illustrate what I've done.
|
|||||||||||||||||||||||
One thing that the control point detectors don't normally generate are vertical and horizontal control lines, though I've seen cpfind this happen on some occasions, with cpfind, so tried again with that. This time I didn't see it generate any line control points, though the first attempt looked much better than the corresponding run with panomatic, with an average error of 1.0 pixels and a maximum of 3.0 pixels. But the verticals were still slanting. So I added some vertical control points in images 4 and 12:
|
|
||||||||
|
|
||||||||
|
|
||||||||
|
|
||||||||
That improved things somewhat, but the east side (away from the house) still looked wonky. Running the mouse over any image shows its partner to show the differences more clearly. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
So I added a couple of control points on that side.
|
|
||||||||
|
|
||||||||
Strangely, that made almost no difference:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Looking at the control points table, the newest control points had more errors than the others. I wonder if it's worth trying further.
Interestingly, none of these additional control points had any effect on the average and maximum errors; they stayed at 1.0 and 3.0 pixels for all images.
| Tuesday, 15 January 2013 | Dereel | Images for 15 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New flash units
|
Topic: photography | Link here |
So: do I repair my Mecablitz or buy a new flash? Spent some time investigating both alternatives.
In particular the new flash units on the market are of interest. The ones that I looked at are:
YongNuo YN-560, advertised on eBay as a “Digital TTL Shoe Mount Power Bounce”, whatever most of that is intended to mean. But the “TTL” is Just Plain Wrong. It's purely manual, without even an “automatic” mode. That's both surprising and a pity. It has a very fast recharge time, only 3 s with alkaline batteries, and a guide number of 58, the same as the Mecablitz. But the lack of any automatic exposure control kills it. YongNuo does have a couple of TTL flashes in their programme, but not for Olympus.
A “Cactus AF50”, for whom I can find no manufacturer. It apparently supports most if not all of the Olympus flash protocol. It gets good reviews from the Four Thirds community. But it has a guide number of 50 and a recycle time of 9 seconds. I suppose it's a possibility, but the recycle time worries me. There's also an AF45 model with guide number 45.
The Vivitar DC-383, another unit where the manufacturer apparently hasn't bothered to publish the specs, but I found the instruction manual on a third-party site. It has a guide number of only 45, and a recycle time of 9 s.
For some reason (reliability?) I can't find many Mecablitz flashes on eBay, or even at Adorama. B&H have a number, starting with the 36 AF-5 for only $119. As the name suggests, it has a flash number of only 36, but that might be enough for many situation.
There are a number of other Mecablitz models at B&H, culminating with the 58 AF-2 digital for $345.
So: what do I really need? I have almost never used many of the features of the 58 AF-1, notably the remote flash (why do people like that so much?) or the high-speed synchronization. The things I do like are fast recycle and wide angle (at least “18 mm” to match my 9 mm lens). But what about flash intensity? There's apparently a connection between that and the tube reliability. How often do I really need a guide number of more than, say, 36? How can I know. Certainly the difference between $119 and $345 is significant. To be pondered.
|
Repairing the Mecablitz
|
Topic: photography, opinion | Link here |
The other alternative, of course, is to repair the Mecablitz myself. That's definitely the cheapest way—if I can open the bloody thing. Got some leads on the oly-e forum, notably this one, which actually shows photos of how it's done—for subscribers only. Basically, the sides of the head slide back when you release a clip on the inside. But how? I've tried with slips of plastic, which go as far as the clip, but can't release it:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder what the correct tools are.
|
Garden flowers in mid-summer
|
Topic: gardening | Link here |
Today's the middle of the month, time for the monthly photos of flowers in the garden. Comparing with this time last year, it's clear that the hot beginning to the month has taken its toll. The Hebe were in full flower this time last year, and now it looks as if all the flowers have died:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Mandevilla flowers show the problem well. The ones in the sun are all burnt, while the small plant in the shade is looking happy enough:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And the Clematis seem to have suffered not just at my hands. Last year the whole south side of the verandah was covered with them, and this time there are only a few blooms left:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Kniphofias, which should have flowered months ago, are finally coming around. I wonder why it took them so long. Is it a different kind, or something to do with where I planted them?
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And other plants, notably Epilobium, Hedychium and Osteospermum, aren't flowering at all. In general, it seems that mid-January is not exactly the best time of year for flowers. About the only positive thing is the Buddleja davidii that Mike Sorrell gave me a year ago, which is already over 2 m high and living up to its common name “butterfly bush”:
|
|||||||||||||||||||||||
|
Network flakiness: new insights
|
Topic: technology, opinion | Link here |
My network connection is bad again. I get the feeling it happens when it's warm. Today it was not timing out, but I was getting impossibly bad connection quality. Went through the whole rigmarole of restarting the ppp process, popping the modem, even rebooting the machine. No help. Then it occurred to me to compare the throughput with the continual cell hops, so I ran a ping every 5 seconds. Bingo!
Every time the connection switches to cell ID 8fc8e4a, no data gets through; it backs up until the connection switches back to another cell. Looking at the first incident, at 17:53:39, it took 58 seconds until it switched back again. And ping sequence 10 took 59.5 seconds, with the following ones evenly spaced at 5 seconds shorter: they all got through at substantially the same time. The second time is similar: the system stayed on the cell for 24 seconds, and after changing back again, the first ping took 23 seconds. That's pretty clearly a problem with the cell. Does Optus still want me to go through this stupid reboot, replace, reinstall rigmarole given this evidence?
| Wednesday, 16 January 2013 | Dereel | Images for 16 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Google false positives
|
Topic: technology, opinion, photography | Link here |
Spent a lot of time today trying to open the head of my Mecablitz, without success. The information I got yesterday wasn't really sufficient, and I went out looking for other information. Google is your friend, right? Well, up to a point. It seems that in the last few years the quality of the results has dropped. What would I like to find? Mecablitz 58 AF-1 service manual. But of the first 10 results I found there, only 3 even included the word “service”, and that out of context with “manual“. Why did I get the other 7? OK, the double quotes still work, so I looked for Mecablitz 58 AF-1 "service manual". That helped: this time 6 of the 10 hits had “service manual”, though none related to the Mecablitz. In addition, one didn't have the text “Mecablitz”. Why is this so difficult?
So went looking in Google Images, where of course I found nothing of any use. I've already commented on the extreme number of false positives there. If I select “Images larger than 4 MP”, I see a cross-section of my life—of the 42 photos on the first page, 27 were taken by me, including our dog Nemo, various photos of the garden, potatoes, cats—in fact, everything except a flash gun. As I noted two years ago, they seem to include all images on a page. On this occasion it was my diary for January 2010, where I made one mention of the flash unit, and didn't include any photos at all. Surely they can improve their algorithms?
|
Mecablitz: impasse
|
Topic: photography | Link here |
All methods to open the Mecablitz have so far failed. Other attempts on the web have been a dead end. I suspect there's some detail that the DSLR-Forum article doesn't mention. It's interesting that a total of four people were actively involved in that thread: two who wanted to open the unit, and two who had managed to do so. But the two who wanted to didn't report any success. Added another question to the thread; I wonder if anybody will read it.
| Thursday, 17 January 2013 | Dereel → Ballaral → Dereel | Images for 17 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Thüringer Bratwurst
|
Topic: food and drink | Link here |
After our success with baking our own bread, we've been thinking for some time about making our own sausages, since the ones we get in Australia taste terrible. And what better to try than Thüringer Bratwurst? We've been researching for weeks, and finally ended up buying some casings on eBay—from Scotland! With postage they were cheaper than any local supplier, and it seems they have a permit to export them to Australia, something that's not easy to get.
The casings arrived a couple of days ago, so the next step was to make a filling. After much comparison came to the conclusion that the basic filling was 2 parts pork shoulder and one part pork belly, so yesterday Yvonne bought some of that. Today we worked out the final recipe. In particular the quantity of salt in the references varied noticeably, from 17 to 22 g per kg. And some wanted eggs and milk, others—including a real live butcher's recipe from Arnstadt—didn't. So I decided not to put in any, until I saw how firm the mixture was, after which I decided that a couple of eggs would help fill the sausages.
That's the easy part. How do you fill the sausages? You can find lots of instructions, but they leave out important details. I thought that we could probably force the filling in through a funnel, but Yvonne didn't. And what I didn't know was that years ago she helped Marie-Louise, an in-law, to make sausages. For that reason she had bought a sausage filler attachment for the mincer yesterday. Unfortunately, it didn't fit: it was just a tiny bit too large. But clearly we need something like that, so after some deliberation and a couple of phone calls, set off to Ballarat to buy a hand-cranked mincer, though we already have an electric one: the electric one seems too uncontrolled for what we want to do.
Back home, and tried to work out how to use the things. I had thought of pushing the mixture down the casings, but fortunately Yvonne has done this before. You pull the casing up on the nozzle and fill like that:
|
|||||||||||||||||||||||
And there our first problem showed itself: the casings only barely fit on the nozzle. We didn't know how much we would need—I had estimated 1.5 metres—but we couldn't get anything like that onto the nozzle, so we had to do it in several steps. This is our very first string:
|
|||||||||||||||||||||||
The first one looked a little strange (top left), and we still have issues getting the length right, but on the whole it's not too bad.
Using a hand-cranked mincer is not as good an idea as I thought it might be. There's a lot of mixture to go into the skins, and maybe the electric mincer would be a better idea after all. But first I need to find a nozzle that fits.
One thing that's relatively irritating: getting the last bit of the filling out of the nozzle. Quite a bit gets left behind, about enough for two sausages. It's fiddly:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Still, finally we were done, and we had far more sausages than I had expected:
|
|||||||||||||||||||||||
And how were they? In the course of the afternoon I had read that they should take 1 minute frying per millimetre thickness. I thought that excessive, but in fact they did need much more cooking than normal sausages. Finally they were done:
|
|||||||||||||||||||||||
They tasted good, though I think I might go for more spices next time. But I had forgotten what a Bratwurst looks like:
|
|||||||||||||||||||||||
Interestingly, on the instruction page from Tongmaster, they answered a question with “Your best with the
belly or shoulder”—the same ingredients. Maybe real EnglishScottish sausages
don't taste that different from Bratwurst.
| Friday, 18 January 2013 | Dereel | Images for 18 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More garden flowers
|
Topic: gardening | Link here |
As often, I missed out some flowers in my recent survey. One of our variegated Bromeliads is flowering, and rather to my surprise I found two Chiles poblanos growing on the sole surviving bush from last year:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's a lot earlier than this year's plants. I wonder if it's normal to treat them as perennials. There's also a strange plant growing on the surface of the pot, probably a weed, but interesting in shape:
|
|||||||||||||||||||||||
Peter Jeremy tells me this is a Liverwort
And there are a number of flowers on the verandah that I didn't expect. The first is obviously a stowaway on a pot of Convallaria majalis. I think it must be a pumpkin. And the second is a small succulent that we've had for a long time, but which hasn't flowered before:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
There's also a mystery plant that my uncle Max gave me for my birthday over 4 years ago. It has particularly nice leaves, and I've been waiting for it to flower. Well, yes, it flowers, but it's hardly worth the trouble:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder what the plant is.
|
Mystery insect
|
Topic: animals | Link here |
Going out onto the verandah, I found something strange on one of my thongs:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At first I thought it was some composite, like some of the beetles I've seen running around joined together, but further investigation shows that it was a butterfly, I think freshly emerged from the pupa:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Opening Mecablitz: progress
|
Topic: photography | Link here |
Got an answer to my question about how to open the Mecablitz today, along with more detailed photos. So I tried that several times all day long, with no success. The real issue is to get the side wall of the head pushed out far enough that the clips clear the body. And you can't see that. Finally pushed in a couple of screwdrivers, and was able to confirm that yes, the clips no longer engaged in the body:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
So tried to push it back. Still no joy. So I got a bigger hammer (actually a shifting spanner) and tapped with that. Finally got the thing off. It has only taken me 6 days. And there's still the rest of the head to go. Why do people make such strange constructions? It has certainly lessened my opinion of Metz.
| Saturday, 19 January 2013 | Dereel | Images for 19 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Lame mouse again
|
Topic: technology, opinion | Link here |
Into the office this morning to find my mouse cursor moving slowly and unevenly. Further investigation seemed to show that it was related to where I placed the mouse on the (active) mouse pad. It looks as if it is dying. After a bit of searching found another one with side buttons (from teevee) and used that for the day. But in the evening the thing was working normally again.
What's wrong here? Should I care? A mouse costs nothing, but where do you find one with which you can simulate the middle button? Since people replaced the middle button with a scroll wheel, I've been remapping the side button as Button 2. And not all mice have that. Did some looking round eBay, but I don't think that's the place to buy a mouse. Looks like a trip to OfficeWorks in the near future.
|
Limiting the flood of photos
|
Topic: photography | Link here |
House photo day again today, and I'm further reducing the number of photos I take. I have up to 277 photos of specific views, notably the oldest:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
How do I present them? In any case, in many areas things don't change on a weekly basis, so I'll limit the photos to the ones that do.
Last week I had problems with the verticals in the “verandah centre” panorama. I attributed them to incorrect leveling of the tripod, so this week I paid particular attention to the leveling. And I still got problems with the verticals! I wonder what causes the issue.
|
More Mecablitz fun
|
Topic: photography | Link here |
So now I have one side off the head of my Mecablitz. The other should go off in exactly the same way, but it wouldn't budge. Possibly it's related to the button on that side for adjusting the direction of the head. There's one on the side I removed, too, but it's a dummy. I have some ideas, but before I damage the unit, it sounds like a good idea to ask again on the forum. And then it occurred to me: the person who answered the question is still looking for a flash tube, after over 6 months. I still haven't heard from anybody who has actually finished the job. This can take a long time.
|
More garden excuses
|
Topic: gardening | Link here |
I've been neglecting the garden recently, but at least I have an excuse: the flies. It seems that there's never good weather for the garden. Did a little weeding, but I think the weeds are getting the better of me again.
|
Researching sausage making
|
Topic: food and drink, opinion | Link here |
It looks as if what I really need for sausage making is a set of filling tubes, one per sausage size. Carola Schlanhof will be visiting us next month from Austria, and she could bring some with her. Did some looking on eBay and found only a small selection at prices of up to € 28 each. Most of them are in Germany and would cost almost as much to send to Austria as to Australia. Once again, maybe, eBay isn't the right place to look for this sort of thing.
Also looked for recipes, and came across this site that might be of interest, though it, too, seems to lack detail.
| Sunday, 20 January 2013 | Dereel | Images for 20 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Riding again
|
Topic: general, animals, opinion | Link here |
I've been remarkably lazy in the last few weeks, to the point where I'm getting worried about my apathy. But today I finally got round to doing what Yvonne does every day, and what I haven't done for over 9 months. Why not? I really don't know. Didn't go very far: Darah was very willing to go fast, but she tired quickly, and we were only out for an hour or so of mainly walk.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I should do this more often.
|
Mecablitz: finally open
|
Topic: photography | Link here |
More investigation of the Mecablitz 58 AF-1, and finally made further progress. As suspected, the button on the left-hand side was in the way, and needed to be pushed in. Then things were relatively simple: the diffuser is held in place by 2 clips on each side. Here it is lying on top of them after having been removed:
|
|||||||||||||||||||||||
Underneath is the reflector, shown here in full wide-angle position, and two pieces of plastic, which may have some function when the reflector is moved back for longer focus settings, but are completely out of the light here. They're apparently held in place only by the diffuser. The tube was also interesting: only half of it is left:
|
|||||||||||||||||||||||
I wonder where the rest has gone.
The next question is: how do I remove the reflector? There are some things that look like clips on the edges, but they're very difficult to recognize. Hopefully further investigation of the following photos (taken from the left side of the left, middle and right of the reflector) will help:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Mystery plant: liverwort?
|
Topic: animals, opinion | Link here |
Peter Jeremy commented on my mystery plant mentioned the other day:
|
|||||||||||||||||||||||
He thinks it's a Liverwort, which sounds likely. Now to find out what kind.
| Monday, 21 January 2013 | Dereel | Images for 21 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Welcome Christiane Bahlo
|
Topic: general, opinion | Link here |
Message today from a certain Christiane Bahlo. That's a name that I know, but it still came as a surprise: even a couple of days ago she was Chris Yeardley. But now she has reverted to her maiden name, an action which, it seems, is handled not by Births Deaths Marriages Victoria, as you might expect, but by VicRoads, because they issue driver licenses. What kind of state we live in!
Anyway, I have 200 references to the old name in my diary. How do I handle that?
|
Updating Microsoft, continued
|
Topic: technology, opinion | Link here |
Today was the start of a new billing month for my Internet traffic, so I was able to continue with my software downloads and updates. In the end I downloaded nearly 1.25 GB of data, including further updates for dxo, my Microsoft “Windows” box. When I postponed 10 days ago, there were three updates outstanding, none of them new. Today it changed its mind: I needed to download Service Pack 2, some 350 MB (or was that 45? It didn't seem to be sure). So it downloaded that, and then it told me that I needed to install another 66 important updates. The more I see this, the more it tells me that the update program has no lookahead: updates dependent on yet-to-be installed updates are not visible.
So I installed the updates, including one for my monitor. That's interesting, because it's the first indication that the system knows what my monitor is. And when I rebooted, it came back with the display frequencies out of range. I wish I understood why that happens. Clearly it can't be because it doesn't know the monitor.
So, another check for new updates. No go:
|
|
||||||||
|
|
||||||||
OK, what does that mean? Followed Help and got one result, which led to a generic page:
|
|
||||||||
|
|
||||||||
Now what use is a 32 bit error code if you can't find out what it means and have to resort to guesswork? Tried again, and it worked. One important update to install. Did that. 2 important updates to install. Did that. No important updates! Just 15 optional, 11 of which looked important enough to me, so started installing them. After 8, the network connection (which has been very flaky today) dropped, and I got... error 80244023. So this means something like “network connection dropped”. Why can't they say so? Numerical error messages went out decades ago. I'm reminded of my attempts to fix it on Tandem.
Still, finally it seems that my machine is up to date. It only took all day.
|
Lilac sick again
|
Topic: animals | Link here |
It's been only a few months since Lilac was last incontinent, but last night it happened again. Last time we thought it might have been because of Zhivago, but she's used to him now. I fear this is going to happen again and again. So does the vet: she gave Yvonne a second injection for next time. But I wonder how much longer Lilac will be with us. She'll be 16 years old on Saturday.
|
New video card
|
Topic: technology, opinion | Link here |
It's been several months since I got my new monitor. And for that time I've had to split my X configuration across two machines, because the display cards in eureka can't handle all the monitors. I've been planning to replace one of the cards for some time now, but research is important, especially since the documentation of these cards is so terrible. I had more or less decided on an nVidia GeForce GT 640 as a good compromise between cost and number of monitors supported. But some of the cards don't have enough outputs. So last weekend I did some research and found a number of cards with the chipset. Here an overview:
| Maker | DVI outputs | Cheapest Price | Seller | |||
| (with postage) | ||||||
| Afox | 1 | |||||
| Palit | 1 | |||||
| Zotac | 2 | |||||
| Galaxy | 2 | $102.78 | Mwave | |||
| Gigabyte | 2 | $108 | PC Case Gear | |||
So last Sunday I ordered a Galaxy from Mwave, and it didn't arrive until today. And it only has one DVI output! I'm sure that the advertisements showed two, but when I went back to look, they have already discontinued the product! The manual on the CD shows two as well, but that's probably generic. What do I do? I can certainly return it, but that'll have me $25 out of pocket, a significant proportion of the card price. First I'll find out whether I can use the HDMI connector: at least one of my monitors has HDMI, so that would be an acceptable solution.
|
Video card: time to tidy up
|
Topic: general, technology | Link here |
The new video card raises an issue that has been looming for some time: I should both tidy up and rearrange my office. I haven't really changed anything in over 5 years, and it shows:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The second photo shows the mess behind the monitors, where I can't get at it very well, looking down from above. The open chassis on the left is boskoop, my Apple, and on the right is the back of eureka. There's also an animated view of the first photo. It serves as the “before” image in what I fear will be a protracted process.
|
DxO: Random processing times?
|
Topic: photography, technology | Link here |
Lately my DxO Optics “Pro” processing times have been particularly good, at least in comparison to what they have been—closer to 25 seconds per image than the previous 45 seconds. Why? I have no idea. I saw this under VirtualBox as well, but there I thought that maybe it had something to do with VirtualBox. I'm beginning to think that it's riddled with heisenbugs. Certainly the screen refresh code is very dubious. Still, as long as it lasts, it's an advantage.
| Tuesday, 22 January 2013 | Dereel | Images for 22 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Upgrading graphics cards
|
Topic: technology | Link here |
So before returning this incorrectly described graphics card to Mwave, I wanted to see whether I could use it to drive three monitors, including the HDMI connection. Tried to put it into dxo, the Microsoft box. No go: it's too wide, so I'll have to try it in dereel. But before I do that it makes sense to migrate what's left of dereel (32 bit FreeBSD 8.1) to a VirtualBox machine. And that required setting up disks, which took a while. Mañana.
|
Smart meter installation
|
Topic: general | Link here |
A power outage of a different kind today: Powercor installed a new, “smart” meter. It wasn't as involved as I had thought, though in retrospect I shouldn't have expected it to be involved. Power was out for only 15 minutes.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The most interesting thing is that to do the work he had to disconnect power. And this is on the live side of the main switch. But there's a fuse up the wall:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's amazing that these things are out in the open like that. I wonder how many people siphon free power off from that point.
The other thing of interest was that he did some kind of ground check with a stake into the ground:
|
|||||||||||||||||||||||
But the ground is as dry as a bone at this time of year, and he had to try three different places in irrigated beds before he got anything like a ground.
So: the docco that Powercor sent told me that I could monitor power usage from inside the house. How? The excuse for docco I got with the meter doesn't mention it, and the installer tells me I have to go out and buy equipment. Wonderful.
Took a look at the device. There are two LEDs that light apparently at random:
|
|||||||||||||||||||||||
Wh is clear. But what does varh mean? The installer didn't know, and it's not in the docco. But then, why should things like this be better documented than computer equipment?
|
Hacking ls, the discussion
|
Topic: technology, opinion | Link here |
My article about adding an option to ls a few months back aroused some interest, and Lim Cheng Soon asked for permission to publish it in Hacker Monthly, where he changed the title to “Hacking ls -l”. In the process discovered a very lively discussion about the article, all apparently dated the day I wrote it and long before the publication in Hacker News. Some are worth addressing:
Some addressed the usefulness of adding the -, option in the first place. Use -h (“human” format), or run it through a perl script like this:
-h doesn't cut it for me. I find it incredibly ugly and difficult to read. The perl script would probably do it. It would take longer, of course. This boils down to a separate question: is it important enough to add another option?. I think it is, but it's quite possible I could be outvoted. So far, though, it's in the tree, and nobody has complained.
The compiler warning, one of the things on my to-do list that I chose not to fix, is gone in gcc 4.7.2. That doesn't help FreeBSD much though, because we're moving to clang, and it has the same error. Nobody in this sub-thread seems to have noticed I wasn't talking about Linux, though they did in other places.
Somebody has already added this functionality, conveniently with a different option letter (-m instead of -,). That's a Tower of Babel issue, and one that concerns me. -m won't work on FreeBSD. It's already taken:
That's what standards bodies are about, of course, but that also stifles innovation.
One person claims that with Linux this will do the job:
It doesn't for me. I'm not sure I want to find out why not.
Finally there's a subthread explaining what I really wanted to express.
Still, I'm surprised about the level of interest this aroused, and even more that I only found out about it by chance.
| Wednesday, 23 January 2013 | Dereel | Images for 23 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Wendy on TV news
|
Topic: general, opinion | Link here |
While my network problems drag on, with no end in sight, Wendy McClelland and cronies have managed to get media attention. Last week the Ballarat Courier published a letter from Christine Crawford, claiming that the authorities are trying to “gag” them. I wish they would! And last night Megan Karbanenko of “Win” News to do a report on the poor, oppressed people of Dereel, who are “outraged” at having this life-threatening radiation tower forced on them:
The funniest part of the thing was where she donned what we're calling a “tinfoil hat”:
|
||||||||||||||||||||||
Of course, it really does offer good protection from the most dangerous kind of radiation: sunlight. But what really annoys me (apart from my abysmal Internet connection) is the arrogance of claiming that she in any way represents the residents of Dereel. I was also left wondering why the press always seem to side with this kind of idiot.
Things didn't stop there, though: there's a discussion on Reddit, and the ABC picked it up, in the process somewhat improving my opinion of the press.
|
Summer pain
|
Topic: gardening | Link here |
Irrigation has been a serious issue this summer, and I haven't really been doing enough about it. Round the front of the verandah, the Alstroemerias are dying back, and the Acanthus are also looking very unhappy:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I had thought that the problem with the Alstroemerias was a blocked dripper, but they're working. It seems that we just need much more water than I had expected. More work ahead.
|
Migrating dereel to VM
|
Topic: technology | Link here |
Today got round to the next stage of my computer restructure: test whether I could really drive my HDMI monitor via the HDMI output. I did that on dereel.lemis.com to avoid disruption to eureka.lemis.com, my main machine. After a bit of messing with the BIOS, the result: yes, I can. That's more than I can say for the DVI connection:
|
|||||||||||||||||||||||
That's not serious, beyond the fact that I can't test in that housing. But maybe I should move the motherboard to something more conventional.
To test X, I needed to run a more recent version of FreeBSD, which will be 64 bit (9.1-STABLE). But one of the reasons I'm still running dereel (32 bit 8.2-STABLE) is because of problems I've had with 64 bit executables, so other machines, notably Yvonne's lagoon, have NFS mounts to dereel. So I needed to migrate dereel to a VirtualBox virtual machine.
That proved more difficult than I thought: I already had a VM called dereel, running an older version of FreeBSD, and somewhere in the course of time I've ended up with disk images in many different places. And VirtualBox, of course, goes along with the Microsoft view of things and shows only the file name, not the path name, unless you go to some trouble to get it to do so. Thus I ended up with three different images, all displayed as dereel.vdi. Finally sorted out that mess, copied the real disk contents to the image, and started the system. No networking! The interface comes up, but there's no traffic. I've seen this before with my gigabit Ethernet card, but I had blamed that on some hardware incompatibility.
Apart from the disk, there appeared to be no difference between the configuration of the old and the new dereel VMs. With a bit of virtual disk swapping, discovered that the kernel I had really does have difficulty with the virtual network card (Intel PRO/1000 MT Desktop), while the old one doesn't. This is the same kernel that had similar problems with my new gigabyte Ethernet card, so maybe I've been a little hasty in blaming the card. Still, there are other virtual Ethernet cards, and the first I chose, PCnet-PCI II, works.
That allowed me to copy a modern system to the alternate root partition, and that came up happily with no issues beyond name and IP address. For once something that went relatively smoothly.
So finally I had dereel up running as a virtual machine, and how about that: things continued to work fine. Ran a tcpdump on dereel to see what was talking to it—only homephone.lemis.com, my SPA 3000 VoIP adapter, which called it for DNS and NTP. Fixed that configuration and there was no more traffic.
So how about shutting it down altogether and see what happens? I expected a hang of some kind, but things continued to run—until Yvonne wanted to reply to a mail message. It seems that Emacs was one of the executables on dereel. OK, as expected. Fired up the dereel VM again. Problem: lagoon had dereel:/home mounted, but the new dereel doesn't have a local /home file system (it uses eureka:/home instead). So things hung.
Back to eureka to continue, and X had hung itself in a T (stopped) state, presumably because of ssh login issues when I had tried to start an xterm on lagoon. So I had to restart X. Problem: the VM had been running under X, so it aborted when I stopped X. And something in the X startup wants an executable on dereel. Deadlock.
Round this point I gave up, shut down the new machine and rebooted the old dereel image, not without a bit of fun, since the disk names have changed between 8-STABLE and 9-STABLE. But I managed. What a pain this is! My own NFS mess has a lot to do with it, of course. Tomorrow I'll take a step back and migrate everything I still need from dereel to eureka.
| Thursday, 24 January 2013 | Dereel | Images for 24 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Smoke without fire?
|
Topic: general, opinion | Link here |
Woke up this morning and smelt smoke. That's bad news round here at this time of year,
especially as today was supposed to be a day of extremecode red fire danger. It
didn't look bad, but there was haze around as well. There was nothing on the broken DSE web
site, so called the Bushfire information line on 1 800 240 667, got an unintelligible
voice menu, hit 1 and got connected to Irene, who told me of several fires also not on the
map, but the closest was
in Mount Pleasant, 30 km
away.
I asked her to report it to the Dereel CFA, but no, I had to call the 000 emergency line for that. That didn't seem to be appropriate, but she said it was the only way. Why can't they connect?
000 told me that they could do nothing unless I saw a pillar of smoke (presumably by day), and that I should call the Bushfire Information Line—this after I had told them that I had been referred by them. Somehow, despite the horrors of 4 years ago and the ensuing Royal Commission, they're still completely disorganized.
Yvonne told me that the Dereel CFA have their own web site. Went looking. Indeed they do: http://www.dereelcfa.org.au/. It's not quite clear what the intention is, but it doesn't give any timely information. It looks like a template web site, with interesting links like Our Local Weather, showing a Weatherzone weather map for most of Victoria and large parts of New South Wales, but no links to bushfire activity. That almost certainly reflects more on the web designer than on the CFA, of course, but does it help anybody?
|
Revirtualizing dereel
|
Topic: technology | Link here |
Spent some time carefully analyzing yesterday's disaster with virtualizing dereel.lemis.com. There were two main issues: the symbolic links to files mounted from dereel, and the status of the virtual machine if X crashes. After much investigation, discovered that there weren't too many executables run from dereel, but that the libraries were a different matter. On the other hand, at least for the time being, I need to run dereel anyway. I'm running a newer version of PHP on eureka, and it's too leet to run some of my older web pages, so I run a second web server on dereel to handle them. As long as I do that, I don't have to worry about the executables on the root file system (which includes /usr/local). The good news was that, although I had /dereel/home mounted on both eureka and lagoon, there was nothing of great importance on it, so I was able to just umount it.
The virtual machine was more of a problem. Yes, there's a command VBoxHeadless which does the trick:
=== grog@eureka (/dev/ttyv5) ~ 3 -> VBoxHeadless --startvm dereel
But there's no output. It's really headless. After heated discussion on IRC, decided that there was no direct way to get it to display console output. The problem appears to be a combination of the fact that the console display under X is bit-mapped, and so far not enough people have been interested in implementing the display. Alternatives include VNC, which of course requires X, and simulated serial ports, which Callum Gibson promptly investigated. By the time he went home he had it working, but it still needs some tweaks.
In any case, for the time VBoxHeadless almost does it. First time round I didn't configure the network card correctly, with the result that the system hung in boot, and I couldn't see what was wrong with it until I repeated under X. I suppose I should investigate the serial console option.
|
Still more heat damage
|
Topic: gardening | Link here |
The recent hot weather is still showing its effects. The shade area, which is well-watered, is looking surprisingly unhappy:
|
|||||||||||||||||||||||
The artichokes we had been looking forward to have mainly not materialized. There are three shriveled up ones barely visible in this photo, and the one half-grown one is probably also not going to be worth the trouble. And some of the Hebes are looking even worse:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
We have drippers in all of these places, but it looks as if specific plants will need extra, so today Yvonne bought a sprinkler which we can move around. But hopefully we won't get much more of this kind of weather.
|
Internode: please reinstall your software
|
Topic: technology, opinion | Link here |
So a few days ago I sent traces of my UMTS network connectiona to Internode support showing that specific cells were associating but not accepting data. Today I got mail back. Here some excerpts, reformatted for legibility:
Maybe I should create a stone circle, go into it with a magic wand and jump up and down 7 times with the wand over my head. Do people no longer understand that everything he's asking for is a workaround for deficient error reporting? And it seems that they don't understand an error report when they see one. They say that this is necessary because the “wholesaler” (as they call Optus) requires it. But in that case there's something very wrong with Optus, too.
In this case, the things they're asking for achieve nothing. The information I supplied shows that the throughput is very varied, depending on the cell to which I'm connected. But today I had two fallbacks to GPRS, so next time it happens I'll repeat this download:
=== grog@eureka (/dev/pts/15) ~ 7 -> ftp http://mirror.internode.on.net/pub/test/10meg.test
“Screencaps” showing that the dongle works at another location. What does a “Screencap” look like? Like this?
|
|
||||||||
But what point does it make to try it in another location with different cells? It's completely pointless. And then they ask me to reinstall the software. I don't know how to do that. As it happens, I will try the dongle on a second PC (eureka) in the near future, once I've repositioned it, but for now it's not possible. I find this Microsoft-oriented guesswork “trouble-shooting” incredibly frustrating.
| Friday, 25 January 2013 | Dereel | Images for 25 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Mail delivery strangenesses
|
Topic: technology | Link here |
Early this morning, Yvonne told me that she hadn't received any non-local mail since last night. On further investigation, neither had I. Went looking around and discovered that my fetchmail config sent the incoming messages to dereel, which then delivered to /eureka. For some reason, it seemed, this was failing, possibly due to NFS lock issues. It was relatively trivial to change the .fetchmailrc to point to eureka, so did that, and pointed Yvonne at her ~/Mail/backup file, where I save everything that comes in, just to protect against procmail issues.
But that didn't work. The files landed in ~/Mail/backup, but not in /var/mail. Spent some time investigating it. The log file looks normal:
The first file made it to /home/grog/Mail/Caughtspam/spamassassin, but the second didn't make it into /var/mail/grog. Why not? At first sight everything looked OK in the mail log:
But there's a message missing from that log. The last few lines should look like this:
Where's the “delivered to mailbox”? No idea, but my eyes were going funny, so for the time being I have disabled procmail and SpamAssassin, which don't seem to help all that much anyway.
|
DxO random processing times again
|
Topic: photography, technology | Link here |
I've already observed that the processing times for DxO Optics “Pro” are variable between extremely slow and glacial. Today I had more strangenesses in the processing times. Recently the times have been in the order of about 25 seconds per image, but today I converted 4 images, and it took 3 minutes, 32 seconds, or 53 seconds per image. Then I converted another two, which took 56 seconds—28 seconds per image. Stupidly, I deleted them, and I had to reprocess them. Not easy: DxO doesn't want to do that unless I change some conversion parameters. So did that, then changed back to the original values, and converted them again. 1 minute, 56 seconds—58 seconds per image. What's wrong with this program?
|
How to put a Microsoft box to sleep
|
Topic: technology | Link here |
Despite the strangenesses with processing times, using a dedicated Microsoft box to process my photos is working out quite well. I put the box to sleep when I'm not using it, wake it up to process photos, and put it back to sleep again afterwards. Well, I try. I've set the power button to put the box to sleep, and discovered that it won't work if I still have an rdesktop session open. But it doesn't always work when I disconnect. Today I have come up with an explanation that may be correct: it only works if the (invisible) screen saver isn't enabled. So if I press a key on the local keyboard, then the power button puts it to sleep. Isn't Microsoft fun?
|
Password security
|
Topic: technology, opinion | Link here |
I'm continually ranting about the stupid rules people make me use to create passwords on web sites. Today I found a site that does a security check on passwords. No idea how accurate it is, but it confirms my expectation that these rules aren't very useful. Here some times for typical passwords (none of which, of course, I use):
| abc123 | instantly | |
| 4711 (typical PIN) | instantly | |
| 4712 | 2.5 µs | |
| grog1 | 15 ms | |
| Grog1 | 229 ms | |
| Iaml33t | 14 minutes | |
| I am leet | 132 days | |
| (FBBG user password) | 130,000 years |
Many password schemes require upper case, lower case and digits, and many also reject punctuation and spaces. But look at the results: they allow Iaml33t (14 minutes) but not I am leet (132 days). And the password I have for the Friends of the Ballarat Botanical Gardens, which also doesn't fit the rules, is simple enough for people to remember and not have to write down. It's nice to have some sort of confirmation that my approach works.
|
Serial console for virtual machine
|
Topic: technology | Link here |
As planned, investigated setting up a serial interface for a VirtualBox virtual machine today, according to the instructions Callum Gibson had worked out:
In the VM configuration, enable a serial port and select “Host pipe” and “Create pipe”. We both put the pipe in /tmp.
Create a file /boot.config an the guest, with the content -Dh (dual console mode, force serial console). This is described in boot(8).
Boot the guest and attach a telnet to the pipe generated in /tmp.
But that didn't work for me. When trying to boot, the VM crashed. So I removed the serial console again and it no longer crashed, but the boot hung. I was able to load the boot from the “CD-ROM” and then boot the system, so it could only be the boot blocks. How did they get overwritten? No idea. After replacing them, I was able to boot again, and the windows displayed
So it read /boot.config, but the message gives the wrong path name. Another thing to fix.
Reconfigured the serial port and tried again. VM crash at boot time. “A critical error has occurred”. The VirtualBox log files showed:
It looks like an Amiga! Still, clearly something's wrong. I'll try again some other time with a newer version of VirtualBox.
But there was still a sting in the tail. I started an Emacs on the VM with the display on eureka, and it hung the X session! No idea why, but it was repeatable, and I hate shooting down X, so I'll defer this one until I rebuild eureka. At any rate, dereel (the VM) carried on running through it all, so some things are working.
|
Leptospermum lives!
|
Topic: gardening | Link here |
While taking photos of plants with heat damage, came across one that has never flowered before:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I had thought it was an Acacia, but it clearly isn't. It's a Leptospermum brachyandrum, and it has been there for nearly 4 years: it's planted in the middle of what was a driveway, and though people had told me to prepare the soil, I didn't listen: “She'll be right”. Well, it has survived, and it's now marginally bigger than when I planted it; maybe now the roots have penetrated the gravel enough that it will grow more quickly.
|
Visitors from Germany
|
Topic: general | Link here |
Xenia and Anna Keppler arrived to visit us for a couple of days. They're the daughters of Christiane Keppler and Andreas Hähnle, who visited us this time last year. They're on a working holiday at various organic farms around the country.
| Saturday, 26 January 2013 | Dereel | Images for 26 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Happy birthday Lilac
|
Topic: animals | Link here |
Lilac is 16 years old today, quite an age for a cat. Sometimes I think she hasn't long to go, but today isn't one of them. Piccola honoured the event by bringing a mouse into the house, which Lilac promptly ate. Piccola spent the rest of the morning looking for it.
|
Still more DxO strangenesses
|
Topic: photography, technology, opinion | Link here |
House photo day again today. DxO Optics “Pro” converted them in fast mode—133 images in 66 minutes, 22 seconds, almost exactly 30 seconds per image. While stitching ran into trouble with the garden north image: the control point detector couldn't link all the images. Further investigation showed surprising chromatic aberration, which I traced back to DxO: it had turned off all the corrections. Why? I've seen this before—it's one of the issues I raised with my collection of error reports last year, and which they told me in no uncertain terms was because I was running in a virtual machine. Well, surprise, surprise, it's still here. It's probably my fault for using a remote desktop. Gradually I'm coming to the conclusion that it's my fault for using DxO, but the results—if you can get them—aren't bad. Time to write a page about my experiences.
So I had to reprocess the photos with the correct settings. It would be reasonable to assume that these corrections take a lot of calculation, so it would take longer. That could also explain why the times vary so much. But no, there was no obvious difference in processing time. That suggests that the inefficiencies in DxO are unrelated to the function of the program.
|
Indian food with the Kepplers
|
Topic: food and drink, general | Link here |
Since Xenia and Anna Keppler are here, had an excuse to cook some Indian food, also giving Anna the chance to make an Indian “taco”:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It was also the first time we cooked Alu masala with our own freshly-picked curry leaves:
|
|||||||||||||||||||||||
They're on the road for another few months, after which Xenia plans to start studying at the Uni Gießen, living in Ober-Rosbach, just 350 metres from where we used to live. That'll be a nice train journey.
| Sunday, 27 January 2013 | Dereel | Images for 27 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye Kepplers
|
Topic: general | Link here |
Xenia and Anna Keppler off again to Melbourne today. They don't travel light:
|
|||||||||||||||||||||||
|
Dismantling Mecablitz 58 AF-1: progress
|
Topic: photography | Link here |
More activity on the Mecablitz thread on the DSLR forum today. This contribution explained how to take the head apart once you have the sides off. The sides are really difficult to remove, but then there's just some adhesive tape holding the inside frame together. Remove that and it comes apart:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The next step is to remove the tube to the back of the housing. To do that, you first need to find the positioning motor, by moving the white card and wide-angle diffuser forward:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's as far as I went today.
|
More wind damage
|
Topic: gardening | Link here |
It hasn't seemed as if there's been much wind lately, but at least two plants have shown evidence: a volunteer Melaleuca has fallen over (to the south), and a Mirabilis jalapa to the north (here after being propped up again with a bamboo rod):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I'm always unhappy when plants get damaged, of course, but the Melaleuca was in a position where it really didn't belong, and the flowers (brushes) are pretty insignificant, so I won't try to save it. And somehow I've had many difficulties with the Mirabilis this year. The leaves on this one suggest that it's not going to recover this year, though since it's a corm, it should be back next year.
|
Dinner with the Yeardleys
|
Topic: general, food and drink | Link here |
Ngu, Tuyết's mother, is returning to Việt Nam on Tuesday, so today we had another dinner at the Yeardleys, where it seems that David did all the work, pulling out all the stops to deliver a 7 kg turkey with 2 kg of stuffing made according to a secret recipe. Another good meal, and once again we ate far too much:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Monday, 28 January 2013 | Dereel | Images for 28 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Still more dry plants
|
Topic: gardening | Link here |
The lack of rainfall this month continues to make itself noticeable. Today it was the Paulownia kawakamii that is looking unhappy, with shrivelled and dropped leaves:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Put a watering pyramid under it and gave it half an hour of water. That should improve things.
|
Alternative to DxO Optics “Pro”
|
Topic: photography, technology, opinion | Link here |
Peter Jeremy asked an obvious question today: why do I bother with DxO Optics “Pro” given all its problems? The answers are simple: I stick with it because the alternative means learning a new product with new bugs, and making comparisons to see which is better. But then Andy Snow suggested bibble, a program that I had once thought of using, but then not followed through, and which he says can do everything, including raw data conversion.
That's simple enough: Bibble doesn't exist any more. It has been bought out by Corel and is now called Corel® AfterShot™ “Pro”. And of course there's a free demo version, so downloaded that and tried it out.
The first part of my objections was soon clear: it's structured differently, though it still has this dark grey on light grey text that makes it so difficult to read. And once again it didn't want to find my photos on SMB file systems, like Olympus Viewer 2 did months ago. So again I had to copy them to a local file system. To be fair, that only happened because I tried it out on Microsoft. There's also a version for Linux, but that would have probably taken longer to set up.
Finally found the knobs to set chromatic aberration and geometry correction, but like with DxO, it relies on lens data—and it didn't have any for this lens. How do you find out which lenses are supported? Ah, that's a well-kept secret. Even the relevant documentation doesn't say, but it looks as if it uses the Panotools lens correction model. And from my recollection, there are very few and random lenses in that database. From experimentation, only 2 of my 5 Olympus lenses are supported.
So I converted anyway. As I had said on IRC previously, I was looking mainly for corrections for chromatic aberration, and also improved shadow detail. I found tick boxes for CA, but not for shadow detail, though it was prepared to give extreme contrast with no shadow detail at all. After Saturday's experiments with the garden north panorama, decided that would be a good one to compare. Here one of the component images, first from DxO, then from Corel. Running the mouse over either image shows its partner to show the differences more clearly.:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly DxO gave better shadow detail. What about chromatic aberration?
|
|
|||||||||
|
|
|||||||||
These images are easier to compare when enlarged to “small”, but it shows clearly that the chromatic aberration is almost gone in DxO, but still very evident with Corel. The difference in position of the antenna is due to Corel's missing distortion correction. But surprise, surprise: the Corel image has much better definition. I suspected that that was due to the pseudo-HDR image processing that I had performed, so I tried it with the standard profiles. Indeed, that was the issue. Here in sequence DxO HDR, Corel, and DxO normal profile:
|
|
|
|||||||||
|
|
|
|||||||||
|
|
|||||||||
It also shows that the detail of the sky is better with DxO, even without the pseudo-HDR stuff. And, as I expected, AfterShot Pro has no advantages for converting raw images. Well, the speed, of course: it's about 5 times as fast. Still, maybe it has other features that would make it useful. I'm not done yet.
|
Internode support going downhill?
|
Topic: technology, opinion | Link here |
My network connection is flakier than ever, so much that I've given up looking at the statistics. Internode support don't seem to be doing anything about it. I've provided them with evidence that it's due to non-responsive cells in the wireless network, but they don't seem to think that that's a problem they can report to Optus. Why? Two years ago I had very similar symptoms, though not the ability to investigate the cause, they contacted Optus and got the thing sorted out pretty quickly. And Max, the technician at that time, knew what he was talking about. In this latest incident, I don't get that feeling, and they claim they're not allowed to talk to Optus until I have changed the modem and “reinstalled” the driver (whatever that means in my context).
|
gnuplot POLA violation
|
Topic: technology, opinion | Link here |
As a result of the network problems, I haven't looked at my network link statistics page much recently. But when I looked today, the graphs were all blank. Why?
Working with gnuplot is a real pain, but I finally got round to looking at it. Date calculations are particularly painful, not helped by the fact that gnuplot timestamps are seconds since 1 January 2000, while Unix timestamps are seconds since 1 January 1970, so there's this continual offset 946684800 (30 years) in the commands. Here's part of a command file I generate for one of the graphs:
The xrange specifies start and end times in gnuplot timestamps, the 39600 is the offset of the local time zone from UTC in seconds, and the first column of /var/tmp/3glinkstats contains Unix timestamps. So what's wrong? Lots of long calculations showed that—not surprisingly—the numbers matched. Out of sheer desperation, started widening the intervals until something came into view. And surprise, surprise! It was interpreting the first column as Unix timestamps and not the old gnuplot timestamps. Remove the - 946684800 and it worked!
But why? No idea. The xrange was still in gnuplot timestamps, so it's not that. Whatever it is, it's a gigantic POLA violation.
| Tuesday, 29 January 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dereel residents want the Radiation Tower
|
Topic: general, opinion, technology | Link here |
Since Wendy McClelland got interviewed on TV a week ago, a number of things have happened, notably the ABC article on the subject. But we were still unhappy that the majority of the residents didn't have their say. Then Greg Nyary, a resident whom I don't know, arranged for the same team to come to Dereel and interview the proponents. It all happened at very short notice, and it wasn't publicized, but I heard about it from two different directions. Today, a workday, at 11:00, not the ideal time to get a crowd. But get a crowd we did: there must have been somewhere between 50 and 70 people there. That's pretty good for word of mouth, and a far cry from the 4 people that Wendy drummed together.
The interviews were well prepared: ordinary people wanting to do ordinary things and being held up by poor Internet access. They interviewed three people: a planner in private practice, a health worker and a bloke from the CFA, each explaining how the current slow, expensive and unreliable Internet connections make their life difficult. Scott and I didn't fit into the “ordinary people” category, so we weren't interviewed.
Unfortunately there's little we can do with it. The newspeople complained about Scott putting up the news broadcast on YouTube, and he ended up having to take it down again. That means, of course, that he's not allowed to put up the new one either.
One other interesting fact that came out of the meeting: Scott tells me that two family members are also having the same kind of networking problems that I've been having recently. We should check the times more carefully.
| Wednesday, 30 January 2013 | Dereel | Images for 30 January 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Computer reorganization: taking the plunge
|
Topic: technology | Link here |
I've been planning to rearrange the machines in my office for months now, at least since I converted my main machine to amd64 over six months ago. There have been many reasons for the delay, but a prime one was the dread of the amount of work and the number of things that could go wrong. But I have new hardware for the machine which will finally allow me to run four displays again on a single machine—maybe. So today I finally reorganized, the first time in the 5½ years I have lived here.
To my surprise, nothing serious went wrong. I started at 10:00, and 3 hours later I had X up and hobbling:
|
|||||||||||||||||||||||
That's only a partial cleanup, of course, of the desk with monitors on it and the desk to the left. The rest needs to be done, but it won't be as terrifying.
The hardware changes included putting in the new nVidia GeForce GT 640 graphics card. To my relief, yes, it can support three monitors, one of them dual-link, so theoretically I can now connect 5 monitors to the machine. But first I need to get the four ones currently connected to work. From experience, that won't be fun, and I was happy enough to continue with the 3 functional monitors for today.
I had planned to put the spare display card (nVidia GeForce 9500 GT) into dxo.lemis.com, and sure enough, in contrast to the new card, it's not too wide to fit. But it is too high. One of the disadvantages of a small machine.
|
The curse of Wendy McClelland
|
Topic: technology, general, opinion | Link here |
While writing my diary for yesterday, discovered a disconcerting problem with the display on my new 27" monitor:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
What's that? It stayed in the same place on the screen while I moved the window, or when I replaced it with a browser window, but it doesn't look like dead pixels. It wasn't until I iconified the screens that I saw what was on the root window:
|
|||||||||||||||||||||||
It's part of the now-removed news item from YouTube, and it was clearly bypassing normal X conventions. When I created another window with a black background, it became clearer:
|
|||||||||||||||||||||||
How did that get there, and why didn't it go away? I was able to get rid of it by switching to a vty and back again. And isn't it amusing that it should be our technophobe Wendy McClelland? You wouldn't need to be overly superstitious to consider it a curse.
|
Network problems without end
|
Topic: technology | Link here |
My network problems continue. This is completely unacceptable. It's been 10 days since I reported the problems and the issue of the non-responsive cell, and I've only had one response asking me to reinstall my software. I've already sent a message expressing my annoyance, but nobody bothered to reply. Called Internode support today and spoke to Goran, who at least contacted the “specialists”, who are too important to talk to me directly. It seems that they did report the problem to Optus (I wonder in what form), and that Optus claims that all is running well and there is no congestion. I wonder how they're measuring that. It seems that they did reset something, possibly the cell in question, but I'm getting very little information out of them.
| Thursday, 31 January 2013 | Dereel | |
| Top of page | ||
| previous day |
|
New X configuration and other surprises
|
Topic: technology | Link here |
Started the day off with the task of getting X running correctly on my four monitors. Tried out the xorg.conf file that I had created yesterday, and how about that! All 4 monitors came up. Not in the correct order, and not at the correct resolution, but it was a good start. In addition, the ΛOC monitor seems to be delivering EDID information again, and it's not the same as the EDID information I downloaded from the web two months ago. Something to investigate when I get the time.
Back to look at the file, and discovered that I had a number of serious errors in it, in particular referring to non-existent devices. How could that have worked at all? Fixed that, and only two monitors came up. The only error message I could find was:
Well, that was on the cards, but I was pretty sure I hadn't changed that definition, and after checking, it was right. And that would only explain one missing monitor, not two. Spent some time scratching my head, and in the end tried the nvidia-settings program, which can also be used to create a config file. It has a nice graphical layout thing, which didn't work for me: it wouldn't let me position the screens where I wanted. Setting “left of” or “right of” placed things at random. There was another button, but it was greyed out, and I didn't look at it carefully.
Went off and did some other stuff, and then tried again. Different results: this time the button was accessible, and proved to be the monitor to which the relationship was to be created, here “X screen 1” at bottom right:
|
|
||||||||
With that, I was able to save to X config file. Well, almost:
|
|
||||||||
Pressing “Show details” gave me a string of errors like:
That's not my fault. It was generated by nvidia-settings. But “Auto Fix” worked, and I had a valid X configuration file, and I was able to start X and use all 4 monitors in the correct resolution. Valid it was, but not correct. Despite the claimed layout, monitors 0 and 3 were transposed, even in the sequence in the config file. OK, that's not too difficult to fix:
Tried with that, and...
And again only two monitors. I get the distinct impression that something in the startup is ignoring or misinterpreting large chunks of the config file. Tried lots of different options, but I couldn't fix it.
But there's a workaround, or there should be. The computer is now next to monitor 0, and the only cable long enough to reach monitor 3 was a VGA cable, so I put monitor 3 on VGA. Not nice: it flickered, and it was diametrically opposed to dxo, for which it is also (via KVM) the monitor. It would have been nicer to put monitor 0 on the VGA output and monitor 3 on HDMI, but the HDMI cable wasn't long enough. And then it occurred to me: 3 years ago I bought a 3 m HDMI cable, which I had never used. And I was able to find it, and it was long enough for monitor 3. So I swapped them over, and with a shorter VGA cable got a much better display on monitor 0, as well as having it more convenient for dxo. That doesn't excuse the nVidia bugs, of course.
Another bug is benign, but confusing:
I thought that this was something to do with the nVidia software, but it proves to be something that just happened to be in my home directory, as it happens a description of problems I had with X nearly 11 years ago. And I've been having fun with X configuration for over 20 years. Some things never change.
I did my testing by starting a second X server, which meant that I still had X running when things went wrong. That works relatively well, but there seems to be a race condition with the mouse. In about half the cases, I got this message:
And no mouse at all, of course. In general the next attempt worked.
There were other random things that I still need to follow up on. The sound output doesn't work, and testing was not helped by the fact that I can never remember which jack to plug the connector into. But after trying all, I couldn't get any sound out. Jürgen Lock suggests that this might be because of the HDMI output on the new graphics card, which may also imply sound hardware, but I'll investigate that some other day.
|
DxO screen refresh insights
|
Topic: technology, photography | Link here |
Had to do some photo processing today, while still playing with X. For some reason I brought up my rdesktop window on another monitor. And the DxO Optics “Pro” screen refresh worked! It seems that the problem is related to the resolution: normally I run it on the 2560×1440 monitor. I wonder if the problem exists with higher resolution displays locally as well. I can't test it, because the monitor has only DVI input, and dxo, the Microsoft box, has only a VGA output.
|
Rain!
|
Topic: general, gardening | Link here |
For the first time in ages we've had a significant rainfall—15.8 mm. The people in Bundaberg would laugh, of course, but it's four times as much as the previous total rainfall for this year. And of course rainfall means power cuts, sort of: this time it was so brief that only the clock in the bedroom failed. But there were presumably a number of power surges. The UPSs screamed, and a switch wedged. Solar power sounds like a better idea all the time.
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 |
