|
| Greg's diary |
| January 2009 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Thursday, 1 January 2009 | Dereel | Images for 1 January 2009 |
| Top of page | ||
| next day | ||
| last day |
| Topic: general | Link here |
Today the notice ran out for the demolition of the swallow's nest, and we removed it. In the process, discovered that there were five eggs in it:
|
|||||||||||||||||||||||
That's a little disappointing; I deliberately waited until the birds were out of the nest before removing it, but I wasn't expecting another set. Still, they had to go sooner or later. I don't think that they're in any danger of extinction.
| Topic: gardening | Link here |
The Buddlejas have marked the New Year by suddenly springing into bloom. Yesterday there was almost nothing, and now there are many balls:
|
|||||||||||||||||||||||
|
Scanner pain
|
Topic: technology | Link here |
Spent some more time looking at alternative software for my scanners, running into all sorts of problems. Downloaded a 50 MB “demonstration” of SilverFast, only to discover that the demonstration didn't include 48 bit support. I had the option of downloading the plugin, but currently boskoop, my Apple, is off the net because of the bug in the networking stack which makes it send out packets with the wrong return address. I can't even just change the address, because then it wouldn't work on the local network.
The real answer is getting NAT working properly on dereel, of course. Until a couple of months ago, that's what I did on eureka, but eureka is no more, and I've rather rashly overwritten the backup disks. I still have the system disk, so I'll have to put a machine together to do a new backup.
Turned my attention to SANE, which, to my surprise, installed without any issues. Getting it to run was another matter. The system didn't like the scanner:
I suspect that this might have something to do with the motherboard, but I'm not very happy with the idea of experimenting with dereel, especially since it's currently doing a level 0 dump of /home, taking an estimated 42 hours, and I don't want to have to restart that. So dragged out eucla, my laptop, which hadn't run under that name since 26 May 2007—it has been running Microsoft under the name pain most of the time—and started upgrading the system. That didn't finish today, of course.
| Friday, 2 January 2009 | Dereel | Images for 2 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
My new glasses (well, the lenses) are ready, just one day too late to get them on last year's allocation from my medical insurance. Into town with Yvonne to pick them up and also to look for a new mixer. As I feared, the Kenwood mixer like the one I borrowed from Chris cost between $400 and $500, and the cheapest that I could find that looked even marginally likely to do the job cost $200. I still need to find a better solution.
Visit from Chris' sister Melanie and her husband Tony and son William in the afternoon:
|
|||||||||||||||||||||||
| Topic: gardening | Link here |
My uncle Max gave me a staghorn fern for my last birthday. It was destined for the covered part of the verandah, where the swallow's nest was until yesterday, and in the interim we had put it in a place which proved to be too hot and windy. It's in its final place now, but it's not looking at all happy:
|
|||||||||||||||||||||||
Hopefully it'll recover now that it's in a better position.
| Saturday, 3 January 2009 | Dereel | Images for 3 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
More of these green beetles have shown up, and I tried to take photos of them. It's not easy with a manual focus lens (here my 50 mm f/1.4 Super Takumar with an extension tube):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I think one of my next acquisitions should be a macro lens with autofocus.
|
“Road Toll”: Bad language, lies, damn lies and statistics
|
Topic: opinion | Link here |
There's a possibly more up-to-date version of this rant here.
The Christmas holiday period, characterized by a morbid attention to road traffic accidents, has just finished. It seems that Victoria took first place in the “road toll” statistics. Given the sums that CityLink charges, I could believe that—if you miss a sign on Batman Ave. in the south of Melbourne's Central Business District, you'll owe CityLink about $13, even more if you don't find out in time and they have to come and get you to pay. But “Road Toll” has nothing to do with toll roads: it's more bad language, the term that the traffic authorities in Victoria use to mean “road accident deaths”. This seems to be a uniquely Australian and New Zealand abuse of the term.
Road accidents are sad, and it's understandable that the authorities are trying to do something about them. But do their attempts make sense? Victoria already spends more police time than any other state in attacking the perceived causes of these accidents. It's clearly not working. What are they doing?
This is a long article. Here are the key points:
According to the reports, Victoria had 304 road deaths in 2008, or 0.83 per day. Over a 14 day period, that would be 12 deaths. In fact, over the “road toll period”, which I assume to be 14 days, 16 people died. Four died on the last day, so for the rest of the period, there were 12 deaths in 13 days, or 0.92 per day. Is this even statistically relevant?
The Victorian Government has created a new 10 year plan called Arrive Alive—certainly an idea with which everybody should agree. But sometimes I think they place more emphasis on “alive” than on “arrive”. The main methods that they advocate—if you can believe the summary on the home page—are limiting the activities of inexperienced drivers (only one passenger allowed) and restricting speeds (thus making it more difficult to arrive, and discouraging people from travelling). No mention of alcohol, no mention of safe driving practices such as adapting to the road conditions or keeping your distance from the car in front of you—a thing at which Victorian drivers are not good.
On 11 December 2008 I drove from Briagolong to Dereel and saw a sign on the central verge of the Gippsland freeway. It read something like: “Speeding: our number one killer”. Unfortunately I wasn't able to get a photo. But what stupidity! I'll get back to that later on.
The government has found a way to blame all issues onto “speeding”, which I interpret as exceeding the legal speed limit—certainly that's what they mean when you get fined for “speeding”. Clearly just about any accident (apart from a few freak exceptions) is going to be made worse with increased speed. But speed is just one of the issues and almost never the cause. My aunt Ursula, a dentist, once said “in case of doubt doctors always put ‘heart failure’ as cause of death—every single person who dies experiences heart failure”. In the same way, every traffic accident involves speed, even if it's less than 1 km/h.
But if you believe official propaganda, you would think that speed is the most important cause. Much of it is misleading. At http://www.vicroads.vic.gov.au/Home/RoadSafety/SpeedingAndSafety/TheFacts.htm you can read:
Small speed reductions save lives
Research shows that the risk of injury in a crash doubles with a 5 km/h increase in travel speed. Travelling at 65 km/h in a 60 km/h zone, you are twice as likely to be involved in a crash. A car braking from 65 km/h will still be travelling at 32 km/h at the point where a vehicle braking from 60 km/h has stopped.
I have a number of issues with these two statements. Firstly, they're unrelated. The first refers to injury and death, as if they were interchangeable, and the second to speed. In this case, they're barely related, as I'll explain. For the first claim, which research shows it? Without substantiation, this statement is barely more useful than the claims that Wendy McClelland makes about the Dereel phone tower.
The second claim is the real issue. It's extremely misleading. It bases on the complicated relationship between speed and distance when braking. At the speeds mentioned, your speed only drops significantly towards the very end of the braking manoeuvre. It's important to understand that, but this document tries to make it into a life-saver. It isn't.
The time it takes to come to a stop depends on how heavily you brake, of course. With tyres (friction) the limit deceleration is the standard gravity of gravitational coefficient 9.81 m/s². A more attainable deceleration would be about 8 m/s². At the speeds mentioned, the bulk of the distance travelled is taken up by the reaction time.
Taking these values, and with the help of a little program, it's relatively easy to come to the following results:
So at 60 km/h it takes you a little over 50 metres to stop. At 65 km/h it takes you 5.8 metres, or about 11% more, to stop. How fast is the vehicle going in the second case 6 metres before it stops? Again, the program can help. “Sample distance” is the distance at which you sample the speed; if I had been more interested, a graph might have done a better job.
That's quite close to what they claim—in fact, they claim less than what my program calculates. So what's the issue? Look at the other values. In each case it will take a little over 4 seconds to stop. In that time you'll travel over 50 metres. The fact that your speed is still relatively high 6 metres before you stop pales into insignificance when you consider how far you travel before you stop.
Presumably you've learnt to drive, very probably at a driving school. And your driving instructor will have reminded you again and again that braking is not the way to avoid accidents. This discussion shows you one of the reasons. My personal horror scenario is driving down a residential road with oncoming traffic and cars parked on either side. Then a child runs out 5 metres in front of me. I can't swerve, because there's nowhere to go. Can I brake? Sure. What will the result be? I'll hit the child.
For reasons like this many residential roads have speed limits like 30 km/h, or (in Australia) 25 km/h. I've even seen 15 km/h. Will they help? Here's the program again:
Independent of the speed I'm travelling at, I'll hit the child before the reaction time expires. Braking doesn't work. Keeping people off the roads does, but it's still far dangerous than “speeding”.
But this is all not the point: how often do you avoid an accident by coming to a complete stop? Almost never, and in those cases where you do, you're probably going very slowly. In VicRoads' example, it takes over 50 metres to stop from 60 km/h. Most surprises are much closer. As any driving instructor will say, braking is the last resort. Consider the situation at higher speeds: at 100 km/h (still allowed in many parts of Victoria, even in the residential area outside my house), your total braking distance is over 100 metres. And at 240 km/h (yes, that's legal in some countries, for example Germany) it's over 400 metres. So at any speed, you can't rely on stopping to prevent accidents.
VicRoads continue with the next claim:
Speeding saves little time
It's true! Speeding is a major contributor to Victorian road deaths and trauma, and yet brings about only minor reductions in travel time. On a 10 km journey, you would save 46 seconds by increasing your average speed from 60km/h to 65km/h, but you double your chances of being involved in a crash.
If you accept the previous claim, then this is a logical consequence. But is that the issue? In areas with a 60 km/h limit you're unlikely to be able to go much faster anyway. The real issue is on long distances between towns. Most Australian highways are very good, there's little traffic, and I find it possible to keep up an average only slightly below the speed limit. I can usually manage the 620 km from Dereel to Echunga in 7 hours. Of that, about 90%—let's say 550 km—is on open roads where I could equally well do 130 km/h, in some places over 160 km/h. At 100 km/h , this part takes 5½ hours. At 130 km/h it takes 4¼ hours—a saving of nearly a quarter.
All we've seen so far have been claims based on shaky theories. What really kills people? If you're drunk and you drive off the road, does the accident get blamed on speed or alcohol? What if your car is in poor condition and a wheel falls off?
The report in The Age gives some details about how people died, unfortunately not for all of them. But in no single report did anybody blame excessive speed as the cause of death. Collisions with oncoming cars or trees are surprisingly common. It's difficult to see how speed can be the main cause of this kind of accident.
The Australian Government maintains a Fatal Road Crash Database, but unfortunately it doesn't seem to contain any indication of the cause of the accident. Maybe they don't keep this information. But without doing so, you can't analyse the problems.
There are many forms of dangerous driving, and I'm not going to go into them all here. Many are exacerbated by excessive speed. This has nothing to do with speed limits. Safe speeds depend on many factors, and they certainly don't come in steps of 10 km/h. For example, traffic conditions greatly influence the safe speed. If visibility is down to 50 metres, you can't safely drive beyond 10 km/h, even if the speed limit sign say 100 km/h. If the road is icy, you need extreme caution (on the Western Highway a sign recommends 40 km/h, which is far too high for these conditions). But what about your position on the road? If you're less than your reaction time behind the car in front of you, and he brakes suddenly, you'll probably hit him. Again the value “2 seconds” comes into play. But on those same freeways where you'll get fined for doing 107 km instead of 100, people continually drive with distances of less than a second from the car in front of them. The police are out there looking for traffic offenders—hundreds of people can confirm that—but they don't seem to worry about keeping your distance.
Drive anywhere in country Victoria and you'll be faced with a barrage of signs with buzzwords like “ONLY SLEEP CAUSES FATIGUE” and “OPEN YOUR EYES” and “TIRED? MICROSLEEP NOW”—or something like that. If “speeding” is the number one killer, it seems that going to sleep at the wheel is number two. But why? I haven't seen this identified as a serious problem elsewhere. And the interesting thing is, yes, I find myself getting drowsy when driving in Victoria too. The long, empty roads are hypnotic at the speeds we're allowed to drive. I've driven similar distances in other countries and never had this kind of issue. Until proof of the contrary, I assume that this is one more negative side of low speed limits.
Another report mentions statistics for various kinds of traffic offence: “more than 900 drink-drivers were caught across the state”, but the same operation caught 8247 “speeding” motorists. There is no mention of anybody being booked for driving too closely to the vehicle in front of him. Clearly it's easier to catch “speeding” motorists.
Victoria may have some kind of record in obsession with speed, but it seems to be an anglocentric problem: the USA and the UK are not much better. In continental Europe, things are different. Despite significantly higher traffic densities, speed limits are considered secondary to other traffic behavioural issues. In Germany there is no general speed limit on the freeways. People regularly drive at speeds round 200 km/h, and the slowest cars drive at about 120 km/h. Some lanes on German freeways have minimum speed limits of 120 km/h. If the policy makers are right, the German freeways should be a graveyard.
In fact, the documentation shows that road deaths on German roads are somewhat lower than in Australia, The Australian Government produced a document comparing the road deaths in Australia with other OECD countries in 2006. The results:
| Australia | France | Germany | Victoria | |||||
| Deaths per 100,000 people | 7.7 | 7.7 | 6.2 | 6.6 | ||||
| Deaths per 10,000 vehicles | 1.1 | 1.3 | 0.9 | 0.9 | ||||
| Deaths per 100 million km | 0.8 | 0.8 | 0.7 | 0.6 |
So despite the draconian speed limits, a higher percentage of the Victorian population dies on the roads than in Germany. That would be surprising if the speed limits were the same: the freeways in Germany are very congested, and it leads to bad behaviour on the roads. The other two values are more indicative of the differences between Germany and Australia: people here obviously have more cars, and they drive further.
It's worth considering the implications of these statistics. The Germans obviously have something going for them that the Victorians don't, and by emphasizing the impact of “speeding”, the authorities are de-emphasizing the real issues.
This doesn't stop people saying the opposite. A few years ago in The Age a reader discussion contained the claim:
It is a significant fact that other countries with higher speed limits have higher road tolls.
Certainly this discussion shows a number of people in favour of the speed limits. Brainwashing works again.
| Sunday, 4 January 2009 | Dereel | Images for 4 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, photography | Link here |
Spent much of today writing up my rant about “speeding”, but also found time to do more work on my photo display page. One of the issues I have had for a long time is that the “big” photos (original size) take a long time to load, so I can't just raise the size of all photos on a page. But the way I solved this was to link to a page that showed all photos of a single day, so (for example) the first photo from 2 August 2008 in the kangaroo album would link to all photos taken on 2 August 2008, not quite what I wanted. After solving the problem of changing the size of individual photos a couple of days ago, it was relatively simple to add larger photos, though there will still be a while before I'm happy with the result, and doubtless I'll find some bugs.
| Topic: photography, gardening | Link here |
More work in the garden today, mainly spraying weeds. There are many more of those green beetles, and this time I caught some and put them in the fridge. After the problems I had with the giant ant, I was careful not to undercool them, but instead put them on a plate that had been chilled in the deep freeze. That worked: they stayed still for about 10 minutes, long enough that I thought that I had killed them after all, but then they woke up. In this case, since they're pests, I then had to kill them.
Things still could have been better: when they're knocked out, they stretch out their legs and fall on their backs, and it's almost impossible to put them in a natural position. Tried with some cotton wool, but the results were only mediocre. At least the focus was OK, and I could confirm the shape, also that some are golden rather than green:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Monday, 5 January 2009 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
Yvonne got a letter from Laurel Gordon today, including a packet of snowdrops and some seeds for a cream-coloured perennial foxglove. Spent some time planting the former, but growing the seeds is more complicated than I thought: one description mentions that you need to keep the seeds uncovered, since they need light to germinate, but it also mentions not keeping them too warm—and this is a British description, so it'll definitely be too warm here, where we're currently having daytime temperatures round 30°.
| Topic: general | Link here |
Spent a disproportionate amount of time today looking for a new kitchen mixer. It really looks like the Kenwood Chef is the only game in town, but they're very expensive—even a model like Chris' A901E, which must be 20 years old, sells on eBay for over $200. As we saw a couple of days ago, you can get a brand new model for $400. So what's the difference?
Went to the Kenwood web site to look, and found little to help. If their products were of the same quality as their web site, I would have looked at alternatives. The front page has a flash-only component that fails silently on my combination of firefox 3 and Flash 10—something more the fault of firefox than Kenwood, except that Kenwood doesn't offer an alternative.
Took a look at the Support tab, in the hope of finding documentation, but there was none. Somehow found my way to the “Products Kitchen” page, which showed a number of categories, including “Accessories - Chef”, but no “Chef”. It doesn't take much intelligence to work out that it's probably under “Food Mixers”—but no such luck: food mixers were only hand-held.
|
|
||||||||
In the meantime, looked at the numerous accessories—34 of them, displayed on 7 pages, 5 at a time, using only 12% of the window area, and in no obvious order:
|
||||||||||||||||||||||
|
|
||||||||
Finally found a browser that would display the flash on the front page, and found a link to a catalogue in PDF format. That was minimal, but gave me some information, only about current models of course. But the quality! Some of the images were almost unrecognizable, like this one on page “08”:
|
|
||||||||
It's nearly 10 MB in size, so you'd think they could put in better images than this.
With a lot of reading between the lines, it seems that the current models are the KM 300 and KM 315, the ones we saw the other day, and that about the only differences are the finish and the price: the KM 315 has a metal bowl and housing, and costs $100 more. I still have no idea how the other models (notably the 700 and 900 series) compare, except that Chris' A901E has a 350 W motor, and the modern ones have a 700 W motor. The accessories—one is a grain mill, which I could use for my brewing—are also specified to work with the new models, but they don't say whether they'll also work with the old ones. That seems to be significant enough to look for a new model, particularly since they seem to keep their value so well.
Further searching found a link to a company called Grays offering refurbished KM 300 and KM 315 models with 6 month guarantees at $204. It turned out that the KM 315 came with no accessories, so the KM 300 seemed a better choice. Signed up with them, retrieving their confirmation email from my SpamAssassin folder:
It's a strange site—finding your way around is almost impossible, and I don't think I would have found the auctions directly from their home page, but I managed to place a bid and win the article for $204. Presumably the poor navigation puts them at a disadvantage even in comparison with eBay
| Topic: general | Link here |
Dave Pinder along this afternoon with 70 bales of hay—despite our intentions, we didn't get anything worthwhile this year.
| Tuesday, 6 January 2009 | Dereel | Images for 6 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
Somehow I'm still involved with researching Kenwood mixers. It started when Grays sent me a confirmation of having won the “auction” for the mixer. The stated price was $204, but they added a “buyer's premium” of $30.60 on top. Further examination of the web site confirmed that they apply a 15% surcharge on all auctions in this category, and other rates in other categories.
Is this legal? I thought that advertised prices in Australia had to show the final price, unlike in the USA, where “applicable” tax is added. I'll find out about the legality; it makes Grays sound very dubious to me, though.
Spent more time researching the Kenwood web sites, and found a number of things:
I mentioned yesterday that I couldn't find a link to “Chef” in the products page. Today I discovered that they're under “Kitchen Machines”—or some of them are. The KM 315 is there, but the KM 300 isn't, though both are in the catalogue. My guess is that the web master has missed out a couple of products.
This problem doesn't exist on the UK web site. The Products page is much easier to read, and includes “Chef and Major under “Food Mixers”, as you'd expect.
The products in the UK are very different from those in Australia. On the positive side, they have more powerful motors, between 800 W and 1400 V, compared to 700 W for the Australian models On the negative side, only the most expensive have all four power takeoffs. I suppose it's always been “normal” that companies develop different products for different countries, but is there really a justification for it nowadays? It certainly makes it a lot more difficult to match accessories.
There are accessory compatibility issues between the models. Found a mincer attachment on eBay with the comment:
Note that this mincer has the small locting pin at the bottom where it fits into the front of the mixer (see photo) . Some (later) models of Kenwood Mixer require attachments with the locating pin to the top where it fits in the mixer.
I checked Chris' machine (A901E, still here, making bread dough for the last time), and noted that it wouldn't fit her model. It almost certainly won't fit the KM 300. With a bit of guesswork, came to the conclusion that there are compatibility issues between accessories with A7xx and A9xx model numbers. So it's a good thing I didn't buy an A700.
The UK web site is better than the Australian one, but it still shows lack of care. One of the FAQs states:
That's incorrectly formatted, of course, and so it displays as:
What are the maximum capacities for my machine?The maximum capacities are as follows. Do not attempt to exceed these as it could seriously damage your machine. Chef Shortcrust pastry - 680g flour Fruit cake mix - 2.72kg total mix Stiff bread dough - 1.36g flour Soft bread dough - 1.3g flour Egg whites - 12
That sounds very dubious. 1.36 g flour? I'd have difficulty mixing that little by hand. Clearly they mean 1.36 kg. Or do they? Why should you be able to make more stiff dough than soft dough? Maybe they really mean the same quantity, but in the first case they did an overly precise conversion from lbs to grams (3 lbs is 1.3608 kg, and 6 lbs is 2.7216 kg, also represented here). The real problem here is that the information is so inaccurate that you can't trust it.
| Topic: opinion, technology | Link here |
It's been about 20 years since I started using the Internet directly (as opposed to my prior email access via Tandem's SMTP gateway). It quickly became a way of life. Then the web came, and we were allowed to do commercial things on the Internet (“Death of the Internet, film at 11” as the old die-hards said). Nowadays even the Australian Government considers “broadband” access a right of all Australians.
But the Internet is still in the future, somehow. It's used by teenagers and “alternative” people, and real people don't use it. Certainly our politicians don't, or they'd come out with considerably better thought-out bills. Those who do use the Internet mainly see it as obfuscated by Microsoft.
Currently the Australian Government is showing its complete lack of understanding by bringing out a bill to censor the Internet. The industry has reacted strongly, of course, and it seems that Mark Newton has been particularly active, to the point where the entire industry is up in arms. Stephen Conroy, the current IT minister, has personally singled out Mark as an evil-doer. There's a good summary of the incident, so I won't go into detail here.
Mark's letter is a little more aggressive than necessary:
SENATOR CONROY HAS INVALIDATED THE GOVERNMENT'S CLAIM FOR A MANDATE BY LYING TO THE AUSTRALIAN PUBLIC ABOUT THE SCOPE OF THE POLICY
In December 2007, Senator Conroy was widely reported in the press as stating that “... anyone wanting uncensored access to the internet will have to opt out of the service.” ... In October 2008, Darren Pauli from IDG/Computerworld obtained confirmation from the Minister's media spokesman that no opt-out will be available, “... the filters will be mandatory for all Australians.” As far as politicians' lies go, this one is breathtaking in its audacity.
I can't agree with Mark here; things change with time. This letter was written to his Federal representative, so he's implicitly insulting her too, which is counterproductive.
Mark mentions the Broadcasting Services Amendment Act, passed 10 years ago, which failed dismally, but he doesn't highlight it. The big thing that he has drawn attention to is the use of a VPN. He's right, of course, but I can only marginally agree with him there, too: you don't need a VPN to bypass censorship: an ssh tunnel is easier to implement and just as effective. It's encrypted, and it changes the port numbers, making it impossible for any sniffer to determine the content. I describe how to set them up in The Complete FreeBSD on page 424.
I send all my mail via an ssh tunnel to an external server—not to break the law, but out of convenience, and (marginally) to protect myself from lawbreakers. I perform my web site updates with rsync and ssh, which has similar properties. Given the anarchy on the net, you'd have to be crazy to transfer any kind of sensitive data (at least passwords) in the clear. So it's only natural that pædophiles and terrorists should use encryption. I should be surprised that there are any left who don't, but I've also ranted enough about the common perception of Internet security, even with people who should know better, like financial institutions.
But people have gone into much more detail than I, with much better researched information. From my perspective, though, it shows that Western society as a whole doesn't understand computers. I rant on this topic continually—the Kenwood web site is the most recent example. People still don't take the Internet seriously. I don't know of any company that has both a web site and printed catalogues or brochures where the web site is of as high a quality as their printed documentation. From the point of view of politicians, they continually look on the Internet as a separate space, requiring separate laws, and they continually overestimate their own ability to enforce the laws that, in their ignorance, they pass. In a time so influenced by the Internet, you'd really be able to expect that the Government understands it. I wonder what computer skills Senator Conroy has. At the very best, I'd expect him to be able to operate “Internet Explorer” and Microsoft “Outlook”.
The big problem with censorship is that the Government is trying to block traffic that people want to perform. They'll find ways around it. They can't even stop spam, traffic that the receiver doesn't want to have; that would be considerably easier to achieve, not with technological tools, but with existing laws. But even that's too much trouble. Why should censorship work?
| Wednesday, 7 January 2009 | Dereel | Images for 7 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
Another short power failure in the middle of the night. I wish they'd get round to burying power cables.
Somehow didn't do much today; writing diatribes doesn't leave much time for other things. I'm trying to build a new system on eucla, and that's taking much longer than I expected, especially since so many tarballs need to be loaded, and my new network tariff gives me most of my bandwidth in the morning.
|
Empire building, SBS style
|
Topic: food and drink, opinion | Link here |
Despite everything, we still watch SBS Food Safari. Some of the recipes are interesting, but the extreme inaccuracy of the claims is irritating. I've already quoted the “French” chef de cuisine who claimed that Speck is a French ingredient. Today they broadcast an episode about the English cuisine, unfortunately without anything new. Or maybe they did: Rollmops. They even go to the nonsensical claim:
Without roast beef, pork pies and rollmops, empires would never have been made.
I'd be more inclined to believe Wikipedia, which states:
Rollmops grew popular throughout Germany during the Biedermeier period of the early 19th century and were known as a particular specialty of Berlin,
Rollmops are found in Dutch, Scottish, German, Latvia, Scandinavian, Czech, Slovak, Polish, Brasil and South African cuisine, among others.
But where do the SBS people get these crazy ideas from?
| Thursday, 8 January 2009 | Dereel | Images for 8 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, opinion, photography | Link here |
Over the last few months, photo postprocessing has established itself as an issue that I haven't got under control yet. I've already changed from JPEG to raw format on my Olympus E-510, and I've tried a number of conversion programs, including dcraw and ufraw, but things still don't seem to be quite right. Strangely, none of this seemed to be a problem with my compact Nikon L1. Today I went and did some comparisons. It took all day, and at the end all I knew was that there are dozens of different “standard” conversions, and that Olympus' own OLYMPUS Master 2 is so horrible that it would drive me mad to have to use it.
First, the photo: of my new verandah. Here's what the Nikon did out of the box:
|
|||||||||||||||||||||||
With the Olympus I can take photos which are stored both in raw and JPEG format. The raw image also has a JPEG “thumbnail” that would blow even my generous definition: it's 1600x1200 pixels, or nearly 2 MP. The thumbnail can be extracted with dcraw (-e flag), and with the supplied software the raw image can also—with considerable pain—be converted to a JPEG. The results for one single photo—from left to right the thumbnail, the original JPEG, and the JPEG converted from the raw image using the Olympus software—look like this
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Compared to the Nikon, the most obvious difference has nothing to do with the image quality: the beam at the top is straight after all. But all three images are less saturated than the Nikon, and there are subtle differences between the three.
By comparison, though, the conversions of the free software packages leave a lot to be desired. dcraw (left) has no significant way of changing the appearance, and ufraw seems to have changed its mind some time between release 0.13 (middle) and 0.15 (right). All three images look significantly different from the original:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, you can twiddle knobs, at least with ufraw, but how? And, in particular, why? Did some investigation, but my guess is that not many Olympus users use ufraw, and that the settings are not optimal. I'll need to invest yet more effort to get this working properly.
Today was the first time I tried in earnest to use the software that came with the camera. Possibly the best thing I can say about it is that I didn't have to pay anything for it. It gave me yet another reason to not want to use modern toy GUI software.
It started off with the inability to read the screen—the text is too small at 1600x1200. This may be a general issue I have with Mac OS X: I can't find a way to change the size of the text in most windows. I had to drop the resolution to a very cramped 1024x768 to be able to read things.
GUI software is intuitive, right? I suppose that depends on what you mean by “intuitive”. My difficulties started with downloading the photos from the camera and continued while trying to process the raw images. Still, there's a “Quick Start Guide” which looks like the results of a class project of a ten-year-old: all the items are formulated as questions, there's an unnecessary dynamic image in the top left of the too-small, non-resizeable window, which uses 100% of CPU time on my machine even when idle, and the actual information is in a minuscule window 5 lines high by about 40 characters wide, equipped with a scroll bar, and requiring 4 different windows to read present what proved to be a bare-bones description. The following
|
||||||||||||||||||||||
|
|
||||||||
One minor labour-saving device was the text:
|
|
||||||||
But how do you do that? It doesn't say. Clicking on one of the unnamed images at the bottom of the screen brings them up as a larger image, after a delay of about 10 seconds:
|
||||||||||||||||||||||
I had to look in “Help” to find the answer. “Help” isn't much help, but it displays in a Safari window in overly large characters, and gave me at least this answer (shown in the image above, if you magnify it a couple of times): the little box at bottom left of each thumbnail is a selection box, and clicking inside it toggles a tick mark.
There are many puzzling things about this display, though: why does the “Navigator” (which I haven't been able to use for anything sensible) display a different photo, and on its side? Why does it display the images multiple times? What are the names of the images? We can see name of the one on the screen at bottom left, but what are the other ones called?
I have partial answers. It seems that the software stores copies of the raw images before converting them to JPEG (a function that appears to be implicit). Next time you come along, it shows both the original and the copy, though it's almost impossible to decide which is which. The one on the screen is called P1085841_1.ORF, which implies that it's a copy of P1085841.ORF. That's not as silly a name as it might appear: the first digits is the month (in hexadecimal), followed by a two-digit day of month and a four-digit absolute sequence number.
The problems are at least that I can't distinguish between the original and copy of the raw image (they're carefully marked RAW so that you can at least distinguish them from the JPEGs). And since I didn't need these copies (over 10 MB each time), I removed them—but the software doesn't seem to have noticed.
To perform a conversion to JPEG, you select the floppy disk (remember them? one of these JPEGs would fill about 4 of them) icon at the top. You get another window offering to store the images in “New Album”:
|
||||||||||||||||||||||
|
|
||||||||
Why does “New Album” have photos in it? I haven't been up to the pain of trying to work out what file naming straitjacket this software wants to impose. find tells me that the original photos are stored in /Users/grog/Pictures/OLYMPUS Master 2/20090108, a directory name directly compatible with my own usage (modulo this stupid directory name with spaces in it). But it stores thumbnails, it would seem, with file names like /Users/grog/Library/Application Support/OLYMPUS/OLYMPUSImageDB/Thumbnails/Users/grog/Pictures/OLYMPUS Master 2/20090108/P1085842.JPG.middle.jpg, and the Albums appear to be described in files with names like /Users/grog/Library/Application Support/OLYMPUS/OLYMPUSImageDB/ALBUM/AlbumID_1231388939.243894, which have a format even more unrecognizable than the name, which proves to be a time_t timestamp. Clearly “you are not supposed to understand this”. One way or another, this isn't the way to keep track of the 15,000 odd photos that I have.
But this screen shows more strangenesses: it suggests a ppi (pixels per inch) value. But what does that mean? Nothing until you display the image, and I'm not displaying it yet. Even then, it only makes sense if you know the size you're displaying. I've seen this nonsense with Photoshop, but I still don't understand what kind of thought processes have caused it to happen.
Then I try to save the image. The default image name is the same as the name of the JPEG image from the camera! Isn't that clever? You have to change the name to something else to keep both of them. While messing around, I ended up with the following directory contents:
=== grog@boskoop (/dev/ttyp2) ~/Pictures/OLYMPUS Master 2/20090108 60 -> l
By this time I had moved the camera JPEG images out of harm's way, but they had names like P1085841.JPG. That does conflict with P1085841.jpg, because of Apples decision to make file names case independent. But here it has created files with a sequential number in the middle, and of course hasn't adjusted the time stamps to the time stamp they were taken. This also shows that the download software has no idea of UNIX permissions, and has set execute and write for everybody.
At the end, both the software and I were thoroughly confused. It clearly didn't like me taking the easy way out and moving the files myself. It didn't see files that were there, and it saw files that were not there. I gave up trying to convert some of the photos, because it didn't see them.
As if that weren't enough, I hate the silly terminology they use. Why do they shout RAW? It's a word, not an acronym. Why do they use the word “develop”? This is processing, not development. Even the analogy with film is broken, because you'd have to think of the raw image as a negative, which has already been “developed”. And titles like “how to Enjoy Slide show?”, apart from strange capitalization and dubious grammar, miss the point: I assume that the heading explains how to create one. I'm always left with a bad taste in my mouth when I use this kind of software.
About the only thing of interest was that the software showed up a minor bug in one of my scripts: it's the only software that I have seen that includes the time of processing into the EXIF data:
| Topic: opinion | Link here |
The battery in my pocket calculator is finally getting so weak that it's difficult to read the display. I bought it for Yana at Frys in Sunnyvale years ago, intending it for her studies, and even when I bought it it was clear that something was wrong with the battery. Under the circumstances, it's a wonder that it lasted as long as it did.
ALDI had scientific calculators on special for $7, so bought a couple of them. They actually look quite similar to my old one—possibly they're made by the same manufacturer. And, like the old one, there's an ON key, but no OFF key. Went looking in the single A3 sheet excuse for documentation, and finally found, in an unlikely place:
|
||||||||||||||||||||||
This seems to be the only place they mention turning the calculator off—when the battery is already dead! It works, of course. 2nd F is the yellow button on top left:
|
|||||||||||||||||||||||
The only problem is, that's the wrong calculator. It's my old one, and the corresponding button on the new one is marked “Shift”. And Shift-ON doesn't do anything useful. Finally noticed the marking above the AC button:
|
|||||||||||||||||||||||
You press Shift-AC to turn it off. So the instructions apply to the wrong calculator. Still, it works, which is more than I can say for most of the ALDI stuff we've bought lately. Today Yvonne took back a bread machine, a mixer, two doorbells and a dual DVB-T tuner. The refund will keep us in food for a few weeks.
| Friday, 9 January 2009 | Dereel | Images for 9 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: photography, general | Link here |
Another day spent mainly writing up diatribes about bad software. The sad part about it is that I am now left with a choice: suboptimal conversion with software that is not too painful to use, or this horrible Olympus stuff, that at least seems to give correct conversions. Researched some alternatives, but there really doesn't seem to be much around that doesn't require lots of mouse pushing and manual labour.
| Topic: gardening | Link here |
A bit of work in the garden in the afternoon.
| Saturday, 10 January 2009 | Dereel | Images for 10 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, opinion, photography | Link here |
Photo day today. With my new-found discoveries, how was I to process the images? Clearly the results from the Olympus software are the best, but using it makes me want to tear out my hair. Decided to try the Microsoft version to see if it was any better. I couldn't find the CD I got with the camera—it's probably buried somewhere in the mess in my office. But that's OK: there's almost certainly a newer version, and there's a free download page on the web. But that wanted my contact details and a name, and sent me a mail message.
The mail message tripped over our greylisting scheme, and didn't want to try again. So I tried re-registering; but no, it told me I had already registered. What would have happened if I had lost the message? Probably what I did: I registered under another name and got sent another message. But that one didn't arrive either; further checking of the mail logs showed that it got through the greylisting and got eaten by SpamAssassin:
How can people rely on email any more? This one was only HTML, and not even correct HTML at that, and they didn't have any reverse DNS. In addition, the URL was open, without any key, and there was nothing saying I couldn't redistribute the URL.
Running it on Microsoft has one advantage: I can use it on my FreeBSD system via rdesktop. As I could have expected, it's pretty much the same as the Apple version, but of course I had even more difficulty finding the files. In the process, found out further details:
The license agreement does suggest paging down to read the rest of the text:
|
|
||||||||
Unfortunately, nothing happens. I'll take the viewpoint that this means that the entire agreement is displayed. Unfortunately, it does limit me: no reverse engineering.
There's a README file, about the only thing that reminds me of real software. But it's presented in Microsoft's “Notepad”, one line per paragraph:
|
||||||||||||||||||||||
How can anybody read that?
I still don't understand the thought processes behind the way this software wants me to work, but it's adamant: it wants to be in charge. You can't do anything behind its back:
|
|
||||||||
“You cannot move or copy images”. That's like telling a film photographer that he can't use his own negative files for his negatives; they've got to be put in special files from Olympus. Imagine having three cameras from different manufacturers, all requiring this kind of nonsense—how could you maintain an overview of your photos?
Using the software is no easier now I've got used to it. I managed to run into trouble with copying from the camera again—something that my little syp script does with no outside intervention—and discovered that, although it does check for duplicate copies, it takes about the worst possible approach to the issue: it doesn't tell you, it copies the files again, adding an appendage in the middle (for example, P1104711.ORF gets recopied as P1104711_1.ORF), and it then forgets about the original files. I went into the directory and deleted the *_* files, leaving just the others—but the software saw an empty directory. I think it must have its own directory, independent of the real directories—but why? This is probably the reason for the restriction on moving and copying.
Processing the images is a pain. As I've discovered, you can mark all the images you want to process together—individually! Imagine coming back with 300 photos and having to tick Every Single One before processing them!
On the positive side, there are ways to mark which photos you want to copy from the camera, at least under Microsoft, with ctrl-left button. That doesn't work on Apple; maybe I just don't know the corresponding key combination.
Conversion really is very slow, even on my Pentium 4 machine. This could be a basic feature of the software; it does a better job, so it might even be justified.
The prohibition of reverse engineering wouldn't be an issue, of course, if the software hadn't been made in this “my way or no way” manner. A simple little program, something like ufraw-batch, that just does the conversion, would be so much easier to handle. But they bundle it with this horrible GUI which makes it almost useless.
Spent some time trying to get it to run under wine, but it failed with some missing library. When I'm feeling up to it, I'll take another look.
Mail from Larry Stamm with observations on the raw conversion dilemma. He's tried many things, but he's not happy either. I need to follow up on the links, and may get back to it. He also pointed me to Andrzej Wrotniak's E-510 settings page.
To make up for the frustration, wrote some more PHP code for displaying the photos. I still don't have proper time lapse capability, but I'm getting more and more interested in comparing the photos I've been taking over the last 15 months. Finally set up something like a “slide show”—if that's the correct term—to display one photo at a time and move quickly from one to the next. It's still pretty shaky, and it show serious issues with differences between one photo and the next, along with issues positioning in the slide show when changing sides. A lot of that is just a matter of tidiness in the code; I could kludge around it, but it's time for fewer kludges.
| Sunday, 11 January 2009 | Dereel | Images for 11 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: brewing | Link here |
Today should have been brew day, and I had done all my preparations yesterday—except for the yeast starter. So another day or two delay before I can mash.
| Topic: photography | Link here |
Yet more work on photographic processing, and spent some time trying to find out how to work around the Olympus processing software. I've decided it's a lost cause; somehow I need to work out why ufraw isn't doing the right thing. Made some progress based only on direct optical comparisons, but it's not enough. I need some software that will compare two images processed from the same original and tell me what's different.
Also more work on the “slide show” software, which now works much better, but the framing looks terrible:
|
|
||||||||
I need to find a way to improve on it. Should I really change the buttons to DVD-player-like buttons?
| Topic: technology, opinion | Link here |
It's amazing what you can find on-line. I mentioned recently that I had bought this new calculator with incorrect instructions from ALDI. Today Jashank Jeremy, Peter's son, came up with the information that it's a rebadged Casio fx-82MS, and even came up with a usable instruction manual for it.
| Monday, 12 January 2009 | Dereel | Images for 12 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: food and drink | Link here |
Mail from James Andrewartha today, pointing out that there's still discussion on naming Chinese vegetables. He writes:
One of my friends made a blog post which links to a working page on the NSW DPI website which claims the difference between buk choy and pak choy is the colour of the stems (as well as flavor and texture), which is how the ones in your photo are organised.
Yes, that makes sense. Well, James' observation does. The idea of NSW bureaucrats telling the world how to name Chinese vegetables beggars the imagination. They don't tell anybody how they think it should be pronounced, which leaves us with three letters that could be pronounced in the same or different ways: [bp][aou]k cho[iy]. The possible pronunciations overlap, especially when you consider the various ways Chinese has been romanized.
| Topic: technology, opinion | Link here |
In the process also found another example of horrible web page layout, this time in the blog post:
|
|
||||||||
That's on my HDTV resolution monitor, so it's nothing unusual.
| Topic: general | Link here |
Phone call from the Optus project manager for the proposed phone tower today—another Greg, who doesn't want his surname published to avoid being inundated by questions really intended for somebody else. Had quite an interesting discussion, in which he showed a remarkable amount of understanding for Wendy and her kind—more than I do, anyway. Got some interesting information from him, including a web site which shows the details of the application and the radiation levels to be expected.
| Topic: technology, photography | Link here |
I thought I had finished my problems with the “slide show” software, then suddenly I got this:
|
||||||||||||||||||||||
The text starting “I need to find a way...” that should have been beneath the picture had been rendered to the right. Confirmed that it also happened on Safari, Google Chrome and Microsoft “Internet Explorer”, so it had to be something with my markup. Checking the HTML found nothing wrong, and neither did the W3 validator. Spent an inordinate amount of time looking through the nested tables that seem to be the only way to create this kind of layout, and found nothing. Finally I made the outside table 100% of the width, and that worked around the problem. Or did it solve it? In any case, the tools are pretty terrible.
| Tuesday, 13 January 2009 | Dereel | Images for 13 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
Yvonne in this morning with the Kenwood Chef that I had purchased online from Grays last week. It didn't match the description: the auction description stated that it was refurbished, but all the accompanying documentation, and the condition of the machine itself, made it clear that it was a brand new factory second. Well, I couldn't tell what made it a “second”—the only distinction from a normal new one was the lack of a glossy carton. That's a slight consolation for the unexpected “buyer premium” that I had to pay.
| Topic: technology | Link here |
Message from James Andrewartha on the topic of HTML debugging:
I find Firebug invaluable for solving HTML layout problems - the inspect tool shows the boundaries of each element as you move the mouse over it.
The screen shots look good, and I downloaded the plugin, but forgot to look at it.
| Topic: general | Link here |
We've been here over 18 months now, so it was high time to finally hang the curtains properly. Spent the usual “far too much” time doing that—I've never paid much attention to curtain hanging hardware before—but it's worthwhile. In particular, we have two doors with curtains in front of them, and until now it was just too much trouble to even use the doors.
It was a pretty warm day today—temperatures up to about 37°—and it brought home to me how efficient a solar water heater can be. Today the hot water out of the tap looked particularly hot, so I measured it: 91°! How can solar heaters make water that hot?
|
Optus mobile tower field visit
|
Topic: technology, general | Link here |
Greg, the Optus mobile tower project manager, turned up today on a site visit, bringing with him 3 other people: Raymond, an RF engineer, and Peter and another Greg, both from DALY (or is that DΛLY or DLLY? They have mixed Greek and Roman letters). Spent some time talking to Raymond about the technology—one thing that was inaccurate in my phone tower page was the power output. The figures there, copied from an MCF document, suggest typical powers of round 20 W, but it can be as high as 100 W—still only about 10% of the power of a microwave oven, of course. It's also interesting to note that nobody is deploying GSM technology any more: soon it'll be necessary to switch over to UMTS.
One thing that came out of their site inspection is that they appear to be reconsidering the location of the tower, and one place that was potentially interesting was our own property. It seems, though, that there's not enough tree cover, and power could be an issue, so it's unlikely to be here. Still, until they come to a final conclusion, anything could happen.
One thing about the visit, though: these are normal, sensible people who understand their job. You'd think that that would be normal, but after my experience with Optus and Telstra over the last year or two, you'd be forgiven for thinking that the companies were run by bumbling idiots. I wonder how much harm they're doing by their externally visible telephone interface.
|
Telstra “broadband”: not soon
|
Topic: general, technology | Link here |
By coincidence, got a call from somebody at Telstra in the evening to talk about my “broadband” options. This appears to be in response to something like a customer satisfaction survey:
So today I received this call to explain to me my options. She told me that I could have “broadband” via wireless, and I had to gradually whittle her down to the point where she admitted that wireless is more expensive than ADSL, and that the speeds are generally lower. She didn't understand my issues with latency, and I suspect she didn't know what the term meant: I even had to explain to her in detail what “upstream” and “downstream” means—she's sitting on the application level and calls it “upload” and “download”. I wonder whether Telstra would accept a complaint that I'm paying not just for the payload, but all the IP overhead as well. She did tell me that the problems I had last year with my Apple had been “resurrected”, and that all Telstra tariffs (“plans”) measure traffic in both directions. But in general nothing new.
| Wednesday, 14 January 2009 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general, photography | Link here |
Today was supposed to be brew day, but somehow I didn't feel like it, so decided to go into town and do some shopping instead. Then I found a reason not to do that either, so spent most of the day in the office. Played around with my “slide show” with the idea of putting in a calendar to select from, and got as far as creating a calendar in it. The good news is that the invocation is as simple as:
But it looks ugly, it's big, and it creates an enormous HTML file. Where do I go from here?
| Topic: opinion | Link here |
Nearly three years ago I had a nightmare experience with the Canadian immigration authorities, one of the unpleasant things that made me decide not to travel any more (that trip was, in fact, the last time I was in a plane). I was so glad to live in Australia and not Canada. Today, though, I heard from Mads Martin Jørgensen that Kaj Arnö has reported similar issues about Australia. You'd think that that kind of treatment belonged in the last century. Is this another indication of the course our new government wants to take?
| Thursday, 15 January 2009 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: brewing | Link here |
Finally got round to brewing today—the first wheat beer since moving here. Things went pretty smoothly: the HLT has started dribbling a little, but not enough to be a nuisance, and the sparge was very slow. I had half expected that, since it's a common problem with wheat mashes, and I had crushed the grain quite finely. But it didn't stick, and I got a good efficiency. About the only real problem—apart from the eternal foam problems when aerating—was that the sparge manifold was blocked. I had to back blow it out, and decanted some unpleasant black stuff, possibly the remains of an earwig, before racking the rest. Hopefully they'll have been properly sterilized by the hour of boiling. Next time I must check my equipment more carefully before assuming it's clean.
| Topic: general, opinion | Link here |
The long sparge gave me time to peruse next week's ALDI “weekly news” flyer. Some of the statements they make there are amazing:
They're featuring East Asian cuisines this week, and I discovered:
There are many different styles of Asian cuisine. Each country has its own unique flavours and ingredients and all use the traditional wok.
How Asia has progressed—eastward—since Roman times!
| Friday, 16 January 2009 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, opinion, photography | Link here |
I had sent a message to Udi Fuchs, the author of ufraw, with a couple of questions; the reply surprised me; apart from the fact that he had changed language, and that he hadn't been able to look at the attachments, he stated:
UFRaw generating underexposed images is a know problem.
Isn't the whole point of the program to generate correctly exposed images? Strangely, it wasn't my issue. But clearly I need more investigation of raw image formats. The current issue of c't (for me; I tend to be a couple of weeks behind) includes a report on “non-destructive raw format converters for mass production”, or some such. I have a high opinion of c't, and I'm sure I can learn quite a bit from this article, but somehow there's a mentality issue. Yes, the ability to convert multiple images without individual attention is important, and it's one of the problems I have with the Olympus-supplied software. But why highlight this ability, or the fact that it doesn't destroy the original? The alternative sounds like madness.
Spent a lot of time reading up on the topics. It still seems to me that the entire digital age has deliberately thrown away the experience of prior generations. The closest Wikipedia article on gamma treats the concept as if it were primarily to do with displaying images on CRTs. And I get the feeling that people have introduced new concepts not because they are needed, but because it makes it easier for them to understand—for example, this weird idea of “resolution” in pixels per inch.
In a similar manner, the c't article also looks for unrelated functionality such as the ability to create web pages from the photos, or sending them as email. What a horrible idea to leave that to the same software that converts images!
| Saturday, 17 January 2009 | Dereel | Images for 17 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, opinion, photography | Link here |
Photo taking (or, more to the point, processing) is still taking up most of my Saturdays. Somehow the digital revolution hasn't brought the ease of use it should have. I've accepted the fact that the Olympus-supplied software gives the best out-of-the-box conversion to JPEG, but only if I use the standard settings. On the other hand, it's glacially slow, and any adjustments cause real pain.
Before even using it, though, I once again got thoroughly annoyed by the tiny fonts. Spent at least an hour looking for a global way to set the font size on Mac OS X. In principle the system has all the information it needs to set a specific font size (and not pixel dimensions): it knows the resolution of the screen, and nearly every monitor nowadays reports its screen dimensions. But somehow there's no uniform way of displaying text at a specific size.
Found a couple of references to a “font panel” in the Help—but nothing to tell you how to find it. It's not in the System Preferences screen, where you'd really not only expect it, but demand it. Found various references on Google, all of which seemed to have gone to trouble to avoid saying how to get to it—and then discovered it's a function of individual “Applications”—and there's no uniform way of bringing it up! What a declaration of software bankruptcy! Maybe that's yet another reason why Apple doesn't have any high-resolution displays: as long as you keep the resolution between 75 and 100 dpi, the problem isn't too extreme.
Back to the Olympus software, and made progress of a kind fighting my way through these silly “Albums” and “Folders”. It took something like 5 minutes to do a chdir function to the directory I want (/dereel/home/grog/Photos/20090117/orig), at each directory traversal waiting for the program to build a display tree with the hundreds of irrelevant entries it felt compelled to show, and then scrolling to the correct directory—and then I discovered that it is possible to just type in the directory name in a little box with a flyspeck marking Address (what does that mean? I thought the bad language for directory was “folder”). And it takes about 2 minutes per image to perform the transformation. OK, this is an old 400 MHz G4 Mac, but the camera itself can perform 5 similar transformations per second. Somehow you can't blame that slowness only onto the computer.
Did some more experimentation with ufraw and came up with some settings that didn't look quite as bad. Here an image of a few months ago that I had somehow missed at the time. Left the Olympus version, in the middle the ufraw version, not “optimized” (i.e. processed by the Ashampoo Photo Optimizer), and on the right the ufraw version after optimization (run the cursor over an image to compare it with its neighbour).
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The overall contrast and saturation is similar, but there are some differences, as shown in the detail of the Camellia in the first two images. The Olympus (first) still looks better.
|
|
|||||||||
|
|
|||||||||
Also, the differences may not be much, but they're enough to completely confuse the Ashampoo optimizer. Here another image: on the first line it's processed by Olympus, then Olympus and Ashampoo, and on the second line by ufraw, then ufraw and Ashampoo:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly there are multiple problems here, but how do I determine what they are?
| Topic: general, technology | Link here |
Greg Larkin has just committed a new port for mythfrontend, so while in a frustrated state, decided to build that. For some reason he hasn't included mythtv-setup, which caused a couple of problems while it tripped over configuration files and a mythconverg database left from prior installations. After that, it seemed to work, and it displays interesting things like signal strength. I suppose I should try using it for a while.
| Topic: general | Link here |
Chris and Fifi over for dinner in the evening. For some reason, we didn't take any silly photos—maybe because they had to leave early to pick up David from the airport, where he was due to land at 1:30 am.
| Sunday, 18 January 2009 | Dereel | Images for 18 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
Last night I had suggested to Fifi that we should all go to the Ballarat Farmer's Market this morning. Chris had correctly pointed out that they wouldn't be home from the airport until 3:30 or so, so it wasn't a good idea. But at 8:45 this morning, we got a call saying they would pick us up at 9:15.
Somehow that became 10:15, and then David went back and insisted that Chris come along, so we didn't get there until about 11:00. Spent some time looking around, during which it occurred to me that this isn't the way I like to buy plants: I need to decide what I want, and then go out and look for it. So we didn't buy any plants. But I got a 16 port 100 Mb/s D-Link switch for $12, something I've been meaning to get since my old one died some months ago, and also some bird netting that we had been meaning to buy. In general, though, “shopping” is something that I see as a necessity, not as entertainment.
| Topic: technology | Link here |
In the afternoon, playing around with MythTV, trying to get cvr2 (soon to become ceeveear) to use ACPI to suspend and awake again as the recordings required. Daniel O'Connor has done some work in this area, and I copied his instructions a little too carefully—they include things like absolute paths and user names. For some reason, Mythbuntu (which Daniel uses) has a user myth, while MythTV typically has a user mythtv, and that's what the Ubuntu packages use. But looking at the scripts also made it clear that I'm going to have to rethink things anyway. In particular, how do I wake up the machine when I want to watch TV from another system? More head-scratching.
| Topic: technology, opinion | Link here |
Spam has been bad lately; as I expected, people have learnt to work around the checks of SpamAssassin, and nowadays it's not much use. But lately I've been getting lots of mailing list subscriptions:
|
|
||||||||
Fortunately, most (but not all) of the subscriptions require confirmation. But why are people doing this? How are people doing this? There are subscriptions in French and German as well. I speak those language, but there is also one in Romanian, and there are related things like “recommendations” and registrations. They appear to be related because they also use fake names like vxqvomxv and tlrpxseq.
| Monday, 19 January 2009 | Dereel | Images for 19 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology | Link here |
Another strange layout issue with my web pages kept me busy most of the day. Although the pages look completely correct and validate correctly with the W3 validator, combinations like this one rendered strangely on nearly all browsers. The worst was—who would have thought it?—Microsoft “Internet Explorer”:
|
||||||||||||||||||||||
Note the scroll bar at the bottom: this image is wider than the screen. The problem seemed to be related to the way I represent these photos: the first one (the nest demolition) is in a table 100% of the width of the screen. For “Internet Explorer”, that's enough; other browsers require the next photo to be in a table as well, which happens when they're at least “small”. But then they're in a table 100% of the width too—so why does it place the second photo to the right of the first?
The problem is reminiscent of the one I had last week. I still don't understand why it's happening, and Callum Gibson, with whom I discussed the matter, thinks it's a basic problem with the browser engine. But it happens in too many different browsers (in fact, all except firefox 3.0, and including Google Chrome), so I suspect it's probably my problem after all. But how do I fix it? Found empirically that adding a non-breaking space before each photo fixed it for all except for “Internet Explorer”, so now I have the following:
In the process, also discovered that my version of “Internet Explorer” is ancient, probably because I don't keep pain on the Internet, so it can't do automatic software updates. Tried to do that now—at least Microsoft “Windows XP” allows you to set up IP aliases relatively easily, and unlike with Apple, they work. But it still managed to fail to download with the message Error Code: 0x80200010, which is surprisingly arcane. Did a bit of googling and discovered it was related with an error name BG_E_NETWORK_DISCONNECTED, which makes sense—the network connection didn't fail, for once (it did many times later in the day), but the rdesktop session went away, possibly because it didn't get keyboard input or something.
| Topic: general | Link here |
Another power failure today, this time in mid-afternoon. It reminds me that Powercor have promised me compensation for the problems I've had, but they haven't delivered.
| Tuesday, 20 January 2009 | Dereel | Images for 20 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, general | Link here |
The weather's still pretty hot—in the course of the morning it hit 41° on the verandah, but by late afternoon it was only about 32°. Spent most of the day in the office, again. At least I made some progress. After establishing that ACPI required More Thought, it became clear that I shouldn't be running two receiver computers any more, so took the plunge, turned ceeveear off and put its tuner into cvr2.
| Topic: technology | Link here |
To my surprise, it Just Worked. Ran the third tuner through its scan, and that was all I needed to do. Then I had three tuners, was able to use them all in parallel and confirm that it worked. That's the way it should be, of course, but it's so seldom that anything in multimedia works the way it should be that it was quite a surprise.
In addition, Callum Gibson told me that VLC can take an entire DVB stream and extract all MPEGs transport streams from it, pointing to this documentation as evidence. It's a bit lacking in detail, so I tried installing vlc—and that worked as well! Then I realized it was on the wrong machine (dereel), and started an installation on cvr2 instead, but didn't get round to doing anything with it by the end of the day.
Freecycle is currently in the process of migrating from Yahoo! Groups to its own server, and we spent a lot of the time changing over. Things are currently a little up in the air.
| Wednesday, 21 January 2009 | Dereel | Images for 21 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology | Link here |
Now that cvr2 seems to be running smoothly, it was time to make it ceeveear again. Maybe I was too encouraged by yesterday's success, which most certainly did not continue today. It started, of course, with finding the locations where name and interface IP address are stored. This is Ubuntu Linux, and it has the host name in /etc/hostname, and interface information in /etc/network/interfaces, all perfectly sensible names. But then there's a file /etc/hosts, which I thought died out with the dinosaurs. Anyway, fixed that, and the system rebooted nicely—but of course, mythbackend didn't start. It was looking for an IP address, not a host name:
|
|
||||||||
OK, that was simple enough; surely it should understand host names as well. So I changed that, and yes, it did get the correct IP address:
But all my upcoming recordings were gone! Spent some time looking through the mythconverg database, and confirmed that the entries were still there (in the table record), but mythweb and mythfrontend both didn't list them. Took a look at the database in a brute force manner:
=== root@ceeveear (/dev/pts/2) /var/lib/mysql/mythconverg 4 -> grep cvr2 *.MYD
(Why does Linux put data files in a lib hierarchy?). That's far too many dependencies for me, but none of them seemed related to the disappearing recordings. Then I found the following in the log, 30 seconds after the preceding (and after I had stopped looking):
So the same mythbackend decided in one place that it was master, and in another place that it was a slave. That seems to be related to the choice of IP address or name. What a crock! Gave up and moved back to IP addressing and the name cvr2. Maybe one day I'll find a better way to do these things, but there are many more pleasant things to do in life.
| Topic: gardening | Link here |
One of them is the garden. I've been following some mystery plants for some time, and today took some photos:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The first two are my Mystery 5, and third is a Gazania. It's interesting to note how much brighter the colours are than the ones I took in March last year.
I don't know what the last one is; it might even be a weed. For the moment, it's mystery 15. There are certainly a number of the plants popping up around it. But it's taking its time to come, and the bud (one of many) is quite fat, so it could also be something pretty. To be observed.
| Thursday, 22 January 2009 | Dereel | Images for 22 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
The weather's still unpleasant. It's cooled down marginally, now only a little over 30°, but the winds were very strong—typical bushfire weather. The wind has torn off large strips of bark from the surrounding gum trees:
|
|||||||||||||||||||||||
| Topic: technology | Link here |
Spent more time working on the MythWeb software. Gradually I'm getting something usable. It would be nice to be able to stream this stuff, though, and to save positions. I thought that the previous version of mythfrontend did that, but I can't see any evidence in the latest release. And somehow mythtranscode doesn't get the timing information correct; I suspect it's using the wrong information from the DVB feed. I've been meaning to investigate that for a long time, but I don't see it coming any time soon.
|
Bushfire info: beware clever web programmers
|
Topic: rant, opinion, technology, general | Link here |
In view of the weather, went looking for bushfire information on the web. The results were horrifying:
There are two sites: http://www.cfa.vic.gov.au/incidents/incident_summary.htm is for bushfires on private land, and http://www.dse.vic.gov.au/DSE/nrenfoe.nsf/LinkView/519C51D981DAE41FCA257257000A5163DC25C965BDA0CAF5CA2573B400013504 is for bushfires on public land. To find out what's happening, you need to go to both.
Look at that second URL! How can anybody remember it? Clearly a triumph of content management system over usability and common sense.
The layout of both web sites is terrible. The CFA site doesn't even tell you directly where the fire is. There's no map, and the “Location” field is detailed enough that you can't easily tell where it is. The only way to find out is to know the region numbers:
|
|
||||||||
On the CFA site, the fires are sorted by (reverse, of course) time only. Even if you know your region number—as near as I can tell from their tiny, untitled maps, I'm on the border between regions 6, 7 and 15—there's no way to find whether there's a fire in the area without wading through all incidents reported in the state, including false alarms and fires that have long been put out. The least they could have done would be to allow sorting by region, type and status.
By comparison, the DSE site is much worse. It's obviously written by a clever web programmer, and it does have a map. The URL is indication enough, but there's more. The main content is in a frame with scroll bars in both directions:
|
||||||||||||||||||||||
|
||||||||||||||||||||||
The window is wide enough to eliminate any requirement for a horizontal scroll bar, but that doesn't worry it. Another triumph of the “don't show them more than you absolutely have to” mentality.
From a content point of view, the page is almost useless. Unlike the CFA site, there's no obvious sequence beyond status. The top 10 fires on 23 January 2009 show two marked Go (if you scroll to the bottom with two vertical scroll bars, you can scroll to the right, losing the start date in the process, and find this means “Going”), 4 marked Co (“Contained”) and 4 marked Co (“Controlled”):
|
||||||||||||||||||||||
But look at the dates! If you can believe the site—I have my difficulties—the 7th fire was reported on 30 October (presumably 2008, but not stated), and there have been no reports since 2 January. There have been no reports about the second fire in Bendigo for the past week.
And how serious was the fire? You can find out more information with the help (if that's the word) of their clever Java programming:
|
|
||||||||
But that requires mouse action for every single fire. Of course, the size of the affected area is reported in the main list, just hidden by the page layout. Yet another triumph of the “don't show them more than you absolutely have to” mentality.
What are these old fires doing on the page anyway? This is a page titled “Fires Today”. As I write this on 23 January, only 9 of the 39 reports (including 6 marked “Not Found”) fires have any report with today's date. 11 are more than 3 days old.
About the most accurate thing on the page is the statement “This state map is automatically generated and may contain errors.” But this is a serious matter. Why do we have to put up with this kind of nonsense from a government agency?
| Friday, 23 January 2009 | Dereel | Images for 23 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, opinion | Link here |
Another day spent ranting. Apart from yesterday's diatribe about the state of bushfire reporting, I had another question to answer, from a bloke called Alfred who had contacted me about problems caused by his Digitrex DVD recorder:
X-Mailer: Microsoft Outlook Express 6.00.2900.5512
NB. What is the reason for you not to write a conventional email like the one I am just writing, but in an attachment style? Never came across this before, hence my question, for we don't correspond exactly covert or subversively, do we?
It's sad that Alfred thinks that Microsoft “Outlook Express” is a “conventional” MUA, but it occurred to me that I should write a better explanation of signing email, and that took me most of the afternoon. I had just uploaded it when John Marshall said on IRC:
The message wasn't confidential, so John forwarded it to me. Yes, it's a validly signed message, and the public key is on both http://pgpkeys.mit.edu/ and the Microsoft web site—everything looks correctly done. It's just strange that—apparently—Microsoft's own MUAs can't evaluate the signature.
One interesting thing about the message is the way it's signed—unlike the procedure that I mention in my web page, it doesn't require MIME. This also means that I need a special incantation (ESC-P) to validate the signature. This is the first time I've seen this.
| Saturday, 24 January 2009 | Dereel | Images for 24 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
Mystery flower 15 has started to flower, slowly and interestingly:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
There's less than 4 hours between those two photos, long enough for me to no longer be able to take a second photo when I discovered the first was partially out of focus. It'll be interesting to see how it progresses.
| Topic: photography | Link here |
Somehow the house photos that I take every Saturday aren't getting easier. Today I decided it's time to retire some views, but that was the easy part. I think I'm gradually getting on top of the procedure, though, and the results with ufraw are good enough that I don't have to endure the pain of the OLYMPUS Master 2.
The layout of the web pages is another thing, though. To fix the last problem with the photo layout, I had made the tables take up 100% of the page width, which really shouldn't be necessary. One result was that the descriptions on the right of the exterior photo page appeared to be centred in the middle of the right-hand part of the page. With the aid of Firebug, discovered that the real issue was that the table column for the photo was too wide (firebug outlines the cell as a blue box):
|
||||||||||||||||||||||
HTML is really a terrible layout description language. It seems that if I make the “big” image tables full width, and leave the “small” image tables natural size, nothing breaks—yet. But it seems to be only a matter of time before the next bug crops up.
| Topic: opinion, technology | Link here |
Mail from “eddie” today, apparently referring to my recent complaints about bushfire web sites:
http://www.cfa.vic.gov.au/incidents/incident_summary_rss.xml
try this one i use google reader
So I tried it. It looks like an even less usable version of the page at http://www.cfa.vic.gov.au/incidents/incident_summary.htm. I see a non-tabular form, with dates which can't decide whether they're Australian or US American, and all the disadvantages I mentioned in the diary page. If there has been no report on a fire in the previous 5 hours (presumably common enough early in the morning), you won't even be able to find out about the fire. About the only advantage that I can see is that it's an RSS feed, if you like that sort of thing. I replied and asked for clarification, but I haven't heard any more.
RSS is one of those strange things that I've never been able to understand. What good is it? About the only thing I can see is that makes up for some deficiencies in web browsers, notably the ability to mark things “read”. I disagree with the polling approach that requires periodic download attempts to find updates, and the static approach means that a lot of the functionality I put in my PHP scripts can't be used.
For the last couple of years I've been going to some trouble to convert this diary into RSS format, and I've been monitoring the number of hits. Because of the way RSS works, you get multiple hits for each feed.
It seems that people who read my diary are of the same opinion. A year ago I had about 400 hits a day on diary.xml and only about 90 on diary.php. The most recent figures (for today) are:
A quick search of the log files shows that only about 20 people read it at all:
=== grog@ozlabs (/dev/pts/1) ~ 7 -> grep "GET /grog/diary.xml" /var/log/apache2/lemis.com.log | awk '{print $1}' | sort | uniq | wc -l
|
“Not a bug, a feature”: episode 4714
|
Topic: opinion, technology | Link here |
A thing that has been annoying me for a very long time is a bug in ls, the program that lists directory entries. I use it to sort my photos so that they're in chronological sequence. In principle, I could go by the file name: all cameras have a naming scheme that increments a number for each photo taken, so an alphabetical sort would be enough—if you're only using one camera. As soon as two cameras come into play, you have a situation like this:
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 62 -> ls -lT
Here the photos taken with Canon 20D, with names like img_5503.jpg, were taken mainly between the photos taken with the Nikon L1 (names starting with DSCN1325.JPG). But that's OK: you can sort by time. The result is:
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 63 -> ls -lTrt
At first glance, that looks like what you want: the images are sorted by time. But all these files are stored in the camera in the antiquated FAT file system format, and it stores time stamps with a resolution of 1 second. What happens if more than one photo is taken in a second? They all end up with the same time stamp.
That shouldn't be a problem, though: the specification for ls states that in this case, the files will be sorted by name. But looking at the list above, something's wrong: img_5503.jpg and img_5504.jpg have the same time stamp, but the names are sorted the wrong way round. And that messes up the natural sequence of the photos.
Today I spent a lot of time in the debugger trying to find this bug. The code (in /usr/src/bin/ls/cmp.c on BSD systems) looks straightforward enough, though obviously written by a Real Man who doesn't believe in comments. I've added some to make it clearer:
This all looked perfectly correct, and it took me a lot of time to realise that the name sort was the other way round from the time sort. By definition, ls -t sorts with this HORRIBLE reverse chronological order. But the name sort doesn't; it's correct. So that makes it the other way round from the time sort.
OK, we can fix that—or can we? ls is defined by the Single UNIX Specification. In the man page for ls you can read:
- -t
- Sort with the primary key being time modified (most recently modified first) and the secondary key being filename in the collating sequence.
In other words, this ridiculous behaviour is mandated by standards!
So what could I do? Add yet another option to ls? We already have so many, and FreeBSD, Linux and NetBSD have additional flags, and they're already diverging—I discovered while checking that I had caused a conflict with the -D option that I added a year or so ago. Linux uses -D to output in a format suitable for Emacs' dired mode (why, I don't know; Emacs gets by without this feature quite happily under FreeBSD), while FreeBSD now uses it to specify a custom listing format. Even the -T option I use in these examples isn't supported by Linux. In BSD it means “display complete time information for the file, including month, day, hour, minute, second, and year”, while in Linux it means “assume tab stops at each COLS instead of 8”. And of course the SUS doesn't define it at all.
In the end, after playing around a bit, I decided that an environment variable (which I am currently calling LS_SAMESORT for want of better inspiration) would be the best solution. This isn't the kind of option you want to use only sometimes; it's the kind of variable that some people want to use all the time. The result:
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 74 -> ls -lrTt img_5518.jpg img_5519.jpg img_5520.jpg img_5521.jpg img_5522.jpg
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 75 -> LS_SAMESORT= ls -lrTt img_5518.jpg img_5519.jpg img_5520.jpg img_5521.jpg img_5522.jpg
Somehow that's still not satisfactory: now everything is backwards. The real problem is that -t sorts the wrong way round, not that the name sort is wrong: these examples all use the -r (reverse order) flag. But the use of -t is so ingrained, and the issue with the conflicting name sort so uncommon, that I don't think it appropriate to try to fix the -t sort. I need to think about this a little more.
| Sunday, 25 January 2009 | Dereel | Images for 25 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology | Link here |
Couple of feedback messages today: Tim Bishop wrote to say that Microsoft MUAs do support S/MIME. That's useful to know, I suppose, but it doesn't help people who use PGP signatures, including Microsoft's security team. I'm left wondering what use it is if their own people don't use it.
Also, Daniel Nebdal wrote with some interesting ideas on how to get firefox to display the DSE current bushfire site with less breakage. It should widen the artificially narrow frame inside the main web page, so that all details can be displayed without scrolling:
I tried it, and it seemed to work while loading, but then it pulled the window in again to the old width. Since it doesn't address other serious issues with the page, I don't think I'll investigate further: I have enough problems with my own layout issues.
I've been using Emacs in various forms for nearly 30 years now, to the point where Emacs key bindings have become conditioned reflexes. All the more surprise today, while writing yesterday's diary entry, when I accidentally caused a bright yellow highlight. I'm used to the brownish ones which represent a region selected by the mouse, but what's the yellow one? I couldn't find a way to find out.
|
|
||||||||
Google Maps seem to have sent their “Street View” people out everywhere, even past my house. The feature works very badly with firefox on FreeBSD, frequently displaying only an empty window, but on Apple with Safari it seems to work:
|
||||||||||||||||||||||
The problems with firefox appear to be the unreliable emulation of the Linux flash plugin.
The photos are surprisingly good—the roof section in the middle is over my computer room. Under those circumstances, it's really amazing that they still misspell “Kleins Road” as “Kliens Road”.
| Topic: photography, gardening | Link here |
Apart from that, a lot of photos: took some of yet another mystery plant:
|
|||||||||||||||||||||||
Also a bee in some of the gazanias:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Also more panorama stuff. This is the view of the sun setting against the trees that divide us from the lagoon, as seen from the new verandah:
|
|||||||||||||||||||||||
| Monday, 26 January 2009 | Dereel | Images for 26 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
More work in the garden today—I really should be doing much more work. While potting a Daphne on the verandah, turned around and saw this about 50 cm behind me:
|
|||||||||||||||||||||||
It was about 30 cm long, including the tail, and completely motionless; it didn't even run away when I went off to get my camera. For a while I thought it was dead, but some minutes later it left again.
| Topic: technology, photography | Link here |
Michael Hughes caught me out with yesterday's discussion of modification times. I had a valid point, but I had oversimplified:
Have you thought about changing the file modification time with the EXIF date taken or date digitized? I have a script that moves my photo in to a directory based on the date taken:
/camera_pictures/<year>/<month>/<day>/<image name>
In fact, I've done something similar myself. But that wasn't the issue. There are multiple considerations:
These photos were taken with the camera set to the incorrect time. My time setting is a best guess. For example, one of the images I mentioned in the page:
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 115 -> ls -lT img_5504.jpg
=== grog@dereel (/dev/ttyp2) ~/Photos/20061223/orig 116 -> exif img_5504.jpg |grep Date
This was before I wrote the exiftime script. Instead, I used touch(1) to set the times manually, and I didn't set the seconds value. This makes things worse, of course, but:
The real issue is with SLRs that can really take more than one photo per second. EXIF doesn't help there: there will still be more than one image with the same time stamp.
So: I chose a poor example. I was aware of this when I wrote the page. There are better examples, but I couldn't be bothered to go and look for one.
|
What good is Flickr?
|
Topic: photography, opinion | Link here |
Over the weekend, Edwin Groothuis had a birthday party and invited a lot of people (myself included, but it was 1000 km away). Juha Kupiainen took a number of photos of the event, and put them up on Flickr. I went to take a look.
Clearly I'm not the typical Flickr “customer”. I spend a lot of time trying to organize my photos, an effort that Flickr customers don't need. But I wouldn't want to give up that freedom. I find Flickr very painful to use:
As I've frequently mentioned and demonstrated, I like high-resolution photos. My “thumbnails” have 67,500 pixels (nominally 300x225), the “small” photos are 270,000 pixels, and the “large” photos have whatever the camera supplies, typically in the 6 to 10 MP range. Flickr is smaller everywhere, though I suspect that part of it is individual choice.
On my screen, the front page displays 9 images, apparently at 240x160 pixels, about 60% of the size of my thumbnails. But in the same real estate, my pages display up to 16 images, assuming that, like on Flickr, they're all in landscape format. There's a lot of space used up for repetitive information like access rights and “All rights reserved”, along with links to (in this case non-existent) comments.
|
||||||||||||||||||||||
How do you access bigger images? On my web site, click on the photo, and that photo will be made larger. That works on Flickr too, but instead of the 270,000 pixel image that I show, it's only 500x333, or 166,500 pixels—again about 60% of mine. But that's not so bad. The problem is, all the other photos are gone! Well, there are two tiny 75x75 thumbnails to the right, with which you can navigate:
|
||||||||||||||||||||||
So to view the pictures in (quite) “small” size, you need to click on the barely recognizable thumbnails, and you get one image at a time. You can get two by opening a new window, but of course windows are out of fashion, and with tabs you can only see one at a time.
But 500x333 isn't exactly big—it's only a little more than double the size of my “thumbnails”. Any bigger photos?
I think that Juha deliberately limits the size of his “large” photos to only 1024x681, and that others may choose significantly larger sizes. In fact, it seems that there are a total of five image sizes; the fifth is a “small” image 240x160 in size. But the navigation! This one comes on a relatively uncluttered window, completely by itself. There's not even a way to navigate without going back to the previous windows.
On the whole, I find the Flickr page so difficult to navigate that I can't be bothered. On my pages I have a flexible way to change the size of individual photos, to navigate easily, and to get more information about the photos, such as the EXIF data. It's there on Flickr, but once again it requires another mouse click to access it. I've also found scripts to give names to images relatively easily. With Juha's permission I've run the photos through my system, which also makes it easier for me to look at them.
So: which approach is better? Mine is here, and Flickr's is here. Clearly lots of people like Flickr, but like so many “modern” things, I still find it irritating. I should probably spend some time looking at what they offer, though: it could be that they offer useful functionality that I hadn't thought of.
| Tuesday, 27 January 2009 | Dereel | Images for 27 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general, gardening | Link here |
Into town today to do a lot of things I've been procrastinating about for some time: bought some more garden fittings, though I'm still having difficulties with the drippers. It seems that all cheap drippers have barbs, and all riser tube is designed for screw fittings.
| Topic: music | Link here |
I've agreed to lend my 1826 Savary jeune bassoon to Aligi Voltan for a year, but getting it to him has proven to be difficult: it doesn't have a case, and the bell was so warped that Guntram Wolf had to nail it to the bass joint, resulting in a single piece 95 cm long—and international parcels are limited to 105 cm. In addition, nobody seems to offer appropriate tubes for this kind of object. Finally got hold of the phone number of Homewrap, who do some of that kind of thing, and called them up. They're not in Ballarat, but they gave me the address of Storage King in Wallis St, so along to them. They don't have that kind of packaging either, but they send me to Pack & Send in Howitt St, so along there, and finally found somebody able to not only supply the packaging, but to, well, pack and send the instrument to Venice, and that for a price that appears lower than what I found last year—only $215 odd, including packaging and materials, while I had been quoted up to $550 last year. It looks like the problem is solved.
| Topic: technology | Link here |
Back home and spent even more time on PHP programming. Somehow this is a can of worms. I can certainly understand why some people prefer Flickr, but the process of converting Juha's photos made some things clear: firstly, most of the work was navigating Flickr's web site to download his photos to my system; processing them locally was very fast by comparison. And secondly, it made them so much easier to look at. That's clear, of course, because I've optimized my system to suit me, but I suspect I'm not the only person who would think so.
|
|||||||||||||||||||||||
| Wednesday, 28 January 2009 | Dereel | Images for 28 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
My uncle Max gave me some strawberry plants for my last birthday, and we quickly discovered that we would need protection if we wanted to harvest any fruit. Today, though, we found what appears to be the result of a kangaroo climbing on one:
|
|||||||||||||||||||||||
I wish I knew a way to keep kangaroos out.
| Topic: general | Link here |
The weather's been getting warm again lately—in fact, warmer than I've experienced in 50 years. About this time in 1959 we had a heat wave that I recall rather differently from what's published on the web. We had up to 48° in Melbourne then, and the highest forecast for Ballarat for today was 40°, though I measured noticeably higher values:
|
|||||||||||||||||||||||
In fact, it seems that the temperatures in Melbourne in 1959 were only about 45°. See next month
But the web reports seem to think that 4 days at 41° are worse than 3 at up to 48°.
| Topic: general, opinion | Link here |
My experiences with American Express are still not uniformly positive. Last month I called up and asked for a form for direct debit; it still hasn't arrived. This month I called up the only number I could find on my statement, 1 300 363 276, and discovered that it was the rewards programme. Finally got to a human, Andrew, who called me “Mr. Lehley”—why do these American Express people have so many difficulties with names? Hopefully I'll get the form this time.
| Topic: technology, photography | Link here |
Another special issue of c't on digital photography arrived today, along with a DVD with tutorials on the GIMP. Put it in boskoop and tried to access it. The “Finder” lived up to the name I've given it, “Loser”: every file I clicked on disappeared from the display, without any other action. When I clicked on directories, I got a separate display, but the name disappeared. The sequence looked like this:
|
|
||||||||
|
|
||||||||
|
|
||||||||
No error messages, of course—they could scare people. From a sane shell interface it was clear that the kernel still knew about the files, but that this badly fitting GUI on top had lost them. Tried to stop and restart “Finder”—does it really stop?—but that didn't help. A reboot did. Clearly Microsoft-class software, possibly even worse.
| Topic: technology | Link here |
That wasn't the only problem. For some time now I've been offered updates for QuickTime and iTunes, and an article I read made me interested in trying out something with the latter. Tried to install and found:
|
|
||||||||
OK, maybe something went wrong with the download. But what? Did the download complete prematurely? Who knows? Tried again, and got the same result, after using 120 MB of my precious download quote (by comparison, the previous total for the month was 1080 MB). What do I do now? I feel so helpless with this Apple stuff.
| Topic: gardening | Link here |
Because of the heat, decided to water the garden again in the afternoon. Yes, it's not a good idea to water when the sun's shining, but it's probably better than letting things die. To my surprise, the LCD of the sprinkler controller was suffering from the heat, and it was barely possible to read it:
|
|||||||||||||||||||||||
While doing some stuff in the garden, heard a loud crack, which I couldn't locate immediately. Later I found it:
|
|||||||||||||||||||||||
A branch of the conifer near the entrance had split lengthwise, probably due to the heat, and fallen on the bonnet of Yvonne's car.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Fortunately it did no damage, but we decided it was prudent to move the cars, and later it fell further, smothering one of my rose bushes, which I had to cut free:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
For a while we were concerned that the whole tree might collapse, and I asked CJ over to take a look. There's definitely a lot of chainsaw work there, but we can't do that during a total fire ban.
| Topic: general | Link here |
In the evening, another thing that we didn't get 50 years ago: a power failure. The power failed at 20:15, and based on my previous experience, I called up immediately. Got a recorded message “We are currently in an emergency situation. If you do not have a life-threatening emergency, please call back later”.
What kind of emergency? Why can't they at least give a summary? Did the emergency apply to our area? We've seen this kind of confusion before. So I stayed on the line, and it led me through to the normal site, where I was able to enter my post code and discover that no outage was reported for our post code. So I held on for the promised 15 minutes, but the line went dead after about 6, and though I held on for another 10 minutes, nothing more happened.
Called back a couple of hours later, and after more non-information, called again and pressed the “life-threatening emergency” button, and was connected to somebody who asked for the exact address and said, yes, the power had failed here, and they didn't have an estimated time for restoration, but that I could follow progress on “The Internet”.
To bed without cooling, and the power finally came back at 0:47, at which time the outside temperature was still 30°. Yes, it's clear that weather like this puts more of a strain on the power suppliers, but that's no excuse. They should plan for it.
| Thursday, 29 January 2009 | Dereel | Images for 29 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: technology, general | Link here |
Into the office to pick up the pieces this morning. All systems had gone down of course. Did I say all? No, one resisted: pain, my Microsoft “Windows” XP box, was hibernated at the beginning, like it usually is, and it was still hibernated at the end. Somehow I find that disappointing.
Discovered that the UPS that was feeding dereel and boskoop had stayed “off” after the power came back, so I had the whole recovery process ahead of me. Started that off, got dereel up and running, and out to attend to a bloke picking up some gas cylinders.
Back inside, and dereel had spontaneously rebooted! Tried again, keeping more of an eye on it, and this time it didn't spontaneously reboot. It took me a while, though, to discover that it had hung. Potentially there was something on /dev/vty0, but I was in X, and there was nothing to be seen.
Big Red Button time. This time I left it in text mode, and sure enough, it hit a page fault in kernel mode—but I had installed a GENERIC kernel, and I couldn't take a dump.
My guess was the background fsck of /home, which has (as I discovered) about 7.5 million files. Run fsck in the foreground and discovered that the main problem was a file called .snap/fsck_snapshot. Subsequent investigation, mainly by Callum Gibson, suggested that this was the problem. Workaround: don't use background fsck on big file systems, exactly the place where you want them.
|
All-time high temperature
|
Topic: general | Link here |
The hot weather continues! This morning the temperature on the verandah had reached 40° by the time I got up, and in the course of the morning it hit 46°:
|
|||||||||||||||||||||||
Things weren't quite as bad as the Bureau of Meteorology had predicted, though: in the late afternoon some clouds came up, and the temperature dropped from 40° to 32° in the course of an hour. But it wasn't enough to save some plants; one of the Jasmine plants looked very unhappy, and even subsequent watering couldn't save some of the new shoots.
| Topic: gardening | Link here |
The rose bush that had been crushed by the falling conifer branch yesterday looks a lot better today:
|
|||||||||||||||||||||||
But then it's been used to rough treatment. Six months ago I cut it down to the ground:
|
|||||||||||||||||||||||
About a third of the branches were damaged and had to be removed, but the rest are looking remarkably happy.
Also made a decision about my mystery plant 15: it's a weed, specifically a thistle. Here it is just before I pulled it out:
|
|||||||||||||||||||||||
Some thistles are pretty, but this one didn't bloom nearly as much as the fat buds promised, and judging by the number that have self-seeded nearby, it could easily become a problem. So it's gone.
In fact, this proves to be Safflower (Carthamus tinctorius), which is used as a saffron substitute.
| Topic: opinion, technology | Link here |
Where's free software (“Open Source”) heading? I believe the era is coming to an end; the BSDs (with the exception of Apple) are suffering, and Linux is keeping its head above water by gradually becoming proprietary. Yes, Linux is still “Open Source”, and it's likely to remain that way, but the direction is increasingly being determined by commercial companies.
Is that bad? That depends on your attitude. Many people will be surprised to discover that I've never been a religious free software fanatic. Making software costs money, and people need to live. So (sorry, rms ), I don't believe it's sinful to ask for money for software.
That's a very different issue, though, from my beefs. I hate Microsoft because the user interface is so abysmal. I like free software because I can hack it the way I like, if I have to. In both cases, I'd rather have software that works the way I want it. And that is behind my myriad rants.
All the more reason for me to be disappointed with Apple. I'm aware that a number of people find my constant attacks on Apple to be offensive. There seems to be something almost religious about their concern; it doesn't address the issues that I'm complaining about, just that I'm complaining about Apple. Maybe they'll be partially placated if they understand that I really, really would like the Apple software to work right. And there are so many mistakes, so many missed opportunities. If I were to guess what platform I'll be using in 10 years time, it could well be Apple. So I have a vested interest that it should work right.
Here's another case where it doesn't:
After two attempts to download the latest iTunes update, decided to go the manual way via Apple's download site. Based on prior experience, I ascertained that my Mac OS X 10.4 is called “Tiger”, and limited my search for iTunes on Tiger for PPC:
|
||||||||||||||||||||||
But that's not the search it performed; it seems to have returned all hits for any of the keywords, a total of 416 spread over 42 slow-loading pages, in no particular sequence. Tried again searching only for iTunes. The first hit was completely unrelated:
|
|
||||||||
But more to the point, this is a web page. Why does it give me a URL which isn't a link? Followed the Download button and got:
|
||||||||||||||||||||||
How do I send mail? They didn't say. Finally found the correct page, but once again I was using the wrong browser:
|
|
||||||||
Again, no option to download for Apple. Considering that Apple bases much of its userland on FreeBSD, it's surprising that it offers me iTunes for Microsoft when I access the page with a browser identifying itself as firefox for FreeBSD. Back to running “Safari”, and was—finally!—able to install the software. But what a pain!
| Topic: technology | Link here |
I've been having a lot of network issues lately. The two failed attempts to install iTunes yesterday weren't the only ones. I've had multiple network dropouts, each of which kills all TCP sessions, and today I downloaded the Ubuntu pocket guide, which worked well, but the unzipped file was invalid (why didn't unzip discover that?). Tried again and it worked, but it seems to be a genuine download error:
=== grog@dereel (/dev/ttyp6) ~/Desktop 41 -> l ubuntupocketguide-v1-1.zip*
=== grog@dereel (/dev/ttyp6) ~/Desktop 42 -> cmp ubuntupocketguide-v1-1.zip*
In that connection, got a message from Wideband, stating:
A number of customers across Australia are experiencing intermittent drop out problems on our Satellite service. These faults have been attributed to the extreme heat that many parts of Australia are currently experiencing.
The outdoor transmission equipment on the Satellite dish is rated to 55 degrees, however with the reflected heat off the roof it is pushing the equipment past the rated temperature.
When the equipment exceeds the rated temperature, it causes the unit to transmit off the designated frequency which causes the unit to go offline.
That's certainly a partial explanation, but only a partial one. It doesn't explain, for example, why I am off the air so often in the middle of the night.
| Friday, 30 January 2009 | Dereel | Images for 30 January 2009 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: music, general | Link here |
Into town today to finally send off the 1826 Savary jeune bassoon to Aligi Voltan. Got there and discovered that I had forgotten the address. Still, I can email that. Also did a few other odds and ends, and got home relatively early.
| Topic: opinion, technology | Link here |
I was a little too hasty in praising pain, my Microsoft box, for staying up yesterday. It wasn't until today that I tried to access it, and I discovered that it had forgotten half of its IP configuration: a week or two ago I had added an IP alias to allow it to communicate with the Internet. That had worked, and it had happily kept both addresses, but today it had lost its main address. I don't know whether that has anything to do with the power outage or not.
| Topic: opinion, technology | Link here |
When copying video recordings, discovered that no fewer than three of them had failed. Further investigation showed that they had all been on SBS. Into MythTV's inscrutable log files and read:
What does that mean? It clearly refers to some undocumented MythTV internals. What's a PSIP? A PAT? Also, clearly, the lone 0x are the result of selecting an inappropriate output format for hex digits, and the lines should be read, for example:
The most important information, almost completely obliterated: “Rescan your transports”. Further investigation confirmed that 885 was the service ID for SBS HD.
So did SBS change its feed parameters without warning? That's not very likely. But that's exactly what they did. No information at all in the programme schedule; for that you have to go to their silly home page:
|
||||||||||||||||||||||
Not there? Well, look at that scroll bar. It's clearly not as important as other SBS staples such as sports results, advertising, or finding marijuana with the aid of Google Earth. Scrolling down, you can find:
|
||||||||||||||||||||||
If you still don't see it, it's in the middle on the right. No information there that this might mean loss of reception. For that you have to go to another page, which is relatively informative.
I could complain about the complete stupidity of the SBS web site administration. But I did that only recently. Sigh.
Off to try to rescan my tuners, and it found them all, and allocated a new multiplexer ID—without any frequency specification! It also left the old information in place, so I now had, for example:
Set the frequencies from the old information, and then removed everything to do with mplexid 9. But after that, trying to display SBS HD on dereel with mythfrontend failed with an obscure error message that looked very similar to the previous ones:
Yes, now there's another TLA there, PMT. But it was able to display some of the other SBS streams. Is this a bug with the current (beta) FreeBSD implementation of mythfrontend? Who knows? Tried a manual recording, and that worked, so gave up worrying.
The worry that remained was that I now had no programme data for SBS. Tried running mythfilldatabase and got the message:
That's all well and good—maybe. Is this really so important? And how do you run shepherd with the --notimetest option? It seems that you have to run it directly, rather than from mythfilldatabase, and in the process got another silly error message:
=== mythtv@cvr2 (/dev/pts/4) ~ 20 -> .shepherd/shepherd --notimetest
At the moment, shepherd runs into hundreds of parse errors and takes about 2 hours to run, so gave up waiting and set this evening's programme to record manually. What a pain!
In the process, it's interesting to note a discrepancy between what SBS say and what they broadcast. On their web site, they show:
|
|
||||||||
There, for regional Victoria, they claim to have a channel SBS News with service ID 882, but what they broadcast calls itself SBS 2, and the channels are off by one. In fact, they don't have anything calling itself SBS News, but they do have SBS 4, which isn't mentioned on their web page:
This is taken from the broadcast stream, so it has nothing to do with my setup.
|
Kitchen experiments
|
Topic: food and drink | Link here |
A couple of months ago we bought a Vietnamese mint tree, which is now growing nicely. What do you do with Vietnamese mint? I've never used it for cooking, but it has a very pleasant flavour. Did an experiment with a Thai dish that really wanted Thai basil—something I haven't been able to grow yet—and it turned out quite nicely.
|
Rice cookers: why bother?
|
Topic: food and drink, opinion | Link here |
In the process, tried out our new ALDI rice cooker. I've been cooking rice manually for over 40 years, and until now I've never found a reason to use a rice cooker. Peter Jeremy suggested I should try one, just for the fun of it, and since ALDI allows returns within two months, and the price ($29) was right, it seemed a worthwhile experiment.
Result, for me at any rate: no advantage, several disadvantages. The advantage Peter mentioned was that you didn't have to wait for it to come to the boil. But you do, sort of, so you can know when your 15 minutes are over.
In return for the convenience the cooker offers, there are several limitations:
About the only potential advantage that I can still see is in making nasi lemak. When doing that in a pot, there's a real danger of burning. Potentially it would work better in a rice cooker. It might also be a disaster. I'll have to try.
One thing that definitely is ALDI's fault is the unbelievable inaccuracy of the measurements on the rice cooker. I've already observed that the rice cooker uses “cups” of 180 ml, and indeed the measure they supply (conveniently marked only “¼”, “½”, “¾” and “1”) measures pretty well exactly 180 ml. There's also a scale on the inside of the cooking vessel that equates cups to 180 ml.
According to the instructions, I should have put 3½ “cups” of water in for my 2 “cups” (which proved to be 300g) of rice. But when I did that, even with the rice in the pot as well, I was a long way from the lowest (4 cup) marking in the almost illegible scale on the Teflon lining. Further investigation showed that the scale is roughly off by 100%: you'll need a litre to fill it to approximately 500 ml. Here's a litre:
|
|||||||||||||||||||||||
This is also by far the most legible I've been able to make it. In the kitchen, I can barely see anything. So clearly the scale is useless.
| Saturday, 31 January 2009 | Dereel | Images for 31 January 2009 |
| Top of page | ||
| previous day |
| Topic: general | Link here |
Finally it's cooler! The temperatures today only hit 36°. That doesn't sound exactly cold, but it's 10° cooler than we had just a couple of days ago. Some of the plants are not happy about it, like the tree fern. As best I can understand it, the fronds were burnt off by the sun:
|
|||||||||||||||||||||||
Saturday photos didn't take quite as long today, which was good in view of the problems I had with the SBS web TV feed conversion, which kept me busy for much of the day.
In the evening, the temperature became decidedly civilized, only 23°. Out on the verandah we found plenty to keep us busy:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Carola Schlanhof, from Wien, will be visiting us later this week—she's been planning the trip for nearly a year. Today she asked us to catch some kangaroos for her, but that's easier said than done. The best we could do at short notice was a wallaby:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And, of course, it didn't want to be caught.
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 |
