|
| Greg's diary |
| February 2010 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Monday, 1 February 2010 | Dereel | Images for 1 February 2010 |
| Top of page | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
Whatever kind of Haworthia I have, whether fasciata or attenuata, it's obviously a lot happier in moister and less bright surroundings, and its flower stalk is now coming out in tiny flowers, only a couple of millimetres across:
Our ornamental grapes are growing at significantly different speeds, perhaps because I left more stems on what is now the shorter of the two. The taller one has now reached the top of the verandah and needed a guide wire to grow back towards the house:
|
|||||||||||||||||||||||
The lemon tree is still not looking happy, but since I've taken it out of the wind it's doing a lot better. This photo is mainly for comparison purposes.
|
|||||||||||||||||||||||
| Topic: brewing, opinion | Link here |
I bought a single bottle of Weihenstephaner Dunkelweizen in Melbourne last month, mainly as a comparison with what I'm brewing at the moment. The claims on the back are remarkable:
|
|||||||||||||||||||||||
“Yeast white (wheat) beer, dark. Permanently cloudy, fine yeasty, malt aromatic. Brewed according to the Bavarian Reinheitsgebot with the experience of a traditional white beer specialist”.
And what is the Reinheitsgebot (“Purity decree”, not a term used at the time)? It regulated prices (I wish they had kept that part) and the ingredients to be used in beer. Barley, hops and water, nothing else. Especially no wheat, which was needed for bread. And that's what white (wheat) beer is made from.
| Tuesday, 2 February 2010 | Dereel | Images for 2 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Censorship and clueless politicians
|
Topic: technology, opinion | Link here |
You'd think that by now that even Australian politicians would have understood that they have little control over the Internet. Maybe they have, but they've found a strange way to demonstrate it. According to a report in the Adelaide News, both sides of the South Australian Government have agreed that critical comments on the web, notably the Adelaide News, about the coming state election, must be signed with name and post code (which assumes that the commenter lives in a country with post codes, and presumably that, unlike the UK, they're generic enough not to identify a single house):
Mr Atkinson described AdelaideNow as "not just a sewer of criminal defamation'' but also "a sewer of identity theft and fraud''.
Are they really that stupid? Or do they really think that the electorate is that stupid? Never mind that I personally disapprove of anonymous comments: this is a state law; even I (in a different state) could offend against it with impunity, at least as long as the web server on which my comment is posted is also in a different state (or, like the Adelaide Now web site, outside Australia: it's located in the UK). People outside Australia are completely out of harm's way.
And there are good reasons not to post your name and post code: one day, some brain-dead bank “security” person might decide that's all you need to apply for a Verified by VISA password.
But why this “law” in the first place? I don't know, but it certainly says nothing about the state of politics in South Australia.
| Topic: animals | Link here |
Woken round 5 am this morning by a noise outside on the verandah. Went out and found another (or maybe the same) possum. This time I didn't have any water, but there was a fly spray can nearby, so gave him some of that. Hopefully it didn't hurt him, but that it will keep him away.
| Topic: general, opinion, animals | Link here |
Into town today, primarily for a blood test at Dorevitch in Victoria St. That worked well: I didn't even get as far as taking a number before somebody took me in for the test, and it must have been over within 3 minutes of entering the door, a far cry for the half hours of waiting I've had elsewhere.
That was the only thing that went well. Yvonne had asked me to bring her back some Indian ink. She had thought that Officeworks would have it, which I doubted, so she told me where there was an artist's supply shop in Lydiard St., so went looking for that, to no avail. Off to Officeworks anyway, and asked for Indian ink. I knew I was in trouble when the girl replied “What's Indian ink?”. And, of course, they didn't have any.
Looked around anyway, and found little of interest until I left the shop. There they had listings of specials—not their own, but of Good Guys and Harvey Norman:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The trick: they have marked their own prices for those items which they sell at the same or lower prices than the specials. This is interesting for a couple of reasons:
The Good Guys always give you a last-minute discount, which really doesn't seem to be in their interest, as I observed last year. This comparison is another case in point: theoretically, the price of, say, the Panasonic DECT phones on the flyer (apparently the same) would be cheaper at Good Guys than at Officeworks. But wait, there's more!
As the second photo shows, they'll give you 5% off any matching price. You don't even need to do your own research: they've done it for you. So if you want to buy the Canon IXUS 801S camera bottom right on the Harvey Norman flyer, you don't have to pay the $177 for which both of them sell it: you have the evidence you need to get it for $168.15. Somehow Officeworks seem to be kicking themselves in the behind with this advertisement.
I'd have liked to take the absence of a flyer for JB HiFi as evidence that they're cheaper across the board. But that's not the case. Went looking for their compact digital cameras and found absolutely no model numbers, just descriptions. Even when I found the IXUS 801S, the only way I could identify it was from the image. Why do they do that? It must be deliberate. Are they trying to make it more difficult to compare? In this case, they're out of the running; they want $199 for it, over $30 more.
After that, to Sebastopol for a haircut and to pick up a parcel from the Post Office. The barber's shop was full, so walked down to the Post Office. On the way, passing a woman with a small dog, another bigger dog came past and jumped on the woman's dog. The canonical thing to do in cases like that is to stand back and let nature take its course, or pour a bucket of water on them. But the attack was so ferocious that I don't think the little dog would have survived if I had done that, so gave it a few well-placed kicks, and with the help of another passer-by, pulled the dog back and restrained it. No sign of an owner, so the bloke in the fish and chip shop called the police, and I sat by holding the dog, waiting for the police to arrive, while most people went away.
Finally the bloke from the chip shop brought me some wire to tie the dog up with, braving his fear of the dog to do so.
|
|||||||||||||||||||||||
In fact, she (a bitch) was quite friendly, and I can imagine that her owners would never believe that she would just attack another dog like that. But that's exactly the kind of dog that is dangerous. She was friendly (submissive) with me because it was clear I wouldn't take any nonsense. But if any animal shows fear, she might attack it, as in the case in point. The same could happen with a small child.
The owner had spent the time calming down in the laundromat next door, and we confirmed that her dog hadn't been hurt. But only a little earlier she had tied her (another bitch) to a street pole while she went into the Post Office. What would have happened if the other bitch had come by then? The ranger came along and picked her up while I was in the Post Office, so I didn't get to speak to him. Another woman who had stayed behind told me that he thought he had seen it before.
In the Post Office, they couldn't find the parcel. Seems they're not too happy with the delivery contractor either, though I have my suspicions about their own organization. Every time I go there to pick something up, it involves a significant amount of searching in many different places. Anyway, the postmaster filled out a form and suggested we'd have it in a couple of days.
Back to the barber—still full. I wonder what's special about today. Still, I can come back some other day. But what an effort for almost no result.
| Topic: photography | Link here |
Somehow I'm still frustrated with many of my photos. When I go into town I take my little Nikon “Coolpix” L1, not my favourite camera. The severe barrel distortion is very evident in the photos of the Officeworks flyers, and despite flash I wasn't able to eliminate camera shake in those photos (1/32 s, but no image stabilization). And the focus is terrible! I tried several times to get a close-up photo of the dog's registration tag, with no success: it was completely out of focus. It was just a matter of luck that it was showing in the one acceptable photo I got.
| Topic: food and drink | Link here |
Chicken tanduri for dinner today, and for the first time we have real Indian atta, so had another go at tanduri nan. Yes, the flour is better, but the results still taste more like Western bread. I think it must be the yeast. Next time I'll try sourdough.
| Wednesday, 3 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Censors back down
|
Topic: technology, opinion | Link here |
Yesterday I wrote about network censorship in South Australia. I was wrong, of course: on reconsidering, this isn't really censorship, thought that's the word most people are using. It's really a restriction of privacy. In a country where free speech isn't guaranteed (yes, that applies to Australia, though the general viewpoint is that it should be tolerated), the requirement to put your name to what you write suggests potential prosecution. That's doubly stupid: first, it makes the government look bad to exactly those people whom it needs to reelect them, and secondly it's as good as impossible to confirm whether the name and postcode given are correct or not, as Russell Coker observes.
All this is water under the bridge now; Michael Atkinson has recognized the adverse publicity (but probably not the uselessness of the law) and announced that it would be repealed after the election—assuming he's still there then. I doubt he will win back many friends. The poll on “Should Michael Atkinson resign as Attorney-General after being forced to repeal his laws censoring internet comment?” was an overwhelming “Yes”—over 91%. That's a stronger view than the than the previous poll about rejecting the law.
I voted “No”. There's more to the job of Attorney General than this kind of stupidity, and clearly he had the support not only of his party, but of the opposition as well. Should we require both parties to abstain from the election? Let him stand up to a more representative poll. Of course, there's still the possibility that his party will dismiss him, but that would make a scapegoat out of him.
“Performance” always seems to be about the number of instructions (whatever that may mean) a processor can perform in a given time, the amount of data I can transfer over a network or to and from storage devices, or the number of frames a video card can display per second. And they're continually on the increase. My first home computer had a Z-80 processor with 2 MHz clock, which corresponded to about 400 kIPS. The machine I'm writing on now has a 4 core AMD Phenom processor running at 2.2 GHz and executing, for argument's sake, about 2 instructions per cycle—a total of about 18 GIPS, well over 20,000 times as fast as the Z-80. The Z-80 was connected to a terminal which initially only handled 1200 bps—it took 16 seconds to write the entire screen—while now I have memory mapped displays which can display a screen with no noticeable delay. The Z-80 had floppy disks, and though I was running CP/M, I had written a format program which compensated for the CP/M sector layout brain damage and allowed disk transfers at up to almost 20 kB/s. Now I have disks that can transfer at up to 50 MB/s, 2,500 times faster. Everything is better.
But is it? Modern software is equal to the challenge. More and more, I need to turn to the web for interaction. And even with network speeds of up to 4 Mb/s, it's slow. Reading through a web forum with, say, 20 entries, reminds me of the Z-80 days. Displaying a big version of my photos with firefox takes up to 20 seconds elapsed time—close to the Z-80 times—and uses about 5 seconds CPU time—corresponding to about 7 hours on the Z-80. And people have come to the conclusion that processors are too fast, and that slower, less power-hungry processors are the way to go.
It's not just computers, either. A year ago I bought a radio/CD player (“Mini HiFi”) from Sony. It's digital, of course, and it can do things like play CDs with MP3s on it. The controls are at times glacially slow, and the strange indicators don't help. It has a red LED which is only on when the device is off, so it's easy to press the power button to turn it on when it's already on. If you do, you have to wait 5 seconds while it displays STANDBY before you can turn it on again, followed by another 5 seconds to change from tuner to CD or back.
What's wrong with this picture? I can only assume that most people aren't as impatient as me. But on the other hand, processor power is enabling new technologies, like computer video recording and display software such as MythTV and mplayer. That same photo that takes 20 seconds to render under firefox comes up in under a second with xv—but that's an old program that nobody uses any more. So why do people put up with slow, web-based software? An obvious reason, which applies to me, is that there isn't anything else. But why?
| Thursday, 4 February 2010 | Dereel | Images for 4 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Weather station software: structural issues
|
Topic: technology, general | Link here |
My weather software is still in a mess. One process regularly crashes after midnight because there is no data available for the day. That's harmless (it gets restarted, and sooner or later the data arrives), but also trivial to fix. I thought.
The problem is that I keep restructuring things, particularly configuration, and it seemed to make sense to have a standardized set of parameters for each process (in fact, it makes more sense for the main process to fork the others), so spent some time restructuring the code for startup and reading configuration. That took longer than I expected.
And today was the first day in nearly 3 weeks that we've had rain, 14 mm of it in the rain gauge. But the weather station didn't report any. Software bug? Steve Woodford's spiders? I can't see that there's anything that I've done in the last 3 weeks that would change the way the software handles rainfall, so it sounds like spiders. Time to get a ladder out—when the rain stops.
| Topic: technology, general, opinion | Link here |
Mail from the Golden Plains Shire for a Hamish Everett today, not for the first time. I've sent two back with the text “not known here”, and it's clearly not getting through to them. Went to their “Contact Us” web page, and found a form that required an email address (and accepted donotuse@example.com), and filled that out. As expected, the copy (sent to my real email address) was eaten by SpamAssassin—all HTML, and with my carefully formatted text (four paragraphs) swallowed up as a single line:
Of course, most government bodies appear to dislike this kind of form at least as much as I do, so it'll be interesting to see when and how they reply.
As I mentioned last month, I think that SpamAssassin has passed its use-by date. The Bayesian analysis seems to award high marks to innocuous messages and to ignore (probably carefully crafted) spam. I hate HTML mail more than most people, but that's exactly the reason not to consider HTML mail to be an indicator of spam. In short, it's only useful for extreme spam, and even then it first has to receive it, giving the sender's MTA the warm fuzzy feeling that things have worked.
There are alternatives, and some people have pointed me at TMDA, the Tagged Message Delivery Agent. Can it do what I'm looking for, specifically get the receiving MTA to refuse the message? I don't know. I can't find any documentation that tells me how it works. There's plenty of (partially conflicting) information on configuring it, and I found some more on Safari in Slamming Spam: A Guide for System Administrators, but it's still a cookbook, not a description of how it works. How I wish people would realize that they have to sell this software, even if it's free.
| Topic: gardening, animals | Link here |
Found an interesting beetle in the garden today. I wonder what it is:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Our plants are handling the heat (which finished with rain today) with differing degrees of success. Our “Gruss an Aachen” rose is feeling very happy:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On the other hands, another rose planted at the same time, “Monsieur Tillier”, is looking quite unhappy for itself, and our ornamental Japanese cherry is almost completely defoliated:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
What's causing this? There are a number of candidates:
The heat and the wind. That's possible, and it has certainly caused damage, but some of this damage looks like things have been cut off with something sharp.
Kangaroos. We've had this kind of problem in the past, and we've blamed it on the kangaroos. But we haven't seen any evidence of kangaroos in the garden in months; maybe the damage last year wasn't all the kangaroos' fault.
Other animals. But what? We've seen rabbits here, and possums, but the tip of the cherry tree is about 1.5 m high, too high for a rabbit or a possum on foot. I don't think that a possum could climb the tree without damaging it further (snapping the trunk).
Caterpillars. But we haven't seen that many, and it's strange that they should attack the cherry and one rose (some distance away) and leave the other rose (“Gruss an Aachen”, next to the cherry tree) untouched. It also doesn't explain damage to the grass nearby:
|
|||||||||||||||||||||||
The “Gruss an Aachen” still had a wire mesh around it; the others didn't. Is this maybe some big mystery animal that leaves no traces? Put the mesh back on the cherry tree; I think I'll have to do it for some others too.
| Topic: general | Link here |
Yvonne's mail parcel has finally arrived—sent from Germany by Jana Desselmann with Deutsche Post's “EuroXPress” about 2 months ago, though the ambiguous date they use (07.12.09) could be interpreted either as 7.12.2009 or 2007-12-9. If people have to be so pedantic as to include leading 0s, why do they truncate things and introduce more ambiguity?
The parcel's interesting for two things, not including getting lost for a couple of days: it contained foodstuffs (a no-no for sending to Australia), but it stated so on the customs declaration (“Weihnachtsgebäck”, which clearly nobody could understand). But quarantine declarations aren't in use in Europe, and in case of doubt customs or AQIS should have asked.
The other thing, of course, is the time it took to get here: two months, reminiscent of the problems I had with c't last year. Things don't seem to have got any better.
| Friday, 5 February 2010 | Dereel | Images for 5 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Weather and spiders
|
Topic: technology, general | Link here |
More investigation of the weather station today to find out why no rain had been reported. Reinstated my memory dump program, which still has problems with reading particular locations—high time I tested this on another system. It showed that the station had really recorded no rainfall at all over the last 36 hours. So it's not a bug in my software.
Decided to investigate the device. I have a spare, thanks to Powercor, so took a look at it. It has a cover (held in place by press tabs at each end), and a bucket that tips alternately in each direction:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Went out to take a look at the one at the top of the pole. Steve Woodford was right. Spiders had built a nest inside:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Cleaned that out and sprayed some insecticide in the cover; I suspect I'll have to do this from time to time, especially if there hasn't been any rain for a while.
| Topic: photography | Link here |
Tried to take some photos of the spiders' eggs inside the cover of the rain gauge. Not easy: they're only a couple of millimetres across, and getting the lens close wasn't easy. I can't do this with the Olympus lenses, because they don't work properly on bellows, and a single Olympus extension tube costs more than most macro lenses, so used the Pentax lenses instead. And then my bellows broke. I've had problems with them for some time (decades, probably). I must once have tightened up the locking screw for the front part too tightly, and broken off part of the brackets that held it on the tracks. Here the bellows seen from underneath, with the front part on the left, and the (intact) rear part on the right. The plastic wedge on top is part of the clamp.
|
|||||||||||||||||||||||
I've “repaired” them (replaced the bits where they belong) in the past, but this time it didn't help; maybe I've lost a plastic wedge. What do I do now? Olympus doesn't do bellows (probably a good thing, looking at the prices for other close-up equipment), and until I can resolve the issues I have with the old Pentax lenses, I don't know if I want to invest anything in them any more. Took a look on eBay and found a couple of second-hand M42 bellows on sale, but neither appear to have a cable-release contact (bottom left, to the right of the clamp screw, on the photo above), and I need that to be able to stop down my 135 mm lens. More head-scratching needed.
| Topic: gardening | Link here |
While taking photos of the damage to the cherry tree, stumbled across another mystery plant:
|
|||||||||||||||||||||||
Well, is it a mystery? It looks like a potato. But if it's a potato, the mystery is how it got there. This is in a somewhat neglected bed that was planted on top of what used to be lawn, using soil purchased over two years ago. How can a potato get there? I'll keep an eye on it and find what it really is.
Another mystery plant is the one that Chris gave us a while back, and that we replanted into a box in the kitchen last week. At the time I was concerned that it might not be happy in the low light conditions. It seems that I didn't need to worry. Here is what it looked like shortly after planting and today:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's interesting how it turns the leaves to face the light.
| Topic: technology, general, opinion | Link here |
Another letter for Hamish Everett today, and of course I had not received any reply to my message to Golden Plains Shire. Sent another one, this time carefully enclosing my paragraphs in <p> tags, to no avail: they were stripped. And the reference numbers were sequential:
Clearly not a service that is used much. I should try again in a few weeks to get a feel for the number of messages sent.
Shortly later, got a call from Karen at the council. Yes, she had got my message yesterday, and she had gone around the departments trying to find who had sent it. That makes sense enough, I suppose. So we agreed I'd send these two back again, and hopefully that will be the last of it. At least my pessimism about the delay was misplaced.
| Topic: brewing | Link here |
Tried out my new Dunkelweizen today. Grrr! It's too bitter, again! That's got to be partially due to the reduced volume, but it's too much to be explained only by that. Is it possible that my own hops, which I used for first wort hopping, and which I had considered to be Tettnang, are in fact something else, more bitter? How do I find out? I'll need to be more careful with my home-grown aroma hops in the future. And I suspect my Farbmalz wasn't dark enough:
|
|||||||||||||||||||||||
| Topic: brewing, technology, opinion | Link here |
Callum Gibson came up with the idea of freezing the beer to get rid of some of the bitterness, quoting the German Beer Institute, which, as the name suggests, is located in the USA. The markup is so completely broken that I could hardly read it:
|
||||||||||||||||||||||
It seems that this is not just the case for me this time; Callum sees it too. But he didn't bother to mention the workaround:
But it's sad that people have to take it for granted that you need to turn off page styles for some sites.
| Topic: food and drink | Link here |
More experimentation with nan today. Yesterday I took some really old (9 months) sourdough starter and mixed it with some atta. Not completely to my surprise, it had risen nicely by today, so mixed in more atta and left it to rise.
I completely miscalculated, and ended up having to add about 50% more atta before the dough was of the right consistency. Rolled it out and baked half of it in the oven, probably too long; it was crispy, which isn't right. But of course it didn't have this yeasty taste of the nan I made the other day. I'll keep the rest until tomorrow.
| Saturday, 6 February 2010 | Dereel | Images for 6 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
One place where we do find “weed” potatoes is in the small succulent garden to the north of the verandah. Before we planted the succulents, it was a mess which had proved to have some potatoes in it, and we harvested a number of them about a year ago. But it seems we didn't harvest all of them, and I've been continually breaking off stems that come out of the ground, conveniently under some large succulent. Today I found one that had escaped me, and was now about 40 cm long. Pulled on it and got more than I expected:
|
|||||||||||||||||||||||
Also found some caterpillars on the ornamental vines:
|
|||||||||||||||||||||||
My inclination was to spray them, but the gardening books say to just take them off unless there are a lot. I found a total of four, so followed the advice, including wearing gloves: the hairs can irritate the skin. That's not all, it seems: one of them bit into my gloves, exuding a quantity of greenish liquid. I don't know what my skin would have thought of that. Next time I'll take a photo.
| Topic: food and drink | Link here |
An experimental dish for dinner today: a cross between my ikan goreng recipe and a fish tikka kebab recipe that we found in an Indian cookbook. I was going to call it “Ikan tikka kebab” to indicate its mixed origin, but unfortunately I came to the conclusion that it wasn't good enough.
Ate the rest of the nan with the fish, this time grilled on the plate of the barbecue. I think that's probably the best way to do it. Once again it was overdone, but only a little. I think I'm gradually getting the hang of this.
| Sunday, 7 February 2010 | Dereel | Images for 7 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Error reporting: for wimps
|
Topic: technology | Link here |
One of the more obvious misfeatures in my weather station software is the comparison between local readings and remote weather stations. My local readings are typically once a minute, while the best I can get from remote stations is every 30 minutes. In some cases (like pressure readings from Ballarat airport), the readings are three hours apart. From a programmatic point of view it's easy enough to interpolate, but what meaning does such an interpolation have? Combined with gnuplot's strange smoothing algorithm, the result looks like a roller coaster. So today worked on changing that, to take the local reading immediately following the remote reading. With less than 60 seconds error I don't even need to interpolate, though I haven't fixed that yet.
Once again I spent hours trying to find out why things didn't work. In case of doubt, they say, the biggest source of errors in UNIX are permissions problems, and that's what it proved to be this time too. But the way you put together web applications, it's difficult to find the errors. The standard environment simply ignores them. My program happily printed out an error message, but it ended up somewhere behind the root window. I think I should call syslog instead of fprintf as a matter of course.
| Topic: gardening | Link here |
I'm finally doing a little more in the garden—leave things just for a while and the weeds take over. One pleasant discovery was the Crassula “Springtime” that I thought had died. I had planted it in the Japanese Garden, and it has already flowered. It's certainly far from dead. Here 10 months ago (top left) and today:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Another surprise was the Aeonium haworthii about which I wrote 6 weeks ago. I don't like them much because of their trumpeting flowers, while Echeveria produces a flower stem from the side:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Or so I thought. Today I found the same Aeonium with flowers coming out between the leaves:
|
|||||||||||||||||||||||
It remains to be seen whether they'll turn into trumpets or not.
| Topic: opinion | Link here |
Discussion on IRC today: Ben, one of our youngest members was extolling the virtues (if that's the word) of a computer game he was playing. The goal of most computer games still seem to be to kill your opponent. Ben defended that: “After all, it's just a game; I wouldn't really kill somebody”. And that's a fairly accepted opinion (and clearly I believe that he wouldn't kill somebody). But it's still violence, and it seems that games like this are becoming more violent as time goes on. c't magazin recently had an editorial on the subject of a new game where part of the game involved killing hundreds of civilians in an airport.
The general opinion on the channel was this is not fun by any stretch of the imagination. But how do you bring this across? One thing on which the vast majority of people agree is that child pornography is bad. But how bad? How much worse than killing? Exploited children may suffer severe psychological damage, but they're not dead. Yet even the possession of child pornography is a criminal offence. Pornography depicting consensual sex is also heavily restricted. Why should people be allowed to market games which require simulated violence, where the target audience is typically children or adolescents? Why are so many films (particularly, it seems, from the USA) so violent?
| Topic: food and drink, general | Link here |
Chris is doing training in Melbourne on Saturdays now (consensual violence), so we've moved our weekly dinner to Sunday. Something meaty for a change: we had started with thoughts about filet de boeuf Richlieu, but Yvonne didn't like the idea of braised lettuce, so she ended up doing her own take on filet de boeuf:
|
|||||||||||||||||||||||
| Monday, 8 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More old laptops
|
Topic: technology | Link here |
I've given up on using the old ThinkPad for the kitchen. Modern bloated web browsers just won't run in 96 MB of main memory. I don't want to buy a new disk for my Dell Inspiron 1100: it needs PATA disks, of course, and they're ridiculously expensive in comparison both to SATA and to cheap new laptops. But Chris had an old Dell Inspiron 8100 with some defect or another, but presumably a functional disk, so today Yvonne brought it over and I took a look. I remember working with this machine a couple of years ago, and that's what the installation of SuSE Linux claimed too—the battery was still functional! For the fun of it installed FreeBSD 8.0 on it, which worked—but I had no X at all! Somehow things aren't getting any easier.
| Topic: general | Link here |
More hot weather. The Bureau of Meteorology has recently been changing their minds about the day's temperatures. Yesterday morning the forecast for Ballarat was:
Yesterday evening they changed the forecast:
And so it was, though not too much cloud. We had a high temperature of 38.5°. Finally, this evening they changed the forecast for Tuesday:
The temperatures aren't overly high for the time of year—this time last year we had temperatures in the high 40s—but it's annoying when you hope the weather will cool down and it gets hotter instead.
| Tuesday, 9 February 2010 | Dereel | Images for 9 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Facebook adopts bank-level security
|
Topic: technology, opinion | Link here |
Something prompted me to look at Facebook again today. When I did, I was presented with a surprising message:
|
|
||||||||
That's doubly interesting. First, I was accessing it from the same browser as usual, so any cookies should still have been there. And secondly, this STUPID idea that only I would know my date of birth. I suppose they can justify it, though, with “if it's secure enough for banks, it's secure enough for us”.
| Topic: general, technology | Link here |
The (belatedly) promised hot weather came, but also a lot of rain. Yes, the rain gauge in my weather station is now reporting rain again; but it's now reporting less, not more, than before. The manual rain gauge collected 11.2 mm of water, but the one on the weather station only reported 7.8 mm. Normally it reports significantly more, not less.
But why? Looking at the design, it seems reasonable. Steve Woodford points to potential issues: wind can cause the buckets to tip prematurely, thus reporting more rainfall than usual. But my excessive measurements seemed to be independent of wind. Spiders can jam up the works, but they're gone now—I think. More thought needed, but the obvious conclusion is that the manual rain gauge isn't superfluous.
| Topic: technology | Link here |
Spent some time installing FreeBSD 8.0-RELEASE on the Inspiron 8100 that Chris gave me. It worked well, but at the end I discovered that it hadn't installed X! Why? At a time when other systems are installing a more and more complete environment, we're cutting back. Went off to try to install it from the ports, and of course first had to install myriad dependencies. It failed building the Perl port. Sigh. There used to be a time when installing operating systems Just Worked. This is progress?
| Topic: general | Link here |
Despite the rainfall, we had problems with the house water supply (from the rainwater tanks). It tailed off to a trickle. Went to investigate and found the filter relatively full of crud, but not, it would seem, enough to cause the problems I've seen. We have two tanks, and I usually have one turned off as a reserve, so turned that on, and things seemed to work again. Maybe there's some blockage in the first tank.
| Wednesday, 10 February 2010 | Dereel | Images for 10 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Satellite dropouts: why?
|
Topic: technology | Link here |
This hasn't been a good month for satellite availability. I've had a total of 45 dropouts, a total of 84 minutes:
The interesting thing, though, is where they're happening. I don't know, but in the past I have been blaming IPStar. But lately Chris Yeardley, also using IPStar, has been joining us on IRC, and during the dropouts on 8 and 9 February she remained connected. So it can't be the satellite itself.
I should contact my ISP, of course. But in the past I've been told it's probably the modem, and that I should disconnect my cables at both ends to allow accumulated charge to drain. I don't need that kind of nonsense. It could be the results of a memory leak in the modem, something that I definitely wouldn't exclude, but the pattern of the dropouts doesn't look right for that. I had intended to power cycle the modem anyway, but I forgot, and after that there were no more dropouts anyway.
Back to looking at the half-complete installation of X on the Inspiron 8100 that Chris gave me. Problem: it could find neither the disk nor the CD drive. Well, we had suspected that the disk controller was defective, so took out the disk and put it in my Inspiron 1150. And it failed there, too. So at least for the moment, it's another dead end.
| Topic: general | Link here |
The satellite dropouts weren't the only problem; we had no fewer than 4 power failures in the afternoon, in the space of 2½ hours. They were all very short—auto recloser, as Powercor calls them—but like the satellite dropouts, they're still a bloody nuisance. Called up Powercor on 132412, and spoke to Daniel, who explained to me that the auto reclosers were doing what they're designed for. He made it sound as if power failures were a feature, not a bug. Pointed out that 4 in such a short time must point to some other problem, and got him to make a report, though I doubt that much will come of it.
Yet another last minute change in the weather forecast. Here today's forecast, as of yesterday morning and this morning:
And indeed it get nearly as hot as yesterday; in fact, the temperature development was quite amazingly similar. Here the last 5 days. Yesterday is the purple line, today the cyan line. In each case there was a similar temperature drop (associated with heavy rain) round 14:00:
Once again we had no water pressure today. Investigation showed that the filter was clogged again, though again not very much; possibly there's a lot of fine stuff in there which causes the problem. Looks like we might have problems with it for a while.
| Topic: gardening | Link here |
After seeing the damage done to my Gruss an Aachen rose, I've put mesh around most of the plants. In the process came across an extraordinary sight:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Something has gone across the Canna leaves with two rows of equally-spaced holes, some of which show traces of what looks like something being dragged across the leaf. What is it? The Thing that came out of the Swamp? I can't imagine anything that would cause that kind of damage. It's also not clear whether it's in any way related to the other damage, but I did find some moist droppings nearby, unlike anything that I'd expect. Maybe it was a cat or something, and unrelated to the damage.
The Hibiscus rosa-sinensis cutting that I got from my uncle Max is now growing happily and has produced its first real flower—the biggest I have ever seen. It's 18 cm across. Most are closer to 10 cm. And one nameless person on IRC liked it so much that he used one of the photos for his screen saver:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||
| Topic: brewing, food and drink | Link here |
The Indian grain moths that I found a while back seem to have escaped and attacked my grain and flour supplies. How do I get rid of them? Nik from Grain and Grape suggested filling the containers with carbon dioxide, but I don't think the seal is tight enough for that (thus also explaining how they got in there). This is a potential disaster; I have about 100 kg of various grain. One thing makes sense: brew another brew as soon as possible. But that'll only use a fraction of the supplies.
| Thursday, 11 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: brewing, food and drink | Link here |
I'm in trouble with my malt and flour stocks: these horrible moths (which I've been calling Indian grain (malt) moths (flies) are everywhere! I'm going to have to get rid of them or throw away about 70 kg of grain and flour. How do you kill them? Sent out a message to the homebrewers' digest, and also did some investigation on the web. That produced some interesting information:
The correct name is Indianmeal moth: “Indianmeal” is an old name for maize meal, and that's where they seem to come from. Nothing to do with India.
It seems that you can kill them by freezing (0° F, or -18° C, which seems to be the reason why European deep freezes have chosen this temperature). A pity we don't have enough space in the deep freeze.
Some also claim that you can kill them by warming to 50° for a couple of hours. That seems unlikely to me: that's only marginally hotter than it gets here outside in the summer, and they seem to have done well. Still, it's nothing I can change right now.
You can also catch them with pheromone traps, but I don't have any source for them here.
In the end, noted that the larvae seem to congregate at the top of the container, so opened it up, chased off all the moths and scraped off all the larvae. There's a very good chance that there are more larvae or eggs in the grain, but I haven't seen any. So maybe I can limit the exposure by getting rid of the moths before they have time to lay too many eggs. L'espoir fait vivre.
| Topic: general | Link here |
The house is in a horrible mess! I've never been tidy, but lately I've been neglecting lots of things. Spent some time today trying to make up for it.
| Friday, 12 February 2010 | Dereel | Images for 12 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
MythTV hangs: hardware or software?
|
Topic: technology | Link here |
Problems with cvr2, my MythTV box, this afternoon. A programme was flagged as “recording” long after it should have finished, and the file was empty (according to MythTV; in fact, it didn't even exist). Spent a lot of time looking for reasons, and it seemed that one of the three tuners had failed; but which one? Spent about an hour removing and replacing them in various combinations, without being able to find the one that was defective; it seemed that the problem was wandering from one to the other. Got to a state where two tuners were working—that's enough for the next week—and decided to consider what the problem might be. Possibly there's something wrong with the cables, or maybe it's just a low signal situation. How do you measure these things?
| Topic: gardening | Link here |
Going round the garden this morning, found what looks like some of the most enormous kangaroo droppings I've ever seen:
|
|||||||||||||||||||||||
No signs of any damage, though. Callum Gibson pointed me at an ABC article on identifying droppings, which I'll do some time soon.
Also more puzzlement and discussion about the perforated Canna leaves. We're considering caterpillars and similar. I didn't look under the leaves before, but at least now there's nothing much to be seen:
|
|||||||||||||||||||||||
Edwin Groothuis suggested it might have been a Wentelteefje, for which there appears to be no English word:
The question is: where is it hiding?
| Topic: brewing | Link here |
Spent some more time trying to free my grain of Indianmeal moths, with some limited success: as I had feared, there were plenty more larvae in the grain itself, but I had left it outside in the light, and that seems to have drawn them out; squashed about 30 of them in the course of the afternoon. What we need now is more light, I suppose.
| Topic: general | Link here |
We've been thinking of subdividing our property for some time, and we've got to the stage where we have decided to involve a couple of estate agents to see what we could get for the old place. Today got a call back from Bram Gunn, the bloke who sold us the house nearly 3 years ago. He had lots of useful information, including the comparison between subdividing—which could involve significant problems with the authorities—and selling the entire property and moving elsewhere: the former has an advantage of $30,000 to $40,000, not to be sneezed at.
We're not very keen on the alternative of moving elsewhere, but for the fun of it took a look around what's currently on the market. There are a whole lot of plots of land for sale at the other end of Browns Road, behind Chris' property. Off to take a look—5 km of dirt road, and despite the photos, pretty dense bush, difficult for fire crews to get to. Then took a look around the area. It wasn't easy looking 3 years ago, and now we're not in a hurry, it's no easier. Subdivision sounds like a doubly good idea.
| Saturday, 13 February 2010 | Dereel | Images for 13 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
House hunting
|
Topic: general | Link here |
More thinking about building a new house today, and into Ballarat to look for plots of land and at display houses. Out to Macarthur Park in Miner's Rest, where we've been numerous times before. This time I wanted to look primarily at the builders we haven't seen before. That was a surprise: they've finished building there, and there was only a single display home still operative, from the National Builders Group. I had a recollection of some pretty average houses from them, but this one (“Beaufort 270E”, 33.55 “squares”, really about 240 DIN m²) didn't look bad. Some things are just ridiculous, of course, such as the 15 m² bathroom, but it looked as if we could make something of it. It's surprising how much more spacious it looks than our house, which is also 240 m²
Then on across the road to Simonds, who had a couple of somewhat boring houses to look at, with a standard of finish that says little about their quality:
|
|||||||||||||||||||||||
Still, even these houses have some good ideas, like the patio that goes round two sides of the house:
|
|||||||||||||||||||||||
Out of interest, asked the girl inside what the houses cost. To my surprise, she said something about $420,000—more than twice the cost of most of the houses. Looked around and saw a price of $175,000 on the wall—but she continued to insist that it was over $400,000. What company needs hired help like that? Left with the feeling that we would probably not be coming back to Simonds.
The new developments in town are in Alfredton, to the West of town, so down there and found a whole heap of display homes in close proximity, including some companies we had never heard of before. First to Cavalier Homes, not the sort of name I would have chosen for a company, and saw what was a relatively boring looking house with rooms scattered apparently randomly around the house. Asked what they had to recommend themselves, and they said that they're a local company—many are based in Melbourne—and that they could get things done more quickly. Certainly the price of the display house ($189,000, but for only 190 m²) was at the low end. But it's difficult to confirm their claim of being local: they have two different domain names, cavalierhomes.com.au and cavalierhomesballarat.com.au, which suggests that they are no different from the others. I can't confirm that, though: their web site is unreachable.
Next was Bryan and Petersen, who advertised with their energy efficiency—they really did have double-glazed windows, the first I've seen in a display home. But they don't insulate the slab, so there's a question how useful it is. With 230 m², the cost was also relatively steep—$235,000. Still, the layout wasn't as painful as some.
Then to homesnow (as near as I can see, they write it in lower case, which doesn't improve things when not written with their fancy font). They advertised with the low price of the house: 304 m² for only $167,900. I think they're including not only the garage, but also the “al fresco” in that sum, and that the real area is closer to 250 m², but that's still pretty cheap. The front door was almost enough for us to give up, but on looking inside they had probably one of the best layouts that we saw. Did a bit of talking with Bill Hancock, their business development manager, and I was left with the feeling (which I haven't yet followed up on) that they charge extra for many things that the others include in the price.
Finally, flagging a bit, to Ryan & Mcdonald, who showed us yet another house looking pretty much the same as the others, the “Monash 35”, allegedly 335.51 m²; after removing the non-living area, they're left with 243 m², about the biggest discrepancy I've seen so far. So it was with the price, too: base price $210,000, price as seen $320,000, including over $10,000 for a silly dual-ported slow combustion heater, which would have been the first thing to get thrown out even if it were free. But they offer full architectural support; the price is based only on the area. It's certainly worth following up on.
|
New Peruvian plants
|
Topic: gardening | Link here |
Our trip took us past the Botanical Gardens, so stopped in to see what was going on there. Apart from a surprising number of old (pre-911) Porsches, there was a small Farmer's Market going on, and bought an Alstroemeria rhizome (if that's the word). They couldn't tell me which species, so they gave me a flower to take with me:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Looking at the Wikipedia page, it's almost certainly Alstromeria aurea, also called “Peruvian Lily”.
It's been over 4 months since we were last in the Botanical Gardens, and I was interested to see how some of their succulents have grown in that time. Here a photo from last September, and from today:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On the face of it, there's not much difference. Also found many more plants on sale; this seems to be a permanent thing. Managed to hold ourselves back and buy only a Zephyranthes candida, which looks almost like a Crocus, but blooms in summer:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Once again there's a connection to Peru: it's calso called “Peruvian swamp-lily”, amongst many other names. But I fear that's all the flowers we'll see this year, though, so we might as well keep it in a pot.
| Topic: brewing | Link here |
My mail message about the Indianmeal moth problem has brought a number of responses, from the obvious “throw it all away” to a significant number of “freeze them”. Now if I only had space to put the stuff in the freezer.
Yesterday's exposure to the sun had drawn out a large number of moth larvae out of one of my grain containers, so today tried four more. It wasn't nearly as successful: I only got a couple. Of course, we can hope that that's all there is, but the alternative is that they just haven't hatched yet. More observation needed.
| Sunday, 14 February 2010 | Dereel | Images for 14 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
DVB-T woes revisited
|
Topic: technology, multimedia | Link here |
Took another look at cvr2, which had mysteriously stopped recording a couple of days ago, and which at first looked as if one of the tuners was defective. Callum Gibson pointed out to me that yes, indeed, there are tools for measuring signal strength. I should have remembered that; I've talked about them in the past in connection with tzap, for which I wrote a lot of the documentation. But that was what I needed on Friday, not today. Put the third tuner card back in the machine, and not completely surprisingly, it worked. So maybe it was a signal strength issue. Some day I should run tzap against all three tuners and all channels and get base values. But that'll take time, especially since I need to stop MythTV to do so.
| Topic: music, opinion | Link here |
While testing cvr2, discovered that they were broadcasting Così fan tutte, so recorded that. They call it just “Cosi”—or do they? They pronounce it to rhyme with “Aussie”. Ugh.
| Topic: gardening | Link here |
Planted the Alstromeria, not before time. It had shoots when we bought it, and they're looking decidedly less happy now. Still, they come out of a rhizome, so they might recover. Found we had three pieces—I thought that there was only one, but maybe it broke. Planted in front of the dining room door, in the other half of the north bed, and in front of the verandah.
Our tomatoes are gradually ripening, but this year we're not getting away without some insect attacks:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Time to look for appropriate insecticides.
| Monday, 15 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
DVB-T signal strengths
|
Topic: technology, multimedia | Link here |
Finally got round to doing some measurements of typical signal strengths for my DVB-T tuner cards. tzap is the only program I know that displays the strength, and it does it in hex, once a second, not the easiest format to work out averages:
=== root@cvr2 (/dev/pts/0) /home/grog 11 -> tzap -a 1 SBS
But what do the values mean? Some of it is documented on the wiki page, but even between the tuners, the values aren't consistent. The example above is from adaptor 1; comparing the output from each adaptor, we get:
It's clear that the snr (signal-to-noise ratio) field is 8 bits wide, and the second 8 bits repeat the first 8 bits. But comparing snr to signal gives no obvious relationship. In addition, the third adapter always returns a status of 1e, not 1f; according to the documentation, this means that it doesn't have a signal. But signal and snr contradict that.
Also, the second adapter (number 1, the first example) shows a strange status after starting tzap. The first FE_HAS_LOCK is always accompanied by a large number of uncorrectable errors (unc), followed by 5 seconds with no errors, and then 6 seconds with the same ber (bit error rate) and no uncorrectable errors. After that, it reports no errors. The values change from run to run, but the pattern doesn't.
Still more playing around with laptops today. Chris had a 2½" external drive that was defective, so she gave me that; it proved that yes, indeed, it was the housing electronics and not the drive that was defective, so finally I seem to have a functional disk drive for my Inspiron 1150. Started installing Ubuntu on it, and left it until tomorrow.
| Topic: gardening | Link here |
A little work in the garden, mainly weeding. The March flies (not very good with calendars) are out in force, and I couldn't go for more than a few minutes without being attacked. That was enough to find a bumper crop of garlic, though.
| Tuesday, 16 February 2010 | Dereel | Images for 16 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
GUI window managers: the poverty
|
Topic: technology, opinion | Link here |
So my Ubuntu installation is now complete. Or is it? Clearly I need to customize it, and I'm getting a bit of routine with that. But the more I work with “modern” GUIs, the more frustrated I get. Even if I ignore programs which seem to have escaped from the Microsoft space, the traditional programs are a problem. What key click do I need to iconify a window? In my fvwm2 configuration it's Alt-Button 2. To lower a window? For me it's Alt-Button 1, and both have been like that for over a decade. But all GUIs that I know require you to move the mouse to a field somewhere on the title bar and click on it. That's particularly annoying when that field isn't visible. No wonder Tabs are taking off: they're working around serious lack of functionality in the desktop manager software.
And, of course, every release of Ubuntu looks different. That's probably the case for all commercial software: they want to have something to sell. But that requires people to learn new things. On the one hand, progress is good, but to me this looks more like market differentiation than progress.
Went around looking for fvwm2. It's not in the standard packages. I could compile it, I suppose, but that's so foreign to the Ubuntu way of doing things that I can see multiple problems there.
| Topic: technology, multimedia | Link here |
More looking at DVB software today, and Callum Gibson pointed me at a couple of tools. The first is femon, which of course has no documentation, but which appears to be related to tzap. It prints out the same status information, but it doesn't require me to stop MythTV first.
The other is dvbsnoop, not your average multimedia software—for one thing, it has a relatively good man page—and which prints out far more information than I can digest at the moment. A pity that some of the examples are wrong:
dvbsnoop -pd 2 -n 1000 -s bandwidth
Analyze bandwidth usage of a PID.
In this example, stop if more than 1000 packets are read. for different output information, you can tweak the -pd option.
All that brings is:
=== root@cvr2 (/dev/pts/0) ~ 74 -> dvbsnoop -pd 2 -n 1000 -s bandwidth
The problem is that you need to specify a PID as well, and you need to find that with other invocations of dvbsnoop:
=== root@cvr2 (/dev/pts/0) ~ 77 -> dvbsnoop -pd 3 -s pidscan
=== root@cvr2 (/dev/pts/0) ~ 78 -> dvbsnoop -s bandwidth -n 1000 -pd 2 512
| Topic: gardening | Link here |
More work in the garden, punctuated by attacks from March flies. Inside instead and investigated a new web site, plantthis.com.au, which gives suggestions of what to plant based on your location. That requires a lot of input, of course, but what I've seen, it doesn't work badly.
| Topic: photography, technology | Link here |
Took a nice photo of Piccola the other day:
|
|||||||||||||||||||||||
The problem, of course, is the door and door frame in the background—a typical application for retouching. I've recently received the c't special on digital photography, which includes articles about retouching with GIMP, so decided to try that.
What a disaster! After an hour of messing around, I still don't know. The article (by Bettina K. Lechner) uses terms I've never heard of. For example, it tells you to select the “Clone tool” from the toolbox (which contains only icons). To find out which is which, you need to hover the cursor over each icon for a few seconds to recognize it:
|
|
||||||||
And that's only after you know that the icon is hidden somewhere on this page. Since this was in German, I thought it might be named differently (it wasn't), and went off on a tangent. The article also includes lots of key combinations (Ctrl-C, Ctrl-B, Shift-Ctrl-N). And repeat. Repeat what? Went looking for GIMP help, but the FreeBSD port doesn't seem to believe in documentation, so it didn't get installed. Looked it up on the web, comparing the German and English versions in case they were different (they weren't, at least not where I was looking), and still I don't understand the underlying principles. But who cares about principles? Just Do It!
What a pain!
| Wednesday, 17 February 2010 | Dereel | Images for 17 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Old laptops: the pain
|
Topic: technology | Link here |
More searching for fvwm for Ubuntu today, and of course found it on the fvwm web site. Continued with the installation on my Inspiron 1150, but didn't get far before a display hinge broke off. I can forget that machine for use in the kitchen. What a pain: I had just bought a new battery for it. But it's no longer clear that the Inspiron 8100 is as broken as I had thought: there's a good reason to believe that the problems I had had were with the disk. It must have been an interface issue, because it caused the disk controller not to recognize the DVD drive either.
Put the disk in the 8100, and sure enough, everything worked. Well, almost everything. I couldn't get any network connection with the old 802.11b card. Nothing obvious in ifconfig: it was up, running, had an IP address, but there was no association information (should there be?). And all the tools behaved normally. It wasn't until I looked in dmesg that I saw:
That surprisingly confusing (and confused) set of messages suggest that the device is like Schrödinger's cat: it's up and down at the same time. Presumably the German-flavoured “Device has been disabled!” is the key: it's not up at all. But there's no way to see that with the normal tools.
I've seen this problem before, but last time I installed Ubuntu (from this same DVD) it worked. So maybe the issue was that the firmware wasn't installed because the card wasn't present at installation time. OK, not a worry, I can install again. That might be a good idea anyway: I found that after suspending and resuming, the machine worked fine, but the display was blank.
But it wasn't that simple. The laptop didn't recognize the DVD. Why not? A bit of comparison on other systems showed that there was no problem with the DVD, and the drive was able to read other CDs. It's a DVD drive, but I didn't have a DVD handy, and I've come to the conclusion that it's a compatibility problem: this is an old DVD drive (the machine must be the best part of 10 years old), and my DVD was a DVD+RW, which is relatively recent. So it looks like I'd have to do the installation on another system. But that would nullify the intention of installing on the current hardware.
Instead, installed FreeBSD 8.0 from CD, something I had tried before and run into trouble with missing X. And that in turn had been a problem because 8.0 is not compatible with the most recent versions of the Ports Collection. So this time installed the Ports Collection from the installation CD. Rebooted the machine. It panicked trying to initialized the 802.11 network card. Without it, I was able to a partial build of X, stopped at 12:00 when we switched to high network tariff.
The whole thing is taking far longer than I had hoped. Maybe I should go and buy a netbook. I now have no less then 7 laptops, and none of them will work properly in the kitchen:
The old Inspiron 7500 is only barely functional. Keyboard defective, floppy (!) and CD-ROM defective, no battery. Basically it's scrap.
I also have an old prototype Dell Latitude. From memory it has about 32 MB of memory, and it can't run X. No battery. Clearly a non-starter.
The ThinkPad 380Z is also too low-powered for the job. 96 MB memory and a 200-odd MHz Pentium II just aren't enough for modern web browsers. The battery, though functional, is also a little flaky.
The Inspiron 5100 is in better shape, but there seems to be something wrong with the battery charging circuitry, and it can't charge the new battery I bought for the Inspiron 1150 (same kind of battery). It also runs Microsoft, and moving that to, say, the 1150 is more pain than I want.
The Inspiron 1150 has a broken display hinge.
I might be able to install on the Inspiron 8100, but I still need to find a way of reading the install disk. This is probably at the moment the best option.
I also have an old G3 PowerBook, but the battery is dead, and so is the charger.
There are solutions to these problems, but they're all a pain. Currently the Inspiron 8100 looks like being the best option, assuming I can get it to resume correctly.
| Topic: general, photography | Link here |
Gary Jones of Jens Gaunt along today to look at the house with a view to selling it. He was clearly not very impressed until he got inside, after which he became quite enthusiastic. He wasn't able to help much with advice about subdividing the property, but it seems that we've got our work cut out.
While he was there, he saw my photo equipment and promptly came out with questions about photography: he has a Sony α330 or α380 (not quite sure which), which has switches for AF/MF both on the lenses and on the camera body, and he didn't understand. Neither do I, but I was at least able to tell him what they meant.
| Topic: gardening | Link here |
The March flies are still a pain. Did a bit of work in the garden, mainly potting things. Contrary to my fears, the Zephyranthes candida is developing lots of new buds:
|
|||||||||||||||||||||||
Put it in a bigger pot, which proved to be very necessary: it was quite root-bound. It'll have to go in the ground in the spring.
| Thursday, 18 February 2010 | Dereel | Images for 18 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Why don't users read documentation?
|
Topic: technology, opinion | Link here |
One of my continual complaints about all kinds of software, whether free or fee, is that the documentation is inadequate. It occurred to me that my moaning the other day about two apparently different issues really had a common cause: documentation.
In the first case, I was complaining about the lack of documentation for the window manager under Ubuntu Linux (presumably it's part of the GNOME project). How do I lower a window? How do I iconify it without looking for the buttons on the title bar (which may not be visible)? Maybe it's significant that the GNOME home page doesn't contain a link to documentation; you have to go to the support page for that. But possibly I can find the answer I'm looking for there.
We had a long discussion on IRC about the subject of software documentation in general, during which the following opinions were expressed:
We don't need documentation, because end users don't read it. Only IT people read documentation.
That's quite an assertion, of course. You can turn it around: end users don't read documentation because there is none. But there's more to it than that; read on.
Some things are documented. Apple's “Keyboard Shortcuts” are “documented” in the System Preferences Menu: Select “Keyboard & Mouse”, and you get a whole collection of keyboard bindings, conveniently encapsulated in a too-small, non-resizeable window:
|
|
||||||||
Some people claim that this is documentation. I disagree. For example, I used the “Save picture of screen as a file” to save this image. Maybe I could have used “Save picture of selected area”. But how do I select it? And where does the file go? I ended up having to look for it, because it's not documented (it was in the directory ~/Desktop, with the file names Picture and 1, separated by a space). If this is documentation, no wonder people don't want to use it.
You can write documentation from at least two perspectives: the perspective of the designer or implementer, and the perspective of the user. Most documentation is written by the implementer or somebody close to him, and it gives rise to monstrosities that I can't understand. What chance does a naive user have?
This last point is very important. Good examples of designer-centric documentation are gnuplot and GIMP. The documentation is probably complete, but it's voluminous, and finding what you want is not easy.
Sure, you might think, that's where books come in. But why? People who write software—especially those who want money for it—are implicitly offering a user-level solution. Why do we get designer-level documentation. No wonder people don't read that kind of documentation; it's akin to following menus to see what they do.
To illustrate the difference, look at the photo of Piccola I was working on a couple of days ago. My question is: “How do I replace the door and frame behind the cat with a uniform area like the wall to the left of her?”. The answers I get from the documentation are:
3. Brush Tools
To be fair to the GIMP project, there's more documentation than that, including some more like what I'm looking for. There are partial guides, such as “Opening files”, or even “Removing unwanted objects from an image”. But that link points me back to the Clone tool, which doesn't seem to be the correct tool to use, and it conflicts with the documentation in the c't article. Still, it's a start.
That c't article is a third kind of documentation: a cookbook. “Click here, move the cursor there, lather, rinse, repeat”. That's not the user perspective either: to use a tool effectively, you need to understand what is happening. It would be really nice to have documentation that looks at the software from the point of view of the user. Then users might read it.
The latest issue of c't arrived today. They've redone the address labels. Agreed, the old ones were not very accurate, but they've introduced two new errors in the new one (on the right):
|
|
||||||||
|
|
||||||||
They've obviously done some thinking about how to write the post code, which is now correct. But they've moved the street number to the position it has in German addresses, and they've changed the case of the title from accusative to nominative (“Herrn” to “Herr”). That requires a bit of explanation to people who don't speak German: there's an implicit “to” in front of the name (“an Herrn Greg Lehey”). On the other hand, as the extract from the following (correctly addressed) letter shows, you address people in the nominative:
|
|
||||||||
So how did this happen? The address is almost certainly a column in a database, and in this case almost certainly with content Herr, and a program that uses this information needs to convert the titles to the accusative in the mailing labels. It looks like they forgot.
| Topic: general, food and drink | Link here |
Into town for a haircut today. Yvonne took the opportunity to have her car serviced, so drove around with her a bit, first to Mega Meats in Latrobe St, where they have a surprising amount of meat at good prices—and helpful service. I think we'll have to investigate more of their products. Returned with much more than our usual amount of meat.
|
New plants
|
Topic: gardening | Link here |
Also to Formosa Gardens, primarily to buy some fertiliser, but returned with another couple of succulents: a Crassula falcata, also called “Propeller Plant”, and a Euphorbia “Diamond Frost”—where do they get these names from?
|
New house?
|
Topic: general | Link here |
Another discussion about subdivision today. We haven't given up, but it looks like it'll be an uphill battle. Over to Chris' place to take another look at the block of land that she had for sale a while back:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's 8 ha, more than we need, but it also has a number of trees on it. But buying it, even at a favourable price, would probably put us over our budget. More head-scratching to do.
| Topic: photography, opinion | Link here |
I took the photos of the plot of land with a polarizing filter. I'm gradually coming to the conclusion that it's not a good idea for panoramas, and maybe not for cloudless skies at all. Yes, it makes the sky even bluer, but it's a steel-blue, and it's blotchy, as is very clear in the first photo.
Call from a bloke today with a problem with his Canon 9900P scanner. He ran into exactly the same problems as I had at the time, but for him it was the first time. I wish I had been able to help him.
| Friday, 19 February 2010 | Dereel | Images for 19 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
cvr2 megacrash
|
Topic: technology, multimedia | Link here |
I've recorded some TV programmes for Chris Yeardley. How do I get them to her? They're 1080i, a total of about 16 GB, so DVD isn't really an option. Chris came up with the idea of giving me a USB-connected external disk drive to copy to.
That's Microsoft-formatted, of course, and USB is still flaky on FreeBSD, so I didn't even try to use that. Instead plugged it into boskoop, the Apple, which immediately recognized it:
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 7 -> ls -al
I won't comment on the file names. What happened next surprised me:
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 8 -> mkdir Videos
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 10 -> ls -ld .
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 11 -> touch foo
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 12 -> cat autorun.inf
So: we have a valid directory, valid entries, but when I try to create anything on the disk I get ridiculous and incorrect error messages. On a hunch, tried:
=== grog@boskoop (/dev/ttyp5) /Volumes/Maxtor1T 13 -> mount -v
So it's mounted read-only. But in that case it should tell me. It does in FreeBSD. Why not in Mac OS? Went to the Desktop and looked through a maze of twisty little menus, all different, but couldn't find anything to help, so finally tried the standard methods to mount it read-write. This works in FreeBSD, and is documented identically in mount(8) on the Apple:
=== root@boskoop (/dev/ttyp6) /Users/grog 2 -> mount -u /Volumes/Maxtor1T
Clearly yet another very untidy part of the Mac OS X implementation. Is it maybe not able to write to NTFS file systems? Gave up and put it on cvr2, the Linux box, which understood it and happily wrote to it—for a while. Then it hung solid:
|
|||||||||||||||||||||||
Brought cvr2 back up without any serious problems, but decided not to try my luck. Instead plugged the drive into pain, my Microsoft XP laptop. And it didn't find the drive!
Spent some time looking at what was wrong there. Went into various menus to identify drives, and they all saw it, but it didn't show up on the “My Computer” screen—only the main disk, the DVD drive and 4 Samba shares.
Yes, I know it's not Samba, but that's the way I think of them.
Looking through the Samba shares, I discovered that one of them (drive E:) wasn't Samba at all—it was the USB disk drive. It seems that:
Started copying that from dereel—Samba share to USB drive. It took an eternity, not helped by the fact that even during copying the directory listing showed the file at full size and with the original modification timestamp, so there was no way of measuring progress, or indeed if data was still being transferred. Chris suggested a 16 GB USB stick instead, so after copying one file, took the disk back and confirmed that it would at least work on the Microsoft box. Well, partially. MythTV creates file names like 2050_20100218225700.mpg, which encode the date and channel, but which are a real pain to understand, so I change the name into something relating to the title and subtitle, for example To-kill-a-Mockingbird. On cvr2 I also turn the transport streams into program streams, and rename the original with a name ending in .orig, for example 2050_20100218225700.mpg.orig. For Chris I left the obfuscated names, which was just as well: Microsoft “understood” the first file name, but had no idea what to do with a file of “type” orig.
Back home and plugged the USB stick into cvr2, and was amazed how fast it copied—for a while. Once again a hang, but this time no display, and rebooting didn't work. Another case of the dreaded file system corruption I last saw only 4 months ago. Tried the tricks I used then, but they didn't work either. xfs_repair returned an “out of memory” condition:
=== root@naan (/dev/pts/0) ~ 6 -> xfs_repair -L /dev/sdb1
Repeated attempts alternated with hangs; clearly I wasn't getting anywhere, so tried mounting the disk read-only. There didn't seem to be much left:
=== root@naan (/dev/pts/1) /mnt 20 -> df /mnt
Further investigation showed massive file system damage, but many of the files were still there:
=== root@naan (/dev/pts/2) /mnt/var/lib/mythtv/recordings 8 -> du -s
That's 23 GB in that directory alone, though many files were broken. Clearly the information from df is inaccurate. But also as clearly, the disk is in a real mess. In view of the repeated problems, decided to rebuild the system from scratch, with a new disk and a new motherboard. I had kept a disk image for this purpose last October, so all I needed to do was to copy it to a new disk. And I even had luck with my MySQL backups: they get done at random times, perhaps because of problems with NFS (the backup directory is NFS mounted on dereel, and interoperability between Linux and FreeBSD NFS seems to be getting continually worse). But the last backups were:
=== root@dereel (/dev/ttypa) /dump/cvr2 169 -> l var/backups/
The system crashed at 15:45, only 90 minutes after the last backup, and I don't think I did any database updates in that time. My guess is that the backup was done after the first reboot, while the /dump file system was still accessible. So I can be reasonably confident of a full recovery. The backup root file system is only 1.7 GB, but it represents a 200 GB disk, and just copying that much can take several hours, so left it run overnight.
| Topic: gardening | Link here |
Our Zephyranthes candida had only developed new buds yesterday, but today they're in full bloom:
|
|||||||||||||||||||||||
That's the same bud I took yesterday.
In the evening, finally found the culprit that has been eating the roses: a wallaby, which I chased away before taking photos. It could be the same one that Carola Schlanhof investigated almost exactly a year ago, and it was in the same place:
|
|||||||||||||||||||||||
It disappeared pretty quickly when I chased it; maybe it'll stay away. Now to catch the Wentelteefjes.
| Saturday, 20 February 2010 | Dereel | Images for 20 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Resurrecting cvr2
|
Topic: technology, multimedia | Link here |
Into the office this morning to find that my disk copy of the cvr2 file system still hadn't completed: the NFS connection had hung again. This is getting beyond a joke—why can't Linux and FreeBSD communicate via NFS? Mac OS X has had no trouble with FreeBSD. I suppose I should see whether Mac OS X and Linux work.
That meant, of course, that I was no further than yesterday afternoon. Copied the disk image to the local disk using scp, and rebuilt the disk from the local copy. It's interesting to note that for large parts of the copy time, the limiting factor wasn't disk (or, in the previous case, net) bandwidth but the CPU power needed to uncompress the image.
Finally got it uncompressed, and tried to mount the disk. Failure. Rebooted. Success. Booted from the disk, which worked, but there were clearly a few strangenesses: from time to time I got a console message, which for some reason didn't get logged, indicating that a specific inode included null block pointers. Decided to put the machine back together and address that problem later, but that didn't work: second time round it ran into more problems and ended up with a “kernel panic”.
Disk out again, back into the test box, and ran a xfs_repair on it. That required a number of iterations, with a mount to replay the log in the middle:
=== root@naan (/dev/pts/0) ~ 3 -> xfs_repair /dev/sdb1
=== root@naan (/dev/pts/0) ~ 4 -> mount /dev/sdb1 /mnt
=== root@naan (/dev/pts/0) ~ 5 -> umount /mnt
=== root@naan (/dev/pts/0) ~ 6 -> xfs_repair /dev/sdb1
It's difficult to recall 9 digit inode numbers, but the 807116596 somehow looked familiar. In any case, after that operation things ran smoothly, and all I needed to do was to restore the database, which ran smoothly. The only issue was that I no longer had the recordings on the computer, but the database knew about them, and it took a bit of convincing to get it to forget them. But by 11:00 we were up and running, with only one recording lost.
| Topic: general | Link here |
Decided to go off to Ballarat again and take another look at the house we had been looking at last week.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, it would probably be OK, with a bit of modification, but they're surprisingly inflexible. Still, well worth considering; you get even less choice when buying a used house. But then Bill came out with all sorts of extra costs that he hadn't mentioned last week; apart from obvious things like lights and floor coverings, there are things like fire rating issues (which in one extreme case added $80,000 to the cost of the house) and possible extras for the slab (foundations) which could run to another $30,000 if we were unlucky. That's two thirds of the base price, though it's unlikely that we would come even close to either of those sums, but quite possibly these extras will make the thing a non-starter. Took back lots more documentation; we'll have to do some careful calculations.
| Topic: food and drink | Link here |
Baking bread today. Lately I've been having trouble with the bread tearing at the edges, like at the bottom side of the top of the bread in this image:
|
|||||||||||||||||||||||
|
|
|||||||||
What causes that? It's common enough—I've even seen it in commercial bread in Germany—but I'd like to get rid of it. Dough too dry? Oven too hot at the beginning? “Too much yeast” doesn't really apply for sourdough, though it could be something like that.
| Sunday, 21 February 2010 | Dereel | Images for 21 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cloning MythTV
|
Topic: technology, multimedia | Link here |
Building on yesterday's success of my recovery of cvr2 from a system image, decided to make another clone for Chris, whose TV has died. In the process, discovered that I had restored from the wrong image: I had one dated 26 July 2009, and another dated 20 October 2009, and by accident I had restored from the former. It worked, though, so did the same for Chris. Her disks are SATA (I'm still using mainly PATA), so I had to change to the new machine I built a few weeks back. That made a big difference in the copying speed. Once again, of course, I had the XFS corruption, so this time ran xfs_repair on it immediately, and noted that I didn't have to replay the log (why not? I thought the dump was made of a live file system).
Booted from the system, and noted that too many things are hard-coded with system name and primary interface address: the new machine happily told me that it had three tuners (it had none), because it was accessing the database on cvr2. And since cvr2 is stored all over the database, it needs to call itself cvr2 as well. Decided to take it over and install at Chris' place.
First, though, I needed to shut down. As I had seen with cvr2, it hung at this point:
Ctrl-Alt-DEL created a message that it was shutting down, but it hung again, and in the end I had to power it down manually. Rebooted and confirmed that it was then able to shut down; the next message was:
Yet another indication, it seems, of flakiness in the Linux NFS implementation.
Took the disk over to Chris' place and put it in the machine, then turned to delicious, her Apple, to update the DNS configuration. That worked, but the changes had no effect: we still had the old configuration. It appears to be another left hand, right hand problem with disagreements between the Apple kernel and the GUI. Went to the GUI, which wouldn't let me change the configuration. I just got a popup screen:
|
|
||||||||
Clicking on OK just brought a repeat of the same message. No way to understand what the problem was, how to fix it, or even exit the menu. Somehow I find these Apples just too frustrating. But the new machine (cvr2, because of the database) runs Linux, and the dichotomy between GUI and kernel isn't that bad there, so decided to install DNS on that. That proved to be more fun: on my cvr2 I had put the /var/cache/apt hierarchy on dereel and accessed it via NFS, so it wasn't present at all on Chris' machine. Back home with a USB stick and brought it back, and then ran into an out-of-date apt data, but finally got named (sorry, bind9) installed. It didn't help much: unlike most Debian-style packages, it just installed the files and left them there. Using the GNOME desktop and a keyboard with the Ctrl and Alt keys in the wrong place didn't help either—it brings home to me how much more difficult it is to work in a strange environment.
Finally finished the configuration and started named—no response. No logging. Even when I started it with debugging, it didn't produce even normal log messages like which zones it tried to load. At the very least the config file could contain some hints about which logging options to use to which effect.
Finally gave up and back home to read the manual. On the way it occurred to me that we already have a name server running properly, on ns1.narrawin.com, so an obvious choice would be to just add the zone dereel.narrawin.com. Did that, and immediately found what I had done wrong:
I had forgotten to update the $ORIGIN line. Fixed that, and everything ran well. But what a pain! People tell me that I should be flexible and be able to use strange environments and keyboards—but why? I've spent literally decades refining my environment to where I can do things easily. Why should I give that up?
And what logging options? I've found that these (courtesy of Edwin Groothuis) work for me:
| Topic: food and drink | Link here |
A number of comments on my bread problems today, including an online article (discussing the issue at the bottom), all much along the lines of what I had suspected: it could be that the dough was too moist, that it had risen too much, that the oven was too hot, that the surface of the bread was too dry (this one certainly wasn't the case). And some suggested “slashing” (cutting grooves into the surface) to make up for it.
The issue is, this doesn't happen every time. I suspect it's due to my recent ideas of glazing the surface by having the grill on at the beginning of the bake. I'll try again without next time.
Chris along for dinner tonight, some of the meat we bought at Mega Meats the other day. Not a revelation: the spare ribs were tough (and I hadn't noticed at the time that they were also pretty expensive), and the lamb skewers were really gristly. We'll need to be pickier in the future.
| Monday, 22 February 2010 | Dereel | Images for 22 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
64 years of continuous revolution
|
Topic: technology, opinion | Link here |
Last week was an anniversary of sorts:
Arguably (very arguably), ENIAC was the start of the computer revolution. 64 years ago, a nice number for subdividing. Consider what has happened in that time, in chunks of 8 years. Here's a first draft of what may become a real article reflecting my views:
1946 to 1954: Some of the basic hardware concepts appear: stored program, binary arithmetic, magnetic core memory, index registers. The IBM 704, introduced in 1954, had almost a modern architecture and instruction set.
1955 to 1962: Programming languages come onto the scene: LISP, FORTRAN, ALGOL and even the beginnings of COBOL. Hardware gets faster, but there is still a marked difference between “commercial” and “scientific” computers. Hardware logic implementations shift from thermionic valves (vacuum tubes) to transistors (“Second Generation”).
1963 to 1970: Hardware logic implementations shift from transistors to integrated circuits (“Third Generation”). Operating systems appear, including UNIX. Languages start to become platform-independent. Disk storage comes into general use, supplanting tape-only systems. IBM introduces the System/360 to unify “commercial” and “scientific” computers. Minicomputers appear, typified by the DEC PDP-8. Cache memory and virtual memory start to appear. The first computer networks appear.
1971 to 1978: Microprocessors appear (“Large Scale Integration”), and with them the first “home computers”. Floppy disks appear, and the last drums die. Magnetic core store starts to be replaced by semiconductor memory. A number of hardware features previously reserved for larger computers, such as cache and virtual memory, become more prevalent. Operating systems start to be written in higher-level programming languages, and are typically multi-user. Video display terminals appear, and access to computers shifts from batch (punch card and printer output) to interactive terminals.
1979 to 1986: The US DARPA awards a contract for the development of the next generation of ARPANET, the Internet, to BBN and the CSRG at UCB. The implementation platform is the BSD version of UNIX. UNIX also becomes the operating system of choice for a number of new startups, notably Sun Microsystems.
IBM introduces the 5150, better known as the “Personal Computer”. Industry emphasis shifts from big computers (which are now called “mainframes”) to smaller computers, typically intended for use by only one person. More and more people from outside the computer industry use computers, and the software industry changes its emphasis to cater for them (“user friendly”). The first graphical user interfaces make it to the market (Apple Macintosh, Digital Research's GEM and Microsoft's “Windows”). Richard Stallman founds the GNU Project and the Free Software Foundation. The punch card dies out.
1987 to 1994: Networking becomes more prevalent, and a number of third party companies appear to provide networking products for desktop computers such as Microsoft DOS. The World-Wide Web appears. Desktop computers all have hard disk drives. CD-ROMs start to appear as a data storage medium, gradually making inroads into tape and floppy disks, particularly as a software distribution medium. New processors for desktop machines appear with features such as cache and virtual memory. The first practical implementations of UNIX for desktop computers appear, notably for the IBM PC and clones, and include free versions derived from BSD UNIX, and also new systems such as Linux.
1995 to 2002: Personal computers become ubiquitous. A strong divide develops between “user friendly” end user software and conventional software designed for people who understand computers. Vendors such as Microsoft discover the Internet and offer support for TCP/IP networking in their operating systems. Web browsers become one of the standard software tools. The dot com boom and bust symbolizes the transition of emphasis in the computer industry from technology to big business.
2003 to 2010: The floppy disk dies out, replaced by USB sticks, CD-ROMs and DVD-ROMs. Otherwise it seems that not much has changed. Things continue to get faster and cheaper, the divide between “user friendly” and “powerful” continues to increase, processors have got faster and cheaper, disk and memory sizes have increased dramatically—but what has really changed? Commercial vendors and the GNOME and KDE projects have changed their user interfaces a couple of times, but somehow I don't see a big difference in what I do or how I do it.
More work on the MythTV box for Chris today, and finally got it working, not without more surprises: the satellite modem has a DHCP server, which conveniently changed the address of the line printer to conflict with cvr2, and cvr2 didn't recognize the file system on the disk I brought over with yesterday's recordings (all 25 GB of them): it was ext4, and the kernel doesn't support it. In the end, booted from the other disk and copied it like that. And yes, I did bring my own keyboard. That makes things a little easier, but only a little.
| Topic: general | Link here |
Called up Andrew Harman , the surveyor whom Bram Gunn had recommended, and talked to him about the potential subdivision of the property. As suspected, it won't be easy, and we'll probably need to apply for reclassification of the zone. The application alone would cost about $6000, the process would take about 18 months, and if successful we'd need another $6000 for the subdivision itself. Not a very pleasant prospect. And it looks as if Chris' plot will be too expensive, so we need to prepare ourselves for the possibility of scrapping the whole project.
| Tuesday, 23 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Network problems: after 3 years, finally action
|
Topic: technology | Link here |
My network problems continue, but at least it seems as if they've finally decided to do something about it. Today I received a mail message from my ISP stating, amongst other things (and reformatting from the original one-line-per-paragraph):
This is an informational email in reference to the current problems you are experiencing with your IPSTAR Satellite Service. IPSTAR is the Satellite provider who controls the management of your service and has advised us as well as other Service Providers that there is a possible fault with their ground stations which is causing services to randomly disconnect thus forcing a \221log off\222 on an active connection.
That's one way of putting it. This “possible fault” has existed since at least the first time I tried to transfer files to Chris Yeardley nearly 3 years ago. During the time I have had my IPStar service, it has been responsible for the vast majority of the 819 outages I have had, and I've reported the problem repeatedly. But the problem is on the increase: 129 of those outages have been in this month alone. They continue:
IPSTAR have now made a commitment to send two senior engineers from Thailand to investigate this fault, we hope they are able to diagnose this fault swiftly and find a resolution. Due to the nature and complexity of this fault IPSTAR had given us an unexpected time frame of mid January to have this problem fixed. As we expected we are yet to hear back from IPSTAR in relation to this problem and are still awaiting the findings of the their two senior engineers. We are continuing our endeavors of pushing IPSTAR for a resolution or an answer on what is causing these small dropouts but are yet to receive any sort of update.
As this fault is outside of our control we are solely reliant on IPSTAR in finding a resolution. Because of this we will suspend your current monthly plan charge as of this email until we/IPSTAR deem this platform fault to be fixed. Upon this fault being fixed by IPSTAR your monthly charge will revert back to your normal monthly plan charge. For validation of this credit please refer to your next monthly invoice.
That's nice of them, anyway. But I wish they would finally fix the problem. And hopefully, when they deem the problem to be fixed, I will be able to agree with them.
Other people have been complaining about this too, and there's a thread on Whirlpool that handles it too.
While searching around for something else, found the text of some additions I had made to adventure decades ago, giving a partial entry into Tandem's Guardian operating system. A sample:
You are in the disc process. The activity here defies description: a young, energetic and obviously slightly stupid dwarf runs around taking single pages out of files in a filing cabinet and putting them onto a pile marked 'CHIT', while taking others off the pile again and putting them into other files. From time to time he disappears through an invisible exit, whereupon a noise like a circular saw can be heard. Shortly later he returns carrying a slice of some indeterminate brown material, which he puts on his pile of CHIT. All the time he complains, invoking magic incantations, and swearing “I'll kill that gnome who lumbered me with TMF“.
Set to to modify the text files, and discovered that FreeBSD removed adventure from the games collection years ago, and didn't even (apparently) replace it with something in the Ports Collection. But both NetBSD and OpenBSD still have it, and was able to compile the latter with only minimal changes to the Makefile. Now to work out how and under what circumstances people got into the rooms.
| Topic: general | Link here |
CJ along to chop down the broken branch today. We had intended to prop up another one, but he was concerned that it wouldn't withstand the wind. We'll have to use two props, and CJ went back home to consider.
| Topic: technology | Link here |
CJ also had a special offer card from http://www.supercheapauto.com.au/. Instructions: “Go to http://www.supercheapauto.com.au/ and follow the links”. CJ's not a computer man, so I went through it with him. Completely broken web site, of course—drop-down menus under images, bad formats, and that under all browsers I tried, including Microsoft “Internet Explorer”. Absolutely no appropriate link to follow. Called up their help line and the representative told me that he didn't know either, and forwarded me to somebody's voice mail. What a pain!
| Topic: food and drink | Link here |
We had intended to eat Pizza today, and I wanted to make the dough with sourdough. That didn't work as planned: my first step in making bread is to mix 150 g flour with 50 g of starter, which works fine. For the pizza dough, decided to stretch it to 250 g of flour. I had expected it to take longer (normally it takes about 5 hours), but by evening there seemed to have been no activity at all. Is this due to the dilution or the olive oil?
| Wednesday, 24 February 2010 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: gardening | Link here |
Gradually the March flies are becoming less of a nuisance (they really don't understand calendars), and now that the broken branch is finally gone, got round to mowing the lawn in the area of the cathedral. Now that it looks as if we're staying here, we're paying more attention to the garden, which definitely needs it.
| Topic: general | Link here |
CJ along with an old gate that Yvonne had manhandled as a harrow, and which he had repaired, but we decided that it wouldn't harm to weld some star droppers on to it, so he took it away again.
| Topic: food and drink | Link here |
My dough still hasn't risen! Why not? I have absolutely no problems when baking bread, and there's no reason to believe there was anything wrong with the starter. Left it all day, and it looked a little larger (wishful thinking?), but still no bubbles.
| Thursday, 25 February 2010 | Dereel | Images for 25 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Moving large files to Microsoft
|
Topic: technology | Link here |
I've been recording video for Chris for a few days now, and I'm still having trouble getting it to her. I moved a fair amount on a Linux disk, but now I'm back to using the USB stick with a FAT-32 file system on it, the one that was involved in the crash of cvr2 last week. Put it in naan, a test box, to be on the safe side, but I didn't have any problems of the nature, which tends to confirm my conclusion that last week's problems were hardware-related. But that didn't mean smooth sailing: the copy failed with the error file too large. The maximum size of a FAT-32 file is 4 GB (well, 1 byte less), and most HD MPEG-2 recordings are larger than that, as was this one.
What to do? The obvious option is split, but Andy Snow pointed me at rar, which has an option to create a self-extracting executable:
=== root@teevee (/dev/ttyp4) /spool/Images 27 -> rar -m0 a -r -sfx -v1000000k violence.exe MVP-2010-02-2*
The -m0 option is really necessary: otherwise it tries to compress, which is both slow and useless with large MPEG files.
Mail from Michiel Overtoom telling me that Zork is available in the FreeBSD Ports Collection: /usr/ports/games/dungeon. But that's not adventure. Peter Jeremy tells me that that's there, too, though, in /usr/ports/games/bsdgames.
|
The Nottles move in
|
Topic: general | Link here |
Lee Nottle and her husband have now moved in across the road, and this morning she came in to say hello. Looks like things will become more active round here.
CJ in with Sue and the gate (or harrow) and tried it out. It proves to be better to leave the droppers on top, though I'm concerned that the mesh might not be up to the task of being dragged around the paddock. We'll see. They then went to cut up some fallen branches, which Sue will use for firewood.
| Topic: food and drink | Link here |
My pizza dough has now finally risen, creating large bubbles, so added some more flour and left it to rise again. Finally made a couple of pizze out of it, and baked.
Not a success. It didn't rise enough, but it did stick to the pan, so just getting it off was a problem:
|
|||||||||||||||||||||||
To add insult to injury, we were using a new kind of anchovy, which are very heavily salted, and the pizza was barely palatable.
| Friday, 26 February 2010 | Dereel → Bannockburn → Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
| Topic: general | Link here |
Off to Bannockburn today to investigate the possibility of subdividing the property. It's the first time I've ever been in a local council in Victoria, and I hope it'll be the last.
To start with, there's no place to talk. There are a couple of stand-up tables in the hallway, and that's where you have to deal with the staff. My issue wasn't confidential, but the whole setup appears to be intended to make things as uncomfortable as possible.
Then, of course, the news that we're not allowed to subdivide. The planner, a surprisingly young woman, explained to me that this was a Rural Living Zone (the original thoughtfully posted on the web as a Microsoft “Word” document), and that the minimum size of the land for this zone was 8 ha—nothing I didn't already know. So I asked her: why am I the only person in this zone with a property that conforms to the zoning? Is this not an indication that the area is incorrectly zoned? She mumbled something about crown allotments, and actually did find one other property of 13 odd ha, and a second that she claimed was 8 ha—but it was split on two sides of a road. On the other hand, there must be at least 100 properties with sizes as little as 2 ha, and there are provisions in the Act for subdivision of Rural Living Zone properties down to this size. But she was completely inflexible: nothing to be done. I said that I wasn't satisfied, and asked what redress was open to, and she suggested that I contact a lawyer! Further investigation showed that yes, indeed, I could appeal to the council, who will be meeting in Linton at 16:00 on 11 March 2010, but she said that she would oppose any subdivision.
What's wrong with this picture? At the very least, I feel I am owed an explanation why this zoning exists. Instead, I'm just told something about Crown Allotments as the excuse why over 95% of the lots in the zone don't conform to the zoning, and the rest must do so under all circumstances. This heavy-handedness doesn't seem to match a democracy, and it's worse than I've experienced elsewhere, even in Germany, where the bureaucrats used to be a law unto themselves. But at least, even there, I could ask for and be given details of how to protest against the decision.
This wasn't like South Australia either. I think the difference here is rooted in the history of Victoria, previously a penal colony. In those days, the authorities were above the common people, and didn't have to answer to them. Somehow I feel that times haven't changed that much.
While in Bannockburn, looked in at the couple of estate agents they have there. Certainly any idea of moving to that area is dead: we're still hoping to get a building block for under $100,000, but the prices there are up to 3 times that, and even 600 m² blocks in Bannockburn go for $120,000. But maybe there are alternatives in our area.
| Saturday, 27 February 2010 | Dereel | Images for 27 February 2010 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
TV programme guides: nothing but pain?
|
Topic: technology, multimedia | Link here |
Going through the TV programmes this morning, found that there was very little on TV. Maybe the Winter Olympic Games? Did some investigation, and found:
|
||||||||||||||||||||||
That's really only visible at original size. But the important thing is that for 4 channels there was no programme at all, neither for today nor for any date in the future. The guide is pretty flaky at best, but this was worse than I have ever seen. Stared writing a message to the Shepherd mailing list when it occurred to me that the channels that had no data were the new ones that have sprung up in the last few months, and that it's been a week—the length of time for which the programme is loaded—since I rebuilt cvr2 with old data. So clearly the shepherd configuration was out of date. It's a pain to configure, but the latest configuration was still intact on the old disk, so copied it across and ran for hours pulling in the new data. That worked, but there are still lots of programmes missing.
Back to play around with MythWeb, and added code to /usr/share/mythtv/mythweb/modules/tv/tmpl/default (diffs here) to detect most missing data:
|
||||||||||||||||||||||
That's not acceptable, of course. What do I do now? I assume that the problem is because I'm in the Ballarat area. Most programmes are the same as in Melbourne; how can I “top up” the listings with the corresponding ones from Melbourne?
| Topic: brewing | Link here |
I'm gradually running out of excuses not to brew, and started crushing malt today. It takes forever, and I'm seriously concerned that this Kenwood mixer isn't up to it, something confirmed by a sudden extreme slowdown—almost stopping—after some time. I turned it off, waited about a minute and turned it on again, and it continued normally, so maybe it was just overheating. But it doesn't give me the warm fuzzies that it will survive many brews. Time to look, yet again, for a viable alternative.
The crushing took so long that I didn't get finished. Yet another excuse to put off brewing. Maybe I should go back to extract for a while.
| Topic: photography | Link here |
Photo day today, and did some more playing around, this time with fusing HDR images with enfuse. I've been doing this for a couple of weeks with the photos of the lagoon (which I didn't do this week because it was raining), but there are more extreme cases, so did some photos indoors, and so took a couple of sequences:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Somehow they come out too dark, but that's possibly the only way it can be done. The most exposed photo of the first sequence was only just enough for the inside, so for the second sequence included one that was significantly overexposed:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In any case, the results aren't too bad, and I have another thing to play around with. There are even some links to good documentation on Tom Mertens' home page.
| Sunday, 28 February 2010 | Dereel | Images for 28 February 2010 |
| Top of page | ||
| previous day |
|
IT teaching at Ballarat
|
Topic: technology, general, food and drink | Link here |
Chris over for dinner—roast beef. Once again cooking times and meat temperatures were a problem, and it took 15 minutes longer than intended. I had aimed for 58°, but in the end settled for 55°; that's still a little on the warm (and slow) side. Next time I should go for 53°.
Chris is starting an Bachelor of IT (Professional Practice) course at the University of Ballarat tomorrow. Discussed what subjects she is studying; it's an interesting combination: Communications (which seems to be interpersonal communications), Networking, Multimedia and Programming. The programming language she'll be learning is Python, which seems a strange choice. I'll watch with interest.
| Topic: brewing | Link here |
More grain crushing today, which I completed without further incident. Interestingly, I found some Indianmeal moth grubs in the malt. They weren't there yesterday, so I assume it's at least a partial coincidence: they had reached that stage in their lives where they leave the grain to pupate. In the course of the afternoon, killed about 30 of them in two containers of grain. That's pretty much what happened 16 days ago, and in the meantime I have had very little. Anyway, it looks as if vigilance helps.
| Topic: photography | Link here |
Responded to a thread in the oly-e.de forum today. The originator was looking for guidance on the purchase of a flash unit, and one choice was the Mecablitz 58 AF-1 O digital that I have. Started writing a reply, checking as I went.
One of the things he wanted to use the unit for was for high-speed photography of skaters. The units in question have a so-called TTL-HSS mode, where the unit can fire at speeds where the shutter only exposes part of the sensor at any one time. It does this by flashing multiple times during the time the shutter traverses the sensor. Clearly any single flash will only cover a part of the sensor that is uncovered at the time the flash fires, but the total flash intensity is no higher, so the guide number diminishes in proportion to the square root of the ratio of slit width to sensor width. And that's dependent on the shutter speed. On the cameras we're talking about (Olympus E-30 and E-3), the shutter takes 1/250s to cross the sensor, so at the highest speed (1/8000s) the ratio is 32:1. So the guide number (81 at 200/24° ISO) should shrink by √32, or about 5.6, so it would be about 14.
In fact, the indicator on the flash unit showed that the maximum distance I could use at 1/8000s was 1.2 metres. The maximum aperture was f/4, so that corresponded to a guide number of only 4.8. Further checking showed that at 1/250s, the maximum distances was 6.9 metres, a guide number of only 28. This only happened in TTL-HSS mode. In TTL mode it showed a maximum distance of 20 metres, in accordance with the calculated guide number of 81.
So: the guide number of TTL-HSS is only about 34% of the guide number of normal TTL (or, presumably, any other single flash mode). That corresponds to an 88% drop in illumination power. Why? Is this the price you pay for careful regulation? Do other flash units do the same thing?
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 |
