|
| Greg's diary |
| October 2006 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Sunday, 1 October 2006 | Echunga | |
| Top of page | ||
| next day | ||
| last day |
Quiet day. Finally got round to connecting up the power point for the fridge in the laundry, which took me far longer than it should have done; I really need to get my tools together.
Australia has some of the most draconian road speed limits I've ever seen. Despite the large distances, good roads and low traffic densities, most states have an absolute limit of only 100 km/h. The authorities seem to think that high speed is the cause of all traffic accidents. That's like saying that heart failure is the cause of all deaths: it's involved, but seldom the main cause. The whole attitude is summed up in a sign that the police have been putting on the freeways lately: “Speeding? What's your excuse?”.
Well, the word “speeding” is silly. Does it mean “excessive speed” or simply “exceeding the limits”? Clearly it's the latter, and that's one of the meanings that the Oxford English Dictionary lists. Others are:
1. a. Success, prosperity, good fortune; the fact of being successful or of attaining one's end.3. The action of aiding, furthering, or setting forward.
I don't need an excuse for travelling fast; I have a perfectly valid reason. Why should I spend longer getting somewhere than I need? The only reason is because it's against the law. The penalties are known, so if I choose to “speed”, it's a tradeoff between these two.
Things are different in Germany: there is no general speed limit, those limits that are set are much higher, the traffic is much denser, and people drive like maniacs. If the proponents of speed limits are right, the road death rate should be much higher. Today I found the International Road Traffic Accident Database, which shows statistics for different countries. In Australia, the road traffic death rate is 79 ppm (i.e. 79 deaths per million population). In Germany, it's 71. To use the same stupid attitude and blame speed for all accidents, this would imply that speed limits increase the likelihood of dying in a road accident.
More seriously, though, there's a lot of interesting stuff here. Why are there so many fatalities in the Czech republic and Greece? That can't be due to excessive absolute speed either.
| Monday, 2 October 2006 | Echunga | Images for 2 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Labour day today, a public holiday, which we spent mainly doing garden work. We're heading for a really dry summer.
Yana is moving into a new house today, planning to stay for 3 years while the owner does some course in Canada. Yvonne loaded up the horse trailer with some stuff for her, including an ancient mattress that reminded me of other times, the first flat that I moved to in May 1973. How time flies!
While researching the meaning of the word “speeding”, decided to sign up—again—for the OED Word of The Day. The message hung in the mail queue for some time:
=== root@echunga (/dev/ttyp5) ~ 102 -> mailq
Decided to try it manually and found:
=== root@echunga (/dev/ttyp5) ~ 103 -> telnet marz.stanford.edu smtp
This worked for a number of different systems, so I assume that their spam detecting MTA has become a little too aggressive.
Looking for references to “Speeding. What's your excuse?”, found this set of photos. These are exactly the signs I meant; the second emblem from the bottom right is the state emblem of South Australia. It also indicates the speed limit (90 km/m) on an almost empty, straight and well-maintained road. Other links also show the “reasoning” from the Government:
At 65km/h, the risk of crashing is double the risk at 60km/h.
If they want people to believe that, they're going to need to present some evidence, or at least reasoning.
| Tuesday, 3 October 2006 | Echunga | Images for 3 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Back to work, a lot of it. Added to my usual load is the ICT Council web site. We've had a surprising amount of interest, but so far no really clear idea of what we're looking for.
Also racked my beer from previous Sunday. That's part of the Brew Day brews.
Tom Maynard sent me another message, referring to my comments on speed legislation, and asking if I had just been booked for speeding. No, nothing of the kind. I stick to the speed limits, despite my objections. That's the difference between respecting the law and agreeing with it. I've been unhappy about the situation for some time, but I can't recall why it came to a head right now. Possibly it was a reaction to some stupid news report, like “Police chased a car 40 km down the South-Eastern freeway in the early hours of this morning before finally stopping it near Murray Bridge. It is alleged that the driver reached speeds of over 140 km/h. The driver was charged for dangerous driving.”. The South-Eastern freeway is in excellent condition, there is almost no traffic, and the weather conditions were ideal. In most countries, this speed would be below the speed limit.
| Wednesday, 4 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
This morning discovered that my carbon dioxide cylinder (fire extinguisher) was empty. It should have lasted at least a year, but it's only been going for a little over a month. Presumably the connectors weren't as gas-tight as they should be. I'll have to see if I can just turn it on when I need beer; the low pressure side doesn't seem to have any issues, and first experiments seem encouraging.
Peter Jones over later to talk about databases, but we also spent a considerable amount of time talking about the weather: today the temperature in Adelaide reached 33°, and even here in the hills it was over 30°. There's little rain in sight, and the general opinion is that this could be the year of the Big Bushfires. And with me planning to go to EuroBSDCon next month. I think I'm going to have to cancel; if I stay here and we have a bushfire, I'm reasonably sure that no serious harm would come to the property, but if I'm away, there would be nobody else able to spend the time to look after it. So I've decided that I will have to cancel my attendance. Damn.
While here, Peter also solved the riddle of why pain, my Microsoft box, didn't print. When installing the software, I was asked for IP address or name of the printer (lp.lemis.com). The software didn't accept the period symbol (.), so I entered it just as lp. But that didn't resolve (because of some problem with the DNS setup), so it just sat there, with no information about what the problem was. With BSD you'd at least get an indication that the name didn't resolve. Changed the name to lp.lemis.com and all was well.
In the afternoon, discussing Apples on IRC. The responses I got confirm my concerns that people no longer care about doing things right; instead, they do what the manufacturers tell them. Finally, however, I at least found a way to remap the modifier keys; it seems it's there not even particularly hidden in one of the menus, maybe only since recently. At least now I have my Ctrl and Alt (sorry, Option) keys where I want them. Getting something corresponding to Microsoft's remote desktop facility still seems difficult to attain. Got some information on Synergy, something like a cross-platform x2x, that might make things easier.
In today's Mount Barker Courier:A P-Plate driver was caught speeding twice in 15 minutes along the South Eastern Freeway by different police patrols on Sunday night.
The 18-year-old from Sellicks Beach was pulled over near Mt Barker at 9:30 pm because he was allegedly travelling at 138kmh [sic] in a 110 kmh speed limit. Then at 9.45 pm he was pulled over near Hahndorf after allegedly driving at 147kmh in a 110kmh zone.
Hills traffic manager Sergeant Brian Schmidt said that the man was arrested after the second incident and charged with travelling at excessive speed and breaching the conditions of his probationary license.
That's really not very different from my made-up report yesterday. Points to note:
None of this should suggest that the driver in question wasn't driving dangerously; but he wasn't charged for that. The legislators should concentrate on direct dangers, not indirect ones.
| Thursday, 5 October 2006 | Echunga | Images for 5 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
I've been a subscriber to c't magazine for decades, and for some reason from time to time it arrives in really poor condition. Last month the cover was torn, and this month it didn't look much better:
It comes from the other end of the world, so I suppose it's not surprising, but I get other magazines too, and it only ever happens with c't. By German law, the subscription department is responsible for getting a good copy to me (several retries if needed), and occasionally I've asked them to do that, but they seem to have the general impression that it's because I choose surface mail rather than paying the significant additional cost for airmail. I suspect that their mailing house has something to do with it.
More programming today, my first ever program written from scratch to access a MySQL server. One of the results was surprising. While trying to log into the server (without valid user information) I got:
It was repeatable (after rebooting), and I was able to confirm that it was really executing /sbin/shutdown, so I changed the permissions on that program, but still got a corresponding error message. Even after fixing the access permissions, it continued trying to shut down. Established it was happening in mysql_close, so set a breakpoint on fork and found:
That looks innocent enough. The code in vio_close is:
But shutdown is a system call that shuts down a socket; why is it invoking /sbin/shutdown to shut down the system? The clue is above: it's at line 357 of misc.c (the fact that it has a line number at all is a smoking gun). In our application we had:
So it's a name conflict. Changed the name to system_shutdown and all was well.
Yvonne into town this afternoon to get the carbon dioxide cylinder refilled. Steve found a damaged valve component in the cylinder:
|
|||||||||||||||||||||||
He assumed that this was the cause of the cylinder emptying, and gave us a free refill as a result. I think he's wrong; the real issue is probably that the connectors between the cylinder and the regulator.
La Tigre had a foal in the evening, a buckskin filly, just what Chris ordered:
|
|||||||||||||||||||||||
It was very big and rather weak, and Yvonne decided to call the vet, who was able to confirm that nothing was seriously wrong.
| Friday, 6 October 2006 | Echunga | Images for 6 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
On with my program today, and finished it, without much of interest to report.
The new foal, whom we're provisionally calling Luna because she was born at the full moon, caused some concerns; she's rather unsteady on her feet. Also had trouble with Samba, who was very unhappy about being separated from La Tigre, and tried to climb over the gates. We were very concerned that she might hurt herself, and finally moved her to the other horses, where she settled down after a while.
| Saturday, 7 October 2006 | Echunga | Images for 7 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent most of the day in the garden today; things are looking ridiculously dry, so spent some time working on the sprinklers. It was more work than I wanted, but less than I had expected. After starting the sprinklers (for the first time this year), the filter clogged up every two or three minutes; the water in the long pipe from the dam was full of mud and sludge, and I had to pump for nearly 30 minutes before it came clear.
Took some photos of La Tigre with her new foal, whom we have named Luna Dorada. This one is Yvonne's favourite:
|
|||||||||||||||||||||||
Yana back in the afternoon. She has discovered Tom Lehrer, whom I had found amusing nearly 50 years ago. Somehow it's disappointing that nothing better of that genre has come up in the meantime.
| Sunday, 8 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another quiet day. We had plenty to do in the garden, but the weather was against us: cold and windy. Did some work on guide ropes for my hops, but didn't finish.
| Monday, 9 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Peters Jones and Denton over for a talk today, during which we rearranged some priorities. Neither of them has a UNIX background, so I'll have to do some documentation to make life easier for them.
Two weeks ago I bought another Apple PowerMac G4 on eBay, this time for only $89. It finally arrived today (“All orders are sent the same day payment is received, except in rare circumstances.”). As promised, the disk was empty, but was able to install my old Mac OS 10.3 disks on it with a minimum of fuss; only the network configuration didn't work: for some obscure reason it got the IP address 169.254.85.8, which makes no sense whatsoever.
| Tuesday, 10 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Mainly work again today. I'm restructuring the synchronization in the application, which consists of multiple processes. The approach uses a shared memory segment with flags in each direction. This means that, where possible, each field can only be modified by a single process and thus doesn't need locking. We're not done yet, and it's possible that I won't make it, but so far it's looking promising.
| Wednesday, 11 October 2006 | Echunga | Images for 11 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Everybody's talking about the weather lately. August was the driest on record. I don't know about September, but so far in this month there has been no rain at all. According to the statistics for Mount Barker, the closest I could find, average October rainfall should be round 68 mm. The results are clear: normally the “fire danger” season, when outdoor fires are prohibited, starts on 1 December, but this year it will be 15 October, at least nominally. Today the temperature hit 31° in Adelaide, and there were over 50 bushfires in the state, and it looks like staying that way (35° predicted for tomorrow), so probably we've already entered the fire danger season. Also, normally we mow our hay in early November, but this year the paddocks have dried out, so we harvested what we could today:
|
|||||||||||||||||||||||
We estimate about 40% to 50% of the normal harvest; hay will be expensive this year.
APC magazine came today, with an article about me as “Australia's grandad [sic] of Unix”, a title to which I'm not really entitled. Still, the article (by Tim Gaden) was good, given the format. Reading it through myself, I found a couple of things that could easily be misinterpreted. In particular:
Greg bought his first Apple a few years ago, hoping to find tighter GUI-OS integration in Macs. He describes Apple as the most successful Unix vendor ever, but was severely disappointed, saying Apple makes the same mistake as Microsoft: making PCs easier and prettier but severing the link between the GUI and the OS).
This was pretty much what I said to Tim a couple of months ago. In retrospect, though, it's not what I'm trying to say. The big issues I have with Apple (and Microsoft) are not so much to do with links as structures. In particular, I can't find any clear organization in either system, and the idea of a path name for a file has all but disappeared—and that when they're the basis of the concept of a URL. Also, this insistence on putting spaces in file names is evil. It breaks all file name list parsers I know, including all UNIX shells, find and Microsoft's COMMAND.COM. If there's one thing that I really hate, it's that—and it's avoidable.
We've been trying for years to sell our old tractor, and today it finally happened. Finally a bit of space in the shed.
On the work front, completed my transition to state-based synchronization. I had hoped to start testing today, but it now looks like tomorrow.
| Thursday, 12 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Started testing my changes this morning, and of course found a few things I hadn't thought of (or, in some cases, known) that meant that I had to go back and make some changes. It's interesting how we're doing the testing, though. Peter Denton is in Hahndorf, only about 17 km away, but he could equally well have been anywhere in the world: using watch(1), I took control of a couple of ptys on his display, and we both accessed the same shells with our own keyboards, maintaining telephonic contact (with head sets, of course). Peter told me the names of his pty, and I enter:
=== root@blackbox (/dev/ttyp9) ~ 4 -> watch -W -n /dev/ttyp0
After that, all text that appears on /dev/ttyp0 also appears on my screen, and I can enter text too. There are a couple of minor restrictions:
The approach very much like the “Agile Programming” that D'arcy Cain was telling me about in Ottawa in May, but with the difference that each of us has a different display and keyboard, so it's more comfortable. It's a pity that it's not better known.
Also writing some style guidelines for coding. I've done plenty of complaining about other people's style in the past, but this is the first time I've tried to write down my own style. Still not finished.
| Friday, 13 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day where I seem to have got nothing finished. But then, maybe that should be s/day/year/. We need dynamic web content for our application, something that I've been meaning to learn up on for 18 months now; back in April 2005 I bought a book on MySQL and PHP, and I still haven't finished reading it. In the meantime I have learnt enough about MySQL, but writing the web pages has still eluded me. That's all the sillier because it's probably all I need to be able to write my own database editor. Today I finally got started in something I seldom find time for: reading the book.
More problems with my fire extinguishers: I had guessed that the problems I had with the other cylinder were due to leaks in the connectors between the tap and the regulator, so with this cylinder I took to leaving the tap off most of the time and only topping up. This showed (by the lack of pressure drop on the high-pressure side) that any leaks were minimal, maybe because I had tightened things more. But today the pressure dropped immediately I turned the tap off, accompanied by a hissing sound. So it looks as if the piston in this cylinder tap has gone too. I suppose they're not designed for frequent opening and closing.
| Saturday, 14 October 2006 | Echunga | Images for 14 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Yet another slow day, spent mainly brewing. Also took a look at the telephone wiring for the house: since leaving MySQL, my second SPA 3000 VoIP adaptor, originally bought for the house, is completely unused. With the last issue of APC, I got a free voucher from mynetfone, a new VoIP provider, so it sounded like a good time to try to wire the second SPA in place. That wasn't helped by my favourite problem, lack of documentation. What is all this?
|
|||||||||||||||||||||||
I know that the terminal block is supplied by Telstra, and that only two lines are connected, not including the one I'm looking at. That's the original phone line for this house, which enters via a separate cable. Where is it? And what are these three lines?
|
|||||||||||||||||||||||
I'm guessing that they're in fact the line I'm looking for, and that the three cables are the incoming cable (which may also be connected elsewhere in the room, but probably nowhere else), one outgoing to my office, and one going to the old part of the house. But that's a guess, and I'll have to check it out to be sure. I paid good money to have this installed some time ago; they could at least have labeled the cables.
| Sunday, 15 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Messages from Achim Patzner and Mads Martin Jørgensen today pointing out that the functionality I praised in watch is also available from screen. I didn't know that. I haven't used screen for about 20 years, when I tried to use it on a Tandem machine to multiplex 80x25 terminal displays. At the time I wasn't very impressed. There were compatibility problems, some of which could have been due to the Tandem environment: the home terminal because a process, and some functions didn't work. Under UNIX there are also issues: it has a control character c-a which is also used by Emacs and bash. watch has a somewhat similar issue with c-g, but that only affects the watching process, not the watched process. One other obvious advantage to watch is that you don't need any preparation. You can attach to any terminal.
I'm setting up a test machine for the ICT Council for South Australia web site upgrade. In the process, tried to start a window from the machine to echunga. Not for the first time, it hung—not just the connection, but the entire X server on echunga, spread across 3 displays. This drives me mad: what appears to happen is:
There appears to be no way to recover this situation except to restart X. This must be a bug, but I can't work out how it happens. Tried installing ssh-askpass in the hope that it would solve the situation, but it did nothing useful: on starting X (with keys loaded), it prompted me not once but three times for the passphrase (once per display), which was completely unnecessary. And then it still hung the session when I tried to access quartet.
Also took advantage of the situation to replace a monitor: my old Hitachi SuperScan 813s are gradually fading, and I also, sadly, have to recognize the fact that 1920x1440 is too high a resolution for normal work, so I dropped back to 1600x1200 with one of the new BenQ monitors.
Apart from that, connected up my Sipura SPA 3000 to the home phone line. My guesses yesterday were almost right: the three cables were the home phone line, but one goes to the Hi-Fi cupboard (where the main phone is connected) and the other to my office. That leaves a phone in the kitchen, and possibly another in bedroom 1, that are before the Sipura box. I think I can live with that, though.
Next step was to sign up with mynetfone, which was relatively straightforward up to the end, where I was presented with a message saying that Mynetfone would contact me within 2 days. I suppose it's reasonable to do some checking, but it was somewhat unexpected.
| Monday, 16 October 2006 | Echunga | Images for 16 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Various things round the property kept us busy today. Brian Schulz came in to bale the hay, which for once went well, and we even ended up with more bales than we had expected—267 from 3 paddocks, compared to about 271 from two paddocks last year. Given the extreme weather, that wasn't too bad.
The irrigation system is giving us trouble, though. The water supply to the pump keeps drying up, though there's plenty of water in the dam. Went up to the dam to confirm, after which the water seemed to flow better, but after putting the pump on to it, discovered a significant vacuum—that can only mean a blockage. The pipe's 5 cm in diameter and about 400 metres long; how do I clean the blockage out? The best bet I have so far is to blow water up the other direction, but that would require a fair amount of preparation.
Also got the promised mail message from mynetfone with all connection details. Amazingly (for VoIP) that and their Sipura SPA 3000 setup instructions got me up and running first time after only 15 minutes of configuration. That should sound bad (a whole 15 minutes to connect a phone), but in fact it's amazingly good compared to what I've experienced in the past. They have had the good idea of showing the original Sipura setup web pages with the things that needed changing highlighted in yellow. That way I could just run two windows on my display, one from the web site and the other from the Sipura device, and cut-and-paste across. About the only thing that caused any confusion was the choice of phone number; Sipura calls it a User ID, and the registration letter calls it a “MyNetFone Number/SIP Alias”. And of course, this is an excellent illustration of where separate windows win hands down over “tabbed browsing”.
On the work front, spent most of the day writing my first PHP application, to display (later update) a MySQL table. It's surprisingly simple.
| Tuesday, 17 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day where I got almost nothing done. Played around a bit with PHP and learnt a bit, but I'm still not finished.
Yana along in the evening to pick up her Apple G4 PowerMac. In preparation, finally connected up a power supply for the old G3 PowerBook so I could back up the data on it. Last time I had looked at these power supplies was when I bought them. I had had difficulties with the connectors and put them on the “too hard” queue. Today discovered that the connectors are almost identical with the connectors on old AT motherboards, so pulled some off an old power supply and created an ugly but functional connection:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Also tried putting a CD-RW burner in the machine; it was equipped only with a DVD-R. That showed things up less well. There are two slots in the front of the machine, but one is 3½*, designed for a Zip drive, and though there's plenty of space, the fittings won't allow it. Replaced the DVD drive and discovered that, though the burner was recognized as a CD-RW, the software appeared not to be able to burn on it, showing a corresponding message. That's the second time that I've had problems with Mac OS X not handling optical media well.
| Wednesday, 18 October 2006 | Echunga | Images for 18 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up early with the brilliant idea of setting Yana's machine up to access the Internet via PPP; I have 10 hours free with my ADSL plan, so she could connect effectively for free. Installation was not easy: I couldn't even find where to start. It would seem that selecting System Preferences/Network would be the place to go, but all I saw there was “Built-in Ethernet”. Spent some time looking in different places, all to no avail. Apple had a three-part article on their web site describing PPP setup, but only the third part was still there. It gave me enough information to convince me that something was wrong, and then I discovered that the System Profiler reported: “Modem model: Modem driver not found”.
With a bit of googling discovered a suggestion to reset the configuration PRAM, which I did to no avail (reboot and hold down the Command(Apple), Option, P and R keys until it beeps at least three times; it will continue that cycle as long as you hold down the keys). Finally, checked the motherboard and found that one of the temperature sensors I had noticed when I get the machine was in fact the connector to the modem, which was missing.
|
|||||||||||||||||||||||
What a waste of time! At the very least, the system could have told me that the modem wasn't installed (I discover it now does in 10.4). But the other PowerMac (boskopp.lemis.com) did have a modem, so I installed that:
|
|||||||||||||||||||||||
The rest was plain sailing.
On the work front, finally managed to build a web page which would present data, let me modify it and send it back to the server. The way it did so reminds me of so many second-rate web pages, and makes it clear why they're second-rate. Now to learn how to do it correctly.
Yvonne took Yana back with her machine, and returned with yet another machine, from the ICT council, which we're going to set up and put in the office as a trial new web server. That'll be interesting.
Phone call from Yana in the evening: her monitor wasn't working. Talked her through it and came to the conclusion that she probably hadn't plugged the data cable in before powering on. That shouldn't stop the display from working, though.
| Thursday, 19 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Came in to the office to find another message about Apples from Achim Patzner. To confirm, put the CD-RW in boskopp.lemis.com, which is identical with Yana's pippin, and it worked. So there must be some system-dependent issue there. pippin is running Mac OS 10.3, and boskopp is running Mac OS 10.4. Noted also that the System Profiler correctly reports that the modem is not installed, and that when I—accidentally—rebooted the machine without a monitor, it didn't work after being reconnected. It's clear that it needs the monitor to be connected to know its parameters, but it could at least fall back to a standard VGA monitor or some such.
In the background, installed software on the ICT council web machine. The method I've thought up for the black box seems better than what I've done in my long-delayed system upgrade procedure. I should probably integrate it.
Still more work on the web pages, interrupted by having to go to Echunga for a haircut. PHP may be pretty straightforward, but you certainly can't say the same for HTML data entry. Made slow progress; maybe I need another book or text after all.
I don't seem to be the only one having trouble with web-based database access. I have a cron job which sends me copies of the TV programme from various Australian TV stations every morning. Today the copy of the SBS programme brought rather unexpected content:
For a while I thought it was a problem on my server, and I went to the trouble to confirm that yes, /tmp/mysql.sock was there and the mysqld was running, but the pathname was all wrong, and I'm using PHP 4.4, not PHP 3. It was reproducible on the original web site.
| Friday, 20 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Mail messages from Monty and Patrik Backman, the latter telling me that they don't want to do the table space implementation, the former to tell me that I should call him at all costs. That I did: it seems that Monty is not happy with the way MySQL is going, and he intends to create his own company, in which there will be a place for me. Just what I predicted a couple of weeks ago. This could be fun.
Spent most of the day working on the web pages, and finally got something that works. I'm not happy with it, though, and it brings home to me how rigid the framework of HTML is. The program works, but it's so cumbersome that in comparison to writing the same code in C. This isn't the fault of MySQL or PHP, both of which work well (though Rasmus is wrong when he says “there is nothing to understand. Just toss a few $'s in front of your variables and pretend it is C”). The HTML/HTTP side of things is the real problem. There must be a better way of handling forms in a client/server environment.
| Saturday, 21 October 2006 | Echunga | Images for 21 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Didn't intend to do much with computers today, but since I haven't seen much progress on the ICT Council Web Site, played around with that a bit. wget is really quite a useful tool in these circumstances; I was able to pull down the entire web site with
Then set to playing around with the HTML, which got bad marks from the W3C validator. With HTML tidy and a couple of Emacs macros was able to convert the first few pages very quickly. That still leaves me with stuff like this:
Probably all of these <font> commands are superfluous, inserted by an HTML editor with tunnel vision. But they're legal HTML, so I'll leave them there.
Spent the afternoon investigating the blockage of the dam water pipe that I had confirmed on Monday. Decided to use the firefighting pump to pump the water out of a drinking trough back up into the dam. That had the double advantage of getting the firefighting equipment in working order; we're expecting to have to use it this year.
Down to Meadows Landscape and Irrigation to buy some fittings, and spoke to Steve Dawes, who gave me some tips about how to go about it: I was going to suck out of a tap, but that won't work with suction. Set up my equipment ready to pump at the tap and at the end of the underground pipe from the dam:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The only thing missing was the suction pipe to the pump, which Steve brought came along later with that and took a look, but found further problems: there's a non-return valve in the line to the trough. Decided to give that up and just put the suction pipe directly into the trough. The other end of the pipe is in the distance: the second photo is a detail of the first, and the dam is just to the left of the shed in the distance.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The good news is that it worked: the level of the trough went down rapidly, indicating that the water must be going somewhere. Yvonne had gone up to the dam to watch the fun, but there wasn't much to see, just a few ripples and some bubbles showing that things were coming out near the blue float:
|
|||||||||||||||||||||||
Stopped after half-emptying the trough and confirmed that the water was coming out freely again. Started up the sprinkler system, and watched it stop after a few minutes. Cursing, checked the pipe, but that was still OK. The problem was the crud in the water, which clogged up the filter worse than I had ever seen:
|
|||||||||||||||||||||||
Decided to flush the thing again, this time with the entire contents of the trough. Some of the crud might have come from the trough, so first emptied it, including a significant amount of sludge:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
After that, and refilling the trough through the filter, pumped it all back again and finally managed to get some water that wasn't too dirty. I suppose we should be happy that it's all ending up in the filter: that's a lot easier to clean than if it deposited itself in the pipe.
| Sunday, 22 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More garden work on the agenda today, planting still more trees. Now that the irrigation system is working again, I need to address the sprinklers. There are two issues: the power supply (same as the one I cobbled together for tomato), which shouldn't be difficult, and a computer to drive the sprinkler relays. Last time I used an old 486 machine which in a still prior life had run my ad-hoc access points, but I wasn't sure which machine that was any more, and anyway I had an old Dell Latitude prototype which I had in previous times used as a kitchen computer. It had been running OpenBSD because I couldn't get FreeBSD to start X on it, but for some reason it wouldn't boot (errors accessing the kernel), and since I no longer wanted the display decided to install FreeBSD on it. Not for the first time, I had difficulty. This time the system (FreeBSD 4.11) recognized the PCMCIA slots, but hung when configuring the Ethernet interface. That was recoverable, but the NFS installation then tried to install on the wrong place, on top of the installation RAM disk, where it promptly ran out of space. There also wasn't enough stuff on the installation disk to do it manually. After a lot of playing around, came to the following successful solution, which almost completely bypasses the installer:
There should already be a device /dev/ad0s1 with the minor number 0x200002. That's really just a name for /dev/ad0s1c; /dev/ad0s1a should be the same except for the 0 in the last place.
Mount the root file system on /mnt:
It may be necessary to first fsck the partition.
=== root@echunga (/dev/ttyp0) /cdrom/floppies 107 -> cat bin.?? > /src/image
The fixit image I have didn't support NFS, so I couldn't NFS mount it.
| Monday, 23 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Had intended to spend today reviewing documentation, but Peter Jones ran into trouble with the web server in Hahndorf. Microsoft “Internet Explorer” said Page cannot be displayed. What a stupid error message! Does that mean that the web server is broken, that the network connection is down, that the page is in a format that “Internet Explorer” can't process? On the face of it, it's the last, but in fact the web server wasn't running, and it took me a while to get it up and running: I thought I had made it an exact copy of the one I have here, but the build options for PHP were different. More things to think about.
Yesterday the remote control for the electric gate at the entrance to the property failed. I suspected a flat battery, and Yvonne rang up the supplier, Total Gate Automation, and spoke to Barry, who told her that the device had a short circuit on the circuit board and needed replacement—for the proud sum of $75.90. We had paid $30 for it only a couple of years ago. Took the device apart and found a flat battery. After replacing it with one from another device, it worked fine. Thank you, Barry.
| Tuesday, 24 October 2006 | Echunga | Images for 24 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More work on the web pages today, spending an inordinate amount of time getting pages layouts right. The more I work with HTML, the less I like it. At least by the evening I had something that looked usable.
In the evening, more work on the sprinklers. We're adding drip lines around the house; this evening I connected up more surface lines to sprinkler 4:
|
|||||||||||||||||||||||
| Wednesday, 25 October 2006 | Echunga | Images for 25 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day where I didn't seem to get anything done. A bit of tidying up what I had done yesterday was interrupted by requirements in the garden, where the sprinkler and planting work still isn't finished.
In the afternoon into town to a board meeting of the ICT Council for South Australia, on the way having time to marvel at the numbers they give buses:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At the meeting we got more time than expected to talk about the web site, but still didn't cover all the matter. I had already published some draft requirements, which interested parties in the SA Open Source movement (so far none) could use to do their own base implementation:
You'd think that these requirements are pretty straightforward, but we discussed requirements at some length without even mentioning more than one or two points. In the end decided to have a 4 hour long meeting to get a professional requirements document done, which will cost real money. I'm obviously in for a learning experience, but at the moment I get the feeling that things are being complicated unnecessarily.
| Thursday, 26 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent the whole day today writing a single PHP page, not even very big. It's designed to set the system date. Twelve years ago, in Porting UNIX Software, I wrote:
UNIX timekeeping is an untidy area, made more confusing by national and international laws and customs. Broadly, there are two kinds of functions: one group is concerned with getting and setting system times, and the other group is concerned with converting time representations between a bewildering number of formats.
PHP goes one level further. It builds on the UNIX timekeeping functions and has similar functions with the same names. For example, the documentation page for localtime claims:
The localtime() function returns an array identical to that of the structure returned by the C function call.
How can an array be identical to a structure? I tripped over that before re-reading it.
But how do you parse a date? No UNIX function does it right. Neither does PHP, though it comes closer: strtotime converts most dates, except dates of the form most commonly used in Australia (today was 26/10/2006). strtotime insists on treating this format as American. I suppose that's a good reason for not using the format at all (I don't), but it's part of the spec, and I had to resort to turning dates around before passing them to strtotime. What a mess!
| Friday, 27 October 2006 | Echunga | Images for 27 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
And I thought I could move on from the web pages! Peter Jones was working on validation techniques, and since they were close to what I had already done, took a look at that too. Ended up with a better structure, but it took all morning, and during testing in the afternoon discovered that there was a strange bug involving strtotime: it refuses to accept times before 9:30. I've had lots of fun with PHP being too “tolerant” of my bugs, but after a lot of searching I can't find anything that I've done wrong.
At lunch time, a nagging feeling in one of my teeth got worse, and called up the dentist to get him to have a look. To my surprise he was able to look at me immediately—I was in the chair within 30 minutes from when I called to make an appointment, 15 of which were spent getting to Mount Barker—and found some gum problems which he handled less helpfully than I had hoped. Looks like the discomfort will stay for the best part of a week.
He wasn't just not helpful. In retrospect, what he did amounted to criminal negligence. He told me there was nothing I could do about the gum problems except clean my teeth more thoroughly, as if the whole discipline of periodontics didn't exist. It took years of periodontic care to get my gums back into shape. I'd mention his name, but in the meantime I've forgotten it.
It's raining again! The first even moderate rain this month, perhaps a millimetre or two, but better than nothing. Of course, that had its down side: back home to find that the power had failed. This was the two-hour variety, which conventional wisdom attributes to possums, but it happens surprisingly often when it's raining. I wouldn't mind so much, but every time something fails; this time it was wantadilla, my main machine, which seems to have a very sensitive power supply. And the big UPS still doesn't like the generator power.
| Saturday, 28 October 2006 | Echunga | Images for 28 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More rain this morning, and—another power failure! This is getting really annoying. And of course wantadilla went down again. I must retire that machine.
Spent the day doing some work on the sprinklers, but mainly with another attempt at tidying up the Mike Smith Memorial Room, which hasn't been tidy in the over 7 years that I've been using it. It's still not tidy, but it's looking better. Another week or two of continuous work and I might get somewhere.
The sprinklers are a nuisance because I can't find the documentation that I'm sure I wrote for them. The main issue is the relay board, of the same kind that is in the brewing temperature control system. Here, though, I don't have a computer power supply to run it, and in any case I need 24 V, though I did some thinking about using the +12V/-12V outputs of a conventional AT computer power supply: I found no less than 16 power supplies while tidying up, most of them functional and some of them brand new. But I the -12 V power supplies only supply 300 mA, which I don't think will be enough. The old board looked like this:
|
|||||||||||||||||||||||
The connectors at top left are 12 V DC for the card and for the pump relay (controlled by the left-most on-board relay). The other relays take 24 V to the solenoids. The junction block at bottom left had ground, ground, and 2 24 VAC inputs. The board on the right had a rectifier and two voltage regulators, one for 24 V, the other for 12 V. Strangely, the solenoids drew enough current that the 24 V voltage regulator overheated; thus the overhaul, but also my assumption that an AT power supply won't handle the load.
Di Saunders around to borrow a portable cassette recorder. I had one, which I bought about 28 years ago for use as my first computer data storage unit (300 bps on cassette tape). The battery compartment still looks a little unexpected:
|
|||||||||||||||||||||||
The unit was still functional as a normal tape recorder, but the drive belts slipped so badly that it wasn't worth the trouble.
| Sunday, 29 October 2006 | Echunga | Images for 29 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Somehow I seemed to spend the whole day today with the sprinkler system, and I'm still not done. In principle I only had a couple of wires to connect, but that took me over half an hour. Part of the problem is that I had to do it standing and leaning into the cupboard, but the biggest problem is the sorry state of connectors. The standard screw connectors that I can get here have a circular hole with a screw on it, offering a good chance for wires to miss the screw. By contrast, the connectors on the relays on the relay board do it right: the cross-section is rectangular, and the screw presses on a plate which covers the whole area:
|
||||||||||||||||||||||
|
||||||||||||||||||||||
|
||||||||||||||||||||||
I feel a rant coming on. So many of my projects are hampered by inadequate connections; this is only the tip of the iceberg.
Took some time off from that to—repair sprinkler lines. Some of them have been above ground and damaged by the lawn mower, and so I had to replace a number of sections.
Then did a little configuration of the laptop I'm using for the sprinkler system. It doesn't have an Ethernet port, and I had tested with a wireless card, but there's an Ethernet cable in the cupboard, so it seemed to be a good idea to put a normal Ethernet card in the laptop. After messing around for half an hour with four old PCMCIA cards, gave up. I don't know whether it's the support, the laptop, or possibly something I did, but I didn't get any of them to work. Put the laptop onto the relay board, and at least that worked immediately—for about 30 seconds. Then the sprinklers sank again, stayed down for a while, and rose again, repeatedly. Checking the solenoid in question, I discovered that it was very hot. I had already been confused by the 24 V requirement, so I checked it. Yes, it's a 24 V part—but it takes AC!
|
|||||||||||||||||||||||
What a waste of time! Clearly the solenoids worked on DC, but presumably they drew far more current than the 190 mA for which they're rated. It explains why the original power supply overheated, and presumably this one—a more sophisticated switcher—was crowbaring. And of course it meant that I could have fixed the original problem two years ago by a simple rewiring of the relays.
So I rewired the relays to use the AC transformer I still had left over. But they didn't work. The transformer also got very hot, which suggests that maybe I've damaged the solenoids with the DC. Still more work.
| Monday, 30 October 2006 | Echunga | Images for 30 October 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Out first thing this morning to turn on the sprinklers, and—there was no water! The pipe from the dam was clogged again. This drives me mad! At least I was prepared this time, so was able to back flush it relatively quickly. But will I ever be happy when things are finally running smoothly again.
More work on web pages, learning more about PHP as I go, including what caused the time problem: I was handing an incorrect parameter to strtotime (time structure array instead of time stamp), though it's not clear why it should have this effect,
Two months ago, Yvonne bought a saddle on eBay in Germany. Today it finally arrived. The surface mail postage was the least of the issue. Australian Customs, which apparently includes the AQIS, the Australian Quarantine Inspection Service, have recently reorganized and drastically worsened their service; but even that doesn't explain the careless manner they handled the whole thing. Read and weep.
19 October: I still haven't received the mail message. I call Anna on 8110 2400, and she said that she had sent it yesterday. Further investigation showed that it had been classified as spam:
And that was the entire message, apart from an attachment:
What can I do with that? Does Customs expect me to go out and buy software to decode the message? I get Anna to fax me the form and discover that it contains the question:
DO THESE GOODS CONTAIN DOG COLLARS INCORPORATING APPARATUS DESIGNED TO CAUSE AN ELECTRIC SHOCK OR PROTRUSIONS DESIGNED TO PUNCTURE OR BRUISE AN ANIMAL'S SKIN?
And yes, it was ALL SHOUTED. What stupidity! Why did they have to contact me separately about that and delay the proceedings still longer? Of course I confirm that this is a saddle and not a dog collar, and that there is no intention of harming the animal.
Today, after another 11 days have gone by, and 13 working days from the time when Customs received the documentation, the parcel finally arrived at Echunga post office. It took 22 days by surface mail from Germany to Adelaide and 31 days from Adelaide to Echunga. On previous occasions the delay at Customs and AQIS has been in the order of 7 to 10 days.
As if that weren't enough, the parcel arrived in so poor a condition that you'd think it had been done deliberately. It had been torn open (not cut) and not stuck back together again:
|
|||||||||||||||||||||||
Inside, the separate package with the saddle girths had been cut open and new adhesive tape applied, but they had not bothered to cover one end. When replacing the girths, they had crushed the accompanying documentation.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The second photo shows clear that Ute stuck things together with opaque brown adhesive tape, while AQIS used clear transparent tape. Some of this had also been used to tape a clearance document to the leather of the saddle:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Despite utmost care, it was not possible to remove it without damaging the saddle—see the remains on the tape in the second photo. I have no idea how we can repair that, but it's clear that AQIS, or just possibly Customs, have something to answer for. I called up Customs and spoke to Caroline, who told me that the new system was much better and they wouldn't be changing it; in particular, it's good for big business—so she says. But why should normal people be subject to this kind of problem? She also explained why the second form came later: it was computer generated. That explains everything, of course.
She redirected me to Australia Post, where Clayton told me that I was entitled to complain, since the goods had been delivered. So we'll have to go and fill out a form. I wonder how much good that will do.
| Tuesday, 31 October 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent some time this morning chasing up the AQIS issue, and got on to Enzo at AQIS, who at least sounded concerned about the problem, the first person who did. He was a bit surprised about the way things had gone; normally saddles are sent to Melbourne for radiation, as I also recall. In any case, looks as if we have more paperwork to fill out.
More work on web pages and things. I don't seem to be getting much done lately.
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 |
