Monday, 1 March 2004	Echunga
		Top of page
		next day
		last day

Who says that keeping a diary isn't productive? Got a mail message from somebody in the SA Police who read my report about my frustration in dealing with the police in relation to drug advocacy last month. He pointed out that the drug enforcement people are frantically overworked and have neither the time nor the energy to chase up “little” things like the issue at hand. It seems that they're completely occupied raiding drug production facilities, and even that isn't making much difference. No wonder they're less than enthusiastic. He appears interested, though. We exchanged a couple of long mail messages, and in the process we noted that the document in question contravened the Controlled Substances Act 1984, section 41 of which states:

Aiding and abetting etc

41 . A person who—

     (a) aids, abets, counsels or procures the commission of an offence against this Act; or

     (b) solicits or incites the commission of an offence against this Act,

is guilty of an offence and liable to the same penalty as is prescribed for the principal offence.

Looks like something's going to happen.

Apart from that, spent some time working on a secret new project. No details at the moment.

Tuesday, 2 March 2004	Echunga	Images for 2 March 2004
		Top of page
		previous day
		next day
		last day

More catchup work today. My software idea is gradually taking shape. Most of the work involved creating a source tree based on the FreeBSD distribution, but with symbolic links to the extras I'm adding. I hope that will allow me to run a CVS update to keep in step with FreeBSD, but still be able to manage my own changes.

Wednesday, 3 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

More work on my secret project today. In the process, I put a bug which scared me: I decided to put the two bits at the top of the second half of a flags word, so I needed a right shift to make an index out of it. I defined the shift as 16 instead of 14–not because I forgot the width of the bit field, but because, without thinking, I implicitly took the width of a halfword to be 18 bits. It's been decades since I last used a machine with a 36 bit word...

Thursday, 4 March 2004	Echunga	Images for 4 March 2004
		Top of page
		previous day
		next day
		last day

To Stirling this morning to meet and negotiate with Jim Johnson, with the result that I joined Rocksoft as the Director of Storage System Development. We've got some interesting projects coming up.

As of 4 March 2024, 20 years later, the domain rocksoft.com is still registered, belonging to Quantum, but nobody's home.

Then into town to meet with Denis Wall of the IT Council of South Australia in preparation for the monthly meeting on Tuesday. It's interesting that the IT Council has so far not come up with any viewpoint on the proposed Free Trade Agreement with the USA, which many of us fear will stifle innovation by further restricting use of intellectual property. In particular, reverse engineering will become more difficult, though it does seem that it will still be permitted for purposes of interoperability. The question is, does the ability to play a DVD from a different part of the world fall under the category “interoperability” or not?

Friday, 5 March 2004	Echunga	Images for 5 March 2004
		Top of page
		previous day
		next day
		last day

So I'm back at work! Spent most of the day getting organized, and not too much coding. The next step is newfs, which requires creating additional structures on disk. Current newfs creates two directories (/ and /.snap), and no files. At the moment it looks as if I'll need to build a cluster file, and spent some time looking at what information I have about the current implementation, which appears to be a directory hierarchy. I suspect it would be much more efficient as a single file managed directly, but it's possible that the current implementation will be adequate until we have time to do some tuning.

Saturday, 6 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Four years ago today I started working with Linuxcare. Things have certainly changed since then! The “Open Source” landscape has changed almost beyond recognition, as has the world. What a crazy millennium!

Didn't do too much work today. I'm trying to make time for doing other things than computers, but wasn't overly successful. Rain kept me from riding, and so I spent most of the day reading and bottling beer.

It seems that video recorders have passed their use-by date: in December, one of ours died, and the repair would have cost approximately half the new price—if we could still buy video recorders. Clearly the future is in digital recording formats, and if I go to Taiwan this coming weekend, I intend to look for something better. Unfortunately, we weren't left that much time: tonight our other video recorder died (also mechanical problems with the tape transport), so we're left with Yana's video recorder and the TiVo, which continually drives me to distraction. Today it hung up and reset the date to 15 May, and also refused to delete a whole lot of empty recordings. There must be a better way.

Sunday, 7 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Didn't do too much today. Spent some time looking for alternative video recorder solutions, and discovered that the whole situation still appears very confusing.

In the afternoon, out riding for the first time in a long time. Made the mistake of doing it after lunch, which is not a good idea after all but the lightest lunches. Still, maybe I can do this more often now.

Monday, 8 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Two power failures in one day! The first was just before I got up, and the second (the two-hour kind) started just before I had to leave for town. I'm getting particularly nervous now, since battunga has been up for over 540 days (and running the same X session since the last boot). Yvonne scared the hell out of me by asking “How do I stop the UPS beeping?” and reaching for the power switch, but I was able to stop her in time. battunga is still up. For once, we weren't the only people hit: a bushfire at One Tree Hill blacked out most of Adelaide.

The trip Into town was to meet with the Rocksoft people and discuss some software issues, but it developed into a full-scale strategy meeting.

In the evening, playing around with a couple of JPEG utilities in the FreeBSD Ports Collection, exiftags and jhead. Both dump image information from a JPEG. In the process, it's clear that xv removes significant quantities of such data. Here's an example with a pretty forgettable photo I took recently. The original, from the camera, is dscn3050.jpg, and the processed version is airlock.jpeg:

$ jhead *
File name    : airlock.jpeg
File size    : 245323 bytes
File date    : 2004:03:05 11:07:02
Resolution   : 1536 x 2048
Jpeg process : Baseline
Comment      : CREATOR: XV Version 3.10a Rev: 12/29/94 (jp-extension 5.3.3 + PNG patch 1.2d)  Quality = 75, Smoothing = 0

File name    : dscn3050.jpg
File size    : 1058921 bytes
File date    : 2004:03:05 21:15:02
Camera make  : NIKON
Camera model : E880
Date/Time    : 2004:03:05 10:45:03
Resolution   : 2048 x 1536
Flash used   : Yes
Focal length : 13.9mm
Exposure time: 0.017 s  (1/60)
Aperture     : f/9.4
ISO equiv.   : 100
Metering Mode: matrix
Exposure     : program (auto)
Jpeg process : Baseline

Not JPEG: info.txt
$ exiftags *
airlock.jpeg:
exiftags: couldn't find Exif data

dscn3050.jpg:
Camera-Specific Properties:

Equipment Make: NIKON
Camera Model: E880
Camera Software: E880v1.0
Maximum Lens Aperture: f/3.4

Image-Specific Properties:

Image Orientation: Top, Left-Hand
Horizontal Resolution: 300 dpi
Vertical Resolution: 300 dpi
Image Created: 2004:03:05 10:45:03
Exposure Time: 1/60 sec
F-Number: f/9.4
Exposure Program: Normal Program
ISO Speed Rating: 100
Exposure Bias: 0 EV
Metering Mode: Pattern
Light Source: Unknown
Flash: Flash
Focal Length: 13.90 mm
Color Space Information: sRGB
Image Width: 2048
Image Height: 1536
Color Mode: COLOR
Image Quality: FINE
White Balance: AUTO
Image Sharpening: AUTO
Focus Mode: AF-C
Flash Setting: NORMAL
ISO Selection: AUTO
Image Adjustment: AUTO
Lens Adapter: OFF

info.txt:
exiftags: doesn't appear to be a JPEG file; searching for start of image
exiftags: invalid JPEG format

It's interesting to note that exiftags didn't find any information at all in the xv output.

Tuesday, 9 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

More work on the file system stuff today, mainly spent finding my way round the CVS tree and understanding how FunnelWeb works: this is the main system that uses it. That will take some understanding:

In the evening to my first monthly meeting of the IT Council of South Australia, which was at 8—I thought. Turns out it was at 8 am, so I was 12 hours late. I've never heard of meetings of this nature at such an early hour. Obviously the typical member doesn't have the hacker nature.

Wednesday, 10 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

I'm planning to leave tomorrow for the AsiaBSDCon in Taiwan, but the flight details still hadn't been finalized. That kept me busy for a while. Apart from that, lots of long phone calls: with Ross about FunnelWeb, which is what the current code is written in. It's a variant of Knuth's Web language, and it seems to have quite a few good ideas, but it breaks all my C-based tools. We need to find a good way to use the advantages without the disadvantages.

Later another long call with Dan Shearer, who as usual had plenty of ideas about everything. Between that, talks with Liz—Linux Australia are planning a couple of press releases—and preparing for the trip to Taiwan, didn't get much else done.

Thursday, 11 March 2004	Echunga → Singapore
		Top of page
		previous day
		next day
		last day

Somehow even the shortest plane journeys take half a day: there's an hour getting to the airport, an hour waiting for departure—especially on an international flight like today—and an hour getting to the destination at the other end. The shortest possible flight is an hour, so that makes four.

That would have been enough if it hadn't been for the fact that my flight details hadn't been confirmed until 30 minutes before I left, after which I called Michael Wu (and, surprisingly, found him in Vancouver). At least I know where to go, but this must have been one of the closest calls.

Today was a little longer than the minimum, with a 6 hour flight to Singapore. Off round lunchtime, and bumped into Daniel Demuth, who had had some health problems and was only able to leave today. Interestingly, he checked in at the same time as me, though his flight was an hour later, and I had taken more time than usual due to the uncertainty of picking up my ticket. Daniel and parents pointed out to me that my flight was on “final call” fully 60 minutes before takeoff. Do the airlines really think they can fool people with nonsense like that? Into the departure lounge, where my portable tripod confused the security people, and had plenty of time to talk to Liz before we had to board.

Uneventful flight to Singapore, then to the Allson on Victoria Street, having more success finding the place than three years ago, when Richard Stallman was staying there, and we ran into a completely incompetent taxi driver.

Over to Raffles City for dinner, and then back early to bed. I think I know my round Singapore better than I do round Adelaide. The Allson, where I stayed for a while about 9 years ago, has become no better.

Friday, 12 March 2004	Singapore → Taipei	Images for 12 March 2004
		Top of page
		previous day
		next day
		last day

There's no point getting up early in Singapore: it's a hacker's city, and it doesn't come to life until 10 am. Up round 8 to finish off my talk for tomorrow, involving a significant fight with grops and the use of Emacs to fix the final PostScript before conversion to PDF, then off to look for breakfast. Got as far as Sim Lim Square, which has a food court underneath, by 9:30, but they were still preparing:

Image title: Food court          Dimensions:          2048 x 1536, 723 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Friday, 12 March 2004, thumbnails          All images taken on Friday, 12 March 2004, small Diary entry for Friday, 12 March 2004 Complete exposure details

 

So on in the direction of Little India and had breakfast at a Nasi Lemak place on the corner of Sungei Road and Serangoon Road, where I had had a beer with Lim Kim Yong last October.

Image title: Nasi Lemak          Dimensions:          2048 x 1536, 606 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Friday, 12 March 2004, thumbnails          All images taken on Friday, 12 March 2004, small Diary entry for Friday, 12 March 2004 Complete exposure details

 

Quite a different approach, and I ended up with a mutton curry, though I'm not sure it's typical. Certainly there was less juice.

Then across the road to a hairdresser with the unlikely name “Haircut 21”, where I have been before. Today was amusing. The only hairdresser there spent an inordinate amount of time doing his job. Though it appeared that he was well trained, he spent a lot of time doing other things, including helping one of his colleagues clean out the shrine near the entrance.

Then to Mohammed Mustafa Samsuddin, where I bought a new watch—again—but this time a good one, for $36, about half of what I paid at the end of January. It's really surprising that it's so difficult to find these things on the Internet, but you can just walk in and find them here. Also bought a couple more shirts and an electric toothbrush: the portable thing from Braun has been giving me electrical problems ever since we bought it, and battery powered ones don't work very well anyway.

Then the obligatory catastrophe. I think I shouldn't leave home any more. Yvonne had dropped a ball-point pen behind her desk, and while searching for it had disconnected the power to battunga, which earlier in the day had claimed:

battunga      up 547+01:07,     0 users,  load 0.07, 0.05, 0.00

Not only that: it didn't want to come up. At some time in the last 1½ years, I had upgraded the system, but not rebooted. When Yvonne tried, the kernel didn't want to know the disks, which had changed from wd to ad. She managed, through much ill-tempered international calls on my part, to get /dev/ad0a mounted, but we didn't have device nodes for it. Finally booted the old kernel, but the new login didn't want to let her log in—or at least that's what I think it is. It's difficult to diagnose these issues from another continent.

Finally, at the airport, tried to connect to the once wonderful Singapore Airport Wireless Network. It now costs money, and their credit card authentication doesn't work (“Please contact your bank to solve this problem”; what do they think I am?). While looking for a phone, found a “PC corner“ with free Ethernet, so used that instead, and finally managed to start X for Yvonne. By that time, it was time to board the plane for Taipei. Not the way I wanted to spend the day: missed lunch with Harish Pillay, and there was also some exhibition on which I had planned to go to.

In Taipei, met up with Warner Losh, who had just come in from Denver, and off to the Academia Sinica, where we arrived round midnight. While checking in, discovered that my plane tickets had been incorrectly issued, and I was booked back on a flight tomorrow at 1405. That will need some attention.

This place is freezing! The temperature in my room was 16°, with no way to warm it.

Saturday, 13 March 2004	Taipei	Images for 13 March 2004
		Top of page
		previous day
		next day
		last day

I was just planning to go down for breakfast at 9 am when I got a call from Michael Wu saying that there was a planned speaker's session at that time. Managed to get some breakfast anyway, where I bumped into Jeffrey Hsu, who is also speaking at this conference. As the name suggests, he's originally from Taiwan, but hasn't been back here for ten years. Reminds me of myself and Australia.

After the briefing, managed to contact Singapore Airlines, and fortunately I was able to change my flight. Nobody knows if any fees are involved, but that seems to be a good sign in itself.

Did my kernel debugging talk, shortened to 3 hours.

At some point I did a demonstration of remote debugging with the help of Warner Losh, connecting two laptops together. Unfortunately I had the kernel sources NFS mounted on the laptop that I was debugging, so when I tried to access them from the other laptop, I hung everything, and the demonstration was over. It seems that out of shame I didn't mention this, though “shortened” might refer to the fact, as might this photo:

Image title: BSDCon 14          Dimensions:          2048 x 1536, 798 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Sunday, 14 March 2004, thumbnails          All images taken on Sunday, 14 March 2004, small Diary entry for Sunday, 14 March 2004 Complete exposure details

 

after which we had a BoF session, then Jeffrey talking about DragonFlyBSD. They have some really interesting ideas there. I should try it.

In the evening, had a banquet. Good food, a far cry from what we get at most conferences. Spent some time talking after that, but most people seemed tired. Early to bed.

Sunday, 14 March 2004	Taipei	Images for 14 March 2004
		Top of page
		previous day
		next day
		last day

There were noticeably fewer people at the conference this morning. Went to Warner Losh's talk about embedded FreeBSD, then a couple of half talks by local people. After that, participated in a panel session on “The future of BSD”, which ended up being a talk on the features to be added to FreeBSD in the next couple of years. It's a pity that nobody thinks beyond this framework. That's not a criticism of the conference, more a general issue of viewpoint.

The panel session was originally planned for two hours, but we finished in one, making it possible to listen to Itojun talk about IPv6 via video feed from Tokyo. It worked better than last time at the AUUG conference in Sydney, but it sill left something to be desired.

In the evening went shopping in the local electronics markets, and discovered that I had made a big mistake by buying DVD+Rs in Singapore: they're about a third of the price here, and I got some DVD+RWs (most expensive kind) for half the price of the DVD+Rs I bought on Friday.

Then to an interesting restaurant, apparently a tourist thing: Kenjiro Cho had brought a Japanese guide book with him, and we found it in there. Still, interesting atmosphere.

After that off with the Japanese for a drink, and ended up in the Ploughman Inn. It was almost empty, and it was obvious that they had just changed owner. She (Anna Wu) came along and told us the story of her life and the history of the pub, in fact everything except for a purported minimum spend of $500 [Taiwan dollars, about 16 USD] per person. That came later, just before the bill. I challenged this, and she got very angry and insulting, which had presumably its desired effect on the Japanese, so we ended up drinking another round in significantly less friendly surroundings than the first.

In the process, discussed at length the differences between the Chinese and the Japanese: I asked what the popular view of the Japanese was about the Chinese, and heard the (for me) rather surprising opinion that they were very strong people, whereas the Japanese felt relatively weak. That makes more sense in view of a conflict like the one we had just had, where the Chinese owner tried on an undocumented minimum spend, and the Japanese acquiesced rather than make a fuss. It seems that many Chinese restaurants in Tokyo have very rude staff, and the Japanese find this somewhat refreshing after the overly polite behaviour which is normal in Japan.

Monday, 15 March 2004	Taipei	Images for 15 March 2004
		Top of page
		previous day
		next day
		last day

Met up with most of the overseas guests today: Sam Leffler and family, Robert Watson and wife, and Warner Losh, to go to the National Palace Museum. They had decided to go by bus and MRT, which meant that we got there just before midday, by which time I had had a series of phone calls asking me to be in Canberra for a meeting on Wednesday. After some deliberation and the unfortunate discovery that yes, indeed, I could be there by then, I decided not to go anyway. There's only so much you can do.

The museum was interesting, though after all people had told me about it, it was an anti-climax: “You haven't seen the Forbidden City until you've seen the National Palace Museum too“. In fact, I saw nothing of the treasures of the Forbidden City. Still, a very interesting day.

In the evening, we had planned to go to the night market (not my first choice, but everybody else wanted to go there). Then everybody else was too tired, so went there alone to meet up with Clive Lin and friends, then back to Clive's place to see how a real Taiwanese hacker lives.

More photos; the raw ones are also here.

Tuesday, 16 March 2004	Taipei → Singapore →	Images for 16 March 2004
		Top of page
		previous day
		next day
		last day

Another day spent with travel, with a little spice added by the meeting in Canberra tomorrow. As a result of the latter, left the Academia Sinica later than planned, and then discovered how bad the traffic here is in week days: it took 90 minutes to get to the airport.

In Singapore found, shown yet another Internet connection, this time with the options of wired, wireless or infrared. I don't do infrared, but was able to confirm that the wired connection was free, and the wireless version required payment. For obvious reasons, didn't check whether the wireless authentication worked.

Wednesday, 17 March 2004	→ Echunga
		Top of page
		previous day
		next day
		last day

What a pain these short overnight flights are! Had dinner round 2:30 am Adelaide time, and somehow managed to get three hours of uncomfortable sleep in before landing in Adelaide, where the lack of priority baggage handling made itself unpleasantly noticeable. And why do people have to stand in the way of the carrousel when their baggage hasn't arrived?

Back home through some of the worst traffic I've ever seen in Adelaide. Yes, it was rush hour, but I've never seen it this bad. At home, in catchup mode, somewhat complicated by the meeting in Canberra, which may mean a significant change in the way AUUG is run.

Thursday, 18 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Up in the middle of the night to go to a breakfast organized by the IT Council of South Australia, taking Yana with me. I usually avoid Hahndorf on a dirt road to the south and west. Yana queried this, since at this time of the day (6:45 am, not yet light) there was no traffic, and I answered that it was more relaxing, since nothing was liable to come out of a side road. The next thing I knew, three kangaroos jumped out of the bush right in front of me, causing me to brake heavily and throw the contents of the rear seat onto the floor. Once again the kangaroo whistles proved completely useless. One point to Yana.

The meeting was interesting. AUUG has barely joined the IT council, and already it looks like the organization will change completely.

In the afternoon, working on a product requirements spec. How useless Microsoft “Word” is! Converted it to OpenOffice and then tried to handle the XML code, in which I got further than expected. There appear to be a number of tools that will format XML legibly (and more than one line per file). The /usr/ports/textproc/libxml port includes a program called xmllint that will, with some persuasion, reformat and indent the text. For example, taking the following original excerpt from the single-line file:

<text:list-item><text:p text:style-name="P24">First offered in the US then subsequently to other English speaking countries, preferably in this order: United Kingdom, Australia, New Zealand, and Canada</text:p></text:list-item>

xmllint turns it into this:

      <text:list-item>
        <text:p text:style-name="P24">First offered in the US then subsequently to other English speaking countries, preferably in this order: United Kingdom, Australia, New Zealand, and Canada</text:p>
      </text:list-item>

That's still far too long, of course. Then I discovered that HTML tidy (/usr/ports/www/tidy), which I use for indenting web pages such as this, can also format XML. With the following invocation:

$ tidy -xml -raw -preserve -w 90 -i doc.xml  > doc6.xml

the result was:

      <text:list-item>
        <text:p text:style-name="P24">First offered in the US then subsequently
        to other English speaking countries, preferably in this order: United
        Kingdom, Australia, New Zealand, and Canada</text:p>
      </text:list-item>

Somehow this still seems to be clumsy; in HTML it would be:

      <li>
        <p>First offered in the US then subsequently to other English speaking
        countries, preferably in this order: United Kingdom, Australia, New Zealand, and
        Canada</p>
      </li>

Still, it's a great improvement, and it's only a short step from that to converting it to groff.

Friday, 19 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Finally caught up with the backlog from Taiwan, and did some more thinking about Funnelweb. It's an interesting concept, but at the moment it's just keeping me from doing real work.

At lunchtime, finally received the Digitrex GKX-9000 DVD recorder that I had bought on Ebay nearly two weeks ago. This appears to be the same as the Apex DRX-9000, barely warmed over for Australian conditions (for example, the date format is the wrong way round). Note that the Australian site claims it will record on DVD-R and DVD-RW media. I don't believe this.

The good news: it works, at least some of the time. Like many modern embedded designs, it looks hurried and rushed. The manual and the on-screen displays were obviously written by somebody who doesn't speak English very well:

5. In five minutes before it is ready to record the scheduled, it appears a Record Prompt dialog, affirm press OK, abolish press CANCEL, and if without operation for a moment it will record the scheduled automatically.

I'm curious about the background of that text; I'd suspect Eastern European.

Within a few minutes of installation, I managed to shoot myself in the foot: the player supports “progressive scan”, a new way to say “non-interlaced video”. I didn't know if my TV supported progressive scan, so I tried it. It didn't. Then, of course, I couldn't access the on-screen display, and the non-intuitive method of access (hint: in case of doubt, try Select and look for the tiny error message at the right of the screen) made it impossible to get a display again. It took a while to find the button I/P on the remote control, which toggles between interlaced and non-interlaced video.

Things weren't exactly plain sailing after that. I'm still trying to work out what's going on, but it seems that

By the end of the evening, I had established:

1. It's rather obtrusive. It has a loud fan, and the power LED is bright blue, about 2cm wide and 5 mm high. I've covered it over.

2. It's very slow. Functions like channel change have a noticeable delay, and some functions (stop recording, power down) seem to require accessing the DVD+RW: the latter takes about 30 seconds. The on-screen text suggests that it's writing to the DVD+RW, even when no write access seems necessary. It doesn't seem to do any harm, tough.

3. Interrupting it while it is recording is a Bad Idea. I managed to make a coaster out of a DVD+R like that. I also managed to hang a recording on a DVD+RW, and the only way to get any reaction was to power cycle it.

4. Power cycling makes it forget the date and time.

5. The user interface is very confusing. To change anything, you need to press Select. Do that at the wrong time and you'll erase what was there before (this applies particularly to things like recording dates and times).

Still, it wasn't very expensive, and with a bit of trouble I'll get over the frustration. I really need to get my act together to use computers to handle HiFi and video.

I've started a page with my experiences. As of now, it's pretty much identical to the text above.

Saturday, 20 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Somehow didn't get much done today. Didn't even get to go riding, though I had intended to. Mañana.

Spent some time playing round with the DVD player, which must have some of the most broken firmware I've ever seen. Surely it must be possible to do things better than that. I must start looking for a good open source project to get involved in.

Sunday, 21 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Talking to people on IRC about DVD playing today. Came up with a number of interesting suggestions: the ports mplayer and mencoder, rip with multimedia/dvdrip, make DVDs from MPEGs with multimedia/dvdauthor. There's also an article on the subject. All to be followed up on.

The Digitrex DVD recorder itself continues to astound. Tried to explain to Yvonne how to use the thing, and discovered we couldn't enter a start time: the thing has this stupid 12 hour format, requiring you to enter first a (modulo 12 hour) time and then to select AM or PM. The previous start time had been AM, and it was now after midday. The recorder refused to accept the time, stating that it had already passed. It also refused to leave the field to allow me to select PM first. What a load of crap! This is also the first VCR-like device that I have ever seen that doesn't clear the recording slots after the recording is finished. The instructions are also horribly vague, so wrote up how to do it. Hopefully we'll get used to it.

Out riding again in the afternoon. Darah has had a girth sore for some time now, and it still isn't gone. Maybe I should leave her alone for a while.

Monday, 22 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Spent most of the day looking at Rocksoft documentation today. It's stored along with the source code in FunnelWeb files, but the Makefile (confusingly called makefile, though it requires GNU make) doesn't build the documentation. Fixed that, sort of, and ended up with 38 individual PostScript files, totaling 912 pages. For some reason, my PostScript merge script didn't work (it created PostScript with missing definitions), so the complete document isn't usable. It's going to be a while before I can get my head around this.

Tuesday, 23 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Today was one of those days where I seemed to get nothing done. Spent some time tidying up my documentation, but still didn't manage to crack the overall documentation issue.

In the evening, more fun with the Digitrex DVD player. We had started watching a DVD last night, and wanted to continue. I had noted the time where we had stopped, and I wanted to find a way to continue from there. As with the TiVo, it wasn't possible, but the Digitrex is worse:

• There's a button GOTO on the remote control. It appears only to function in the EDIT menu, which appears to be accessible only from the DISC OPER menu. At any other time, GOTO seems to be invalid.

• If you select DISC OPER during normal play, the recorder resets to the beginning of the recording.

• On page 16 of the manual, it suggests that you can use the NAVIGATION menu to enter a specific time. This hasn't worked for me; I can only enter a Chapter or Title.

In the end, found the data the same way I have to on the TiVo: by manual search. It's frightening how little people use the possibilities of new technology. It should be easy enough to store (cache) the position in the recorder and have it remember it next time the disk is inserted.

Wednesday, 24 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Spent much of today wondering about documentation, and how to approach the 912 pages of documentation I formatted a couple of days ago. I'm still trying to work out what literate programming is. On the one hand I end up with an idea which I consider flawed, that you want to separate documentation and program:

The “program” then becomes primarily a document directed at humans, with the code being herded between “code delimiters” from where it can be extracted and shuffled out sideways to the language system by literate programming tools.

On the other hand you end up with statements like:

A more casual approach to literate programming would be that a program should be documented at least well enough that someone could maintain the code properly and make informed changes in a reasonable amount of time without direct help from the author. At the most casual level, a literate program should at least make its own workings plain to the author of the program so that at least the author can easily maintain the code over its lifetime.

It's difficult to argue with that. However, the article goes on:

The philosophy behind WEB is that an experienced system programmer, who wants to provide the best possible documentation of his or her software products, needs two things simultaneously: a language like TeX for formatting, and a language like C for programming. Neither type of language can provide the best documentation by itself; but when both are appropriately combined, we obtain a system that is much more useful than either language separately.

I suppost that a lot depends on just how you split up the documentation. What good is detailed technical documentation of a piece of code if it's not easily available when I'm looking at the code? Well, I could look at the FunnelWeb version and edit that, but then at least etags, indent and Emacs C mode are not available. I'm sure many other things are missing too, but I haven't got that far.

Took a look at the source files to consider manual conversion to C. That doesn't look like the way to go. Neither does repeating the documentation both for the TeX output and for the C code output; that appears to require defining the comments as a macro and hiding them, which obfuscates the original. It seems that it would be better to maintain everything as comments in a C file and have some script to extract them for text formatting.

Thursday, 25 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Still more work looking at the source code today, and spent some time looking at what would happen if I put the “literate programming” documentation back in the .c source file. That took a while, but at the end I had noted:

• The documentation was in fact spread across three different places in the source file, so reading the original (.fw) source file is not an easy way to go.

• As a result, some functions were documented twice (a short version with the function, and a long version elsewhere). I didn't find any contradictions, but this seems to be inviting them.

• The documentation printout of the file I was converting had a total of 59 pages. By adding the documentation to the source file, I increased its size by 450 lines, or about 8 pages. That seems a good relationship.

This isn't so much a criticism of the software at hand as it is of the method of documentation. In the process, I thought a bit about the Vinum documentation and came to the conclusion that it was probably no better.

In the afternoon to Stirling to have my eyes tested, by coincidence about 100 metres from where I had coffee with Jim Johnson earlier this month. I had been dissatisfied with my previous opthalmologist after three separate attempts to get a new pair of glasses (not reported in my diary) lasting nearly three months from October to December last year. Today I discovered that the lenses in my glasses were the wrong way round (swapped left/right). At least I made the right decision to change.

Friday, 26 March 2004	Echunga	Images for 26 March 2004
		Top of page
		previous day
		next day
		last day

Spent most of the day continuing my quest for documentation. Having converted one source file yesterday, I went to read the results. The very first function had no documentation whatsoever, and it called library functions that were in a different directory. Spent some time working out Makefiles to build everything, in the process coming to the conclusion that if you're using GNU make, you should:

• Call the Makefile GNUmakefile.

• Create a file Makefile that invokes GNU make. This can make things easier if you forget to explicitly invoke GNU make.

After a few minor glitches I was able to format the library documentation. I now have the glorious total of 4423 pages of beautifully formatted documentation, and there are more directories to investigate. If I spent two minutes per page reading each page, eight hours a day, five days a week, it will take me nearly a month to get through it. There is no indexing in the documents, so I would have to memorize just about everything, and for that two minutes per page is pretty fast.

I've been having my concerns about “literate programming” for a while, and I have now come to a realization: it's designed for people used to reading books. I don't think that that is a good paradigm for software development any more. In particular, some form of online indexing, such as hyperlinks or (preferably) etags make life so much easier.

Saturday, 27 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Made a recording with the Digitrex recorder on a DVD+R today. The quality was impossibly bad; only the first half displayed at all. This is the second DVD+R I've tried to burn on the machine, and the other one was just as bad. By contrast, the (cheaper) DVD+RWs work fine. Either I have a dud batch of DVD+Rs, or the Digitrex doesn't like them. To check the latter hypothesis, spent some time looking at mplayer to see if it can understand the result. mplayer is somewhat frustrating, and the first installation didn't install a GUI (though I'm not sure that this is a problem).

I was still pondering this when other problems became evident. We had been having sporadic trouble with our water pump (a long-standing problem; I'll never buy Davey again) for some time: the pump would stop and wouldn't cut in again when the pressure dropped, requiring power cycling. Wayne Graetz, the plumber, is due to come in for some work on Friday, but today the pump wouldn't pump after a power cycle. Further investigation showed that the tank was empty, and still further investigation confirmed that at least one cause was that Yana, despite information to the contrary, had been taking, in her words, ten to fifteen minute showers. I had asked her in the past to restrict it to 5 minutes, and I have my serious doubts as to whether they were only fifteen.

Then it occurred to me that we had a second tank at the bottom of the property, and I was able to confirm that it was full, probably 10,000 litres—certainly enough to keep us going until the rains set in. Sent Yana out to work out how to get the water into the other tank. This proved to be severely hampered by Yana's complete ignorance of how pumps work, and she spent some time learning. No water this evening.

Sunday, 28 March 2004	Echunga	Images for 28 March 2004
		Top of page
		previous day
		next day
		last day

Up relatively early this morning, without a shower, and found Yana bumbling around looking for ways to get the water from one tank to the other. Based on her investigations, started attaching some fittings to the tank, only to discover that the size was not as reported. All we had needed to do was to take the feed pipe from the fire water tank, including adaptors, and put them on the rain water tank.

Did that and pumped about 80% of the water to the other tank, which took about 4 hours and three tanks of petrol for the fire pump. They could have dimensioned the tank a little larger. Getting the Davey electric water pump to work was a lot more difficult than I had expected: this silly Hydrascan system seems to react in a very unfriendly manner to any air in the system, which is possibly the reason why it had failed so frequently recently. Finally got things going, though. Took some photos of the mess. After all that, we had some heavy rain, a total of 20 mm, so we could have got by without the trouble.

After that, didn't feel like doing very much. A few days ago I had severely trimmed my beard, and still had the cut-off hair left over, so put it up for auction on eBay, causing some mirth on the IRC channels:

** groggy has changed the topic on channel #<censored> to
+   http://cgi.ebay.com.au/ws/eBayISAPI.dll?ViewItem&item=3283911043
* dkaiser falls off his chair
<dkaiser> HA
* groggy wonders if it will sell.
<groggy> No, but it could end up with no bids.
<tpot> eww
<tpot> (-:
<dkaiser> groggy: do you have a photo of self sans beard ?  before and after?
<bje> groggy: does this mean you'll be taken off the watch lists now?
<groggy> bje: Who knows?
<groggy> In fact, I haven't shaved completely.  It's just a normal length now.
<dkaiser> groggy: i realized they were just larger version of the ebay photo
<groggy> More like in http://www.lemis.com/grog/old-office.html
<dkaiser> groggy: oh, i was just curious since I don't think a picture of a clean
+        shaven you exists anywhere...  not that i care, just thought it'd be different to
+       see.  kind of like when i showed people my shaved head photos
<groggy> dkaiser: I did shave it off completely about 18 months ago.
*** benh (~benh@202.46.137.12) has joined channel #<censored>
* groggy points benh at the /topic.
<groggy> Nearly an hour, and still no bids.
<benh> heh
<anton> benh: interested?
<benh> anton: hrm, nah, thanks
* groggy tries to sell anton his beard.
<anton> what advantages does this beard have over one i can grow?
<jk> (which he could also afford)
<groggy> anton: It brings me money.
<groggy> anton: And it doesn't need trimming (any more).
<hughhalf> groggy: I think you need some before and after shots to reassure people
+        of the authenticity :)
<jk> there's an alarming increase in the incidence of fraud in the worldwide beard
+        market
<groggy> hughhalf: No way.
<hughhalf> :)
<groggy> hughhalf: Well, in deference to your name, half way.
* groggy points at http://www.lemis.com/grog/grog-passport-thumbnail.jpeg
<hughhalf> Should post it to slashdot.  "Bastion of BSD shaves Beard!"
* groggy eggs hughhalf on.
<jk> sounds like a job for tpot
<hughhalf> Submitting now to the BSD section :)
<jk> heh
* groggy watches the "BSD is dying" troll crank up again.
<hughhalf> groggy: Any quote for the story ?
<hughhalf> Accustomed to the quiet normalacy of Sunday afternoons drinking Coopers
+         Pale Ale, the Australian BSD community was shocked when it was revealed shortly
+         after lunchtime that Greg "groggy" Lehey had shaven his beard and put the
+         remains up for auction on eBay.  Groggy, best known for his many contributions
+         to FreeBSD declined to comment on the authenticity of the item on eBay, or to
+         provide a current photo for verification purposes.
<hughhalf> posted
<bje> hugh: Fear not; I will bid and wear it.
* hughhalf chuckles
<bje> Well, to be fair, groggy has just made airline travel so much easier for himself
<hughhalf> Indeed :)
<tpot> is a bsd hacker without a beard still a bsd hacker?
<bje> tpot: I am!
<tpot> well for only $1 it can be yours!
<bje> Could Advance Hair Studio graft it on?
<bje> Err, "Advanced".
<groggy> hughhalf: Let me know if it actually makes it.
<hughhalf> groggy: Shall do, I don't think my prose was particuarly newsworthy but
+       we'll see how we go I guess :)
<groggy> hughhalf: I'd guess it depends on how bored people are.
<hughhalf> Might be a slow news day :)

Monday, 29 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

More work on the documentation today. Spent some time on the phone discussing the documentation concept with Ross. A lot of it makes eminent sense, in particular that externally visible functions should be documented in more detail than internal ones, and they should be documented in a different place. The problem I see is that it's only defined there, and that the lack of documentation in the source is a real problem for me at least. I suspect that part of the problem is working around other inferior tools; it seems that I'm the only person to use UNIX-based editors, and the others don't offer high-level functions such as ctags or etags and version control interfaces. As we agreed, we shouldn't change things to require the use of such editors, but it would be desirable to not restrict their users with workarounds for other problems. Simply moving the documentation to the .c files won't help either: the FunnelWeb files do serve a useful purpose for the external documentation.

In any case, Ross pointed me to some of the documentation to read, reducing my load to only about 300 pages. It's still difficult to get used to this different way of doing things, and I didn't get as far as I wanted.

Tuesday, 30 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

Into town this morning for another meeting, mainly about “performance” (a term used to include such atypical functions as robustness; I have the task of finding a better term), but also discussed our source code base and decided that we should consider restructuring. Now comes the difficult part: how? It's surprising how excited software people get about source code structure and management. Spent some time discussing the matter and I get the feeling that we're not communicating very well, so tomorrow I'll spend some time looking at how convert the files to a different format.

I now have two bids on my beard. Somebody suggested that I should donate the proceeds to charity, which sounds like a good idea, but for some reason (bug?) eBay won't let me append to the description.

Wednesday, 31 March 2004	Echunga
		Top of page
		previous day
		next day
		last day

More work on the source code conversion today, taking up most of the day. We've also been thinking of how to check for regressions, and I had the bright idea of comparing the object files. That was a non-starter: they were even different lengths. Looking at the assembler output didn't help much either. It was full of stuff like this:

 .LCFI4:
 .LM2:
        pushl   __stderrp
-       pushl   $80
+       pushl   $301
        pushl   $1
        pushl   $.LC0
 .LCFI5:
        call    fwrite

After some investigation, discovered that the numbers in question were line numbers. Other differences were the file names in various places. Spent some time eliminating these differences, but I also managed to introduce some real differences, so it'll be a while. At any rate the conversion is complicated enough that I don't believe that a program can do it. On the positive side, I don't expect to spend more than 2 or 3 hours per module, and we can do the changeover incrementally.

A couple of years ago I complained about the Trading Post web site. Today I got a message from a competitor, Cracker, who offer free classifieds and thought that I was being unfair to them by increasing Trading Post's web visibility, but not theirs.

They're right: they, too, have a web site that doesn't render correctly. This seems almost to be becoming the norm. Increasingly, web designers seem to be making assumptions about browser setups, in particular about the font size. I display my browser on wantadilla:0.0, which has a 2048x1536 resolution, so I use a font size of “20 point”; I don't know where the browser designers get their metrics from, but it's closer to 12 point. A large number of web page layouts can't handle that and truncate or overlap fields. It took me a while to realize that “Ade” in the top right meant “Adelaide”.

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.