Greg's diary--June 2003

Sunday, 1 June 2003	Echunga
		Top of page
		next day
		last day

Spent some time with Yana today looking at the MP3 player. For some reason it won't mount properly on her machine: it seems it has a quirk which means that it doesn't understand 6 byte SCSI commands, and release 4.7 of FreeBSD, which is what she was running on firefly, her laptop, didn't compensate for the problem, so we had to do it on my laptop (sydney, which runs 5.0-CURRENT). Another OS upgrade to do. Over NFS and 802.11b it takes forever.

Mail's been collecting again. Somehow there doesn't seem to be a good alternative to processing it every day.

Monday, 2 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

A strange thing happened today: I had nothing urgent to do. That so surprised me that I couldn't find anything non-urgent to do, though God knows there's enough to do. Spent some time half-heartedly tidying up my mail, then decided it was time to experiment with some cooking. When in Singapore I like eating nasi lemak for breakfast, so spent the afternoon playing around with that.

Finally got firefly upgraded to 4-STABLE. It didn't help much. The MP3 player still doesn't work:

umass0:   USB MP3, rev 1.10/1.00, addr 2
da0 at umass-sim0 bus 0 target 0 lun 0
da0: < USB MP3 1.03> Removable Direct Access SCSI-0 device
da0: 650KB/s transfers
da0: 125MB (256000 512 byte sectors: 64H 32S/T 125C)
umass0: BBB reset failed, TIMEOUT
umass0: BBB bulk-in clear stall failed, TIMEOUT
umass0: BBB bulk-out clear stall failed, TIMEOUT

Looks like I'll have to do an MFC.

Tuesday, 3 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Another day spent catching up. Maybe I should unsubscribe from a lot of mailing lists. At any rate, got some work done on Vinum, which is now not looking as neglected as it once was.

Wednesday, 4 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Finally found time to write one of the many papers I'm trying to get finished this month. I have a paper and a tutorial for AUUG 2003, the latter (on kernel debugging) also for BSDCon '03, and an article for acm queue. As if that weren't enough, I should be chasing the newspapers for writing an article about BSD: the SCO nonsense makes it an appropriate time.

Christiane Yeardley came along in the afternoon with her new three-horse float full of horses for us: La Tigre back after nearly 18 months, Serafina, a filly by Cali out of Seren, and the second horse we have called El Mago Negro, also by Cali and out of Magia Negra:

https://lemis.nyc3.digitaloceanspaces.com/grog/Photos/20030605/big/Horses-in-float.jpeg

Image title:	Horses in float	Dimensions:	2048 x 1536, 1229 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 Thursday, 5 June 2003, thumbnails		All images taken on Thursday, 5 June 2003, small
Diary entry for Thursday, 5 June 2003		Complete exposure details

Chris also brought some assorted old computer equipment with her, including an old monitor for Diane Saunders and an old 1999 vintage Dell Dimension machine, which had at least a dead power supply. It turns out that, although it looks like an ATX power supply and uses the same connector, it's wired differently. What a stupid thing to do! Decided that it wasn't worth ordering a new power supply, since we weren't sure if that was all that was dead, and started looking for a new machine to put together tomorrow. Finding your way through motherboard hell is difficult, but the prices are good: you can get a motherboard with built-in graphics, processor and memory for about $250. The on-board graphics got Chris thinking, and she ended up deciding she could do with two 19" monitors as well. Looks like the horses are going to take second place this time.

Thursday, 5 June 2003	Echunga	Images for 5 June 2003
		Top of page
		previous day
		next day
		last day

More work on Vinum in the morning. Some strange things have happened in FreeBSD release 4, and people are reporting panics badly. Had some trouble getting the system up and running; for some reason I had managed to put in an ifconfig executable which panicked the system, and then I ended up with a configuration file without a kernel debugger.

Before I could finally start testing, Chris Yeardley returned from a computer buying spree: effectively a complete new replacement for the Dell Dimension I mentioned yesterday. Athlon XP 2200+, 1 GB of memory and an MSI motherboard with onboard everything. Well, that's what I thought. It did have an onboard nVidia chip set, four USB ports and an RJ45 connector for Ethernet, even a PHY behind it, but it seems that the "optional" Ethernet chip was missing. Finally put it together; getting the fan on is becoming more and more difficult, and I ended up having to put a spare on because I was just not able to fit the fan that came with the processor. Connected Chris's disk with Microsoft 2000, which started booting and then decided it couldn't find the boot device.

Connected another disk and installed FreeBSD, which had no problem at all, but recognized the processor as a 1.3 GHz Duron. Didn't feel like removing the fan, which would have also meant having to remove the motherboard, so did a bit of investigation and discovered we were running the FSB at 100 MHz. After going up to 133/266, we got the right frequency (1.8 GHz), but the processor was still calling itself a Duron. Didn't have time to investigate further; mañana.

Friday, 6 June 2003	Echunga	Images for 6 June 2003
		Top of page
		previous day
		next day
		last day

Up earlier than planned this morning: somehow I had managed to advance the alarm clock by an hour. In to find some answers to my message about the Duron/Athlon question, one pointing out that the text Duron must be coming from the processor, since the only occurrence in the source code is in a comment. Also a URL which wasn't very conclusive. Finally found some information on AMD's web site which confirmed that the processor code (0x681) was indeed an Athlon. Did a verbose boot which also showed me the sizes of the internal cache:

Jun  6 08:33:09  kernel: FreeBSD 5.0-RELEASE #0: Thu Jan 16 22:16:53 GMT 2003
Jun  6 08:33:09  kernel: root@hollin.btc.adaptec.com:/usr/obj/usr/src/sys/GENERIC
Jun  6 08:33:09  kernel: Preloaded elf kernel "/boot/kernel/kernel" at 0xc0673000.
Jun  6 08:33:09  kernel: Preloaded elf module "/boot/kernel/acpi.ko" at 0xc06730b4.
Jun  6 08:33:09  kernel: Calibrating clock(s) ...  TSC clock: 1804856130 Hz, i8254 clock: 1193688 Hz
Jun  6 08:33:09  kernel: CLK_USE_I8254_CALIBRATION not specified - using default frequency
Jun  6 08:33:09  kernel: Timecounter "i8254"  frequency 1193182 Hz
Jun  6 08:33:09  kernel: CLK_USE_TSC_CALIBRATION not specified - using old calibration method
Jun  6 08:33:09  kernel: Timecounter "TSC"  frequency 1804103284 Hz
Jun  6 08:33:09  kernel: CPU: AMD Duron(tm) (1804.10-MHz 686-class CPU)
Jun  6 08:33:09  kernel: Origin = "AuthenticAMD"  Id = 0x681  Stepping = 1
Jun  6 08:33:09  kernel: Features=0x383f9ff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR,SSE>
Jun  6 08:33:09  kernel: AMD Features=0xffffffffc0400000<AMIE,DSP,3DNow!>
Jun  6 08:33:09  kernel: Data TLB: 32 entries, fully associative
Jun  6 08:33:09  kernel: Instruction TLB: 16 entries, fully associative
Jun  6 08:33:09  kernel: L1 data cache: 64 kbytes, 64 bytes/line, 1 lines/tag, 2-way associative
Jun  6 08:33:09  kernel: L1 instruction cache: 64 kbytes, 64 bytes/line, 1 lines/tag, 2-way associative
Jun  6 08:33:09  kernel: L2 internal cache: 256 kbytes, 64 bytes/line, 1 lines/tag, 8-way associative
Jun  6 08:33:09  kernel: real memory  = 1040121856 (991 MB)

The cache sizes are correct for an Athlon version 8 (Thoroughbred). It looks as if this particular processor chip is an Athlon masquerading as a Duron. Did a make world -j3 on it, which completed in 25 minutes, the fastest I've ever built a FreeBSD 5 world, so it's really just the name that's confusing.

Terrible weather today. For the first time since I had the satellite receiver, the rain was so heavy that it interrupted the reception. In the time from 9 am to 4 pm, seven hours, we had 60 mm of rain, 7% of our annual rainfall.

More work on Vinum bugs. The 4-STABLE panic turned out to be not in Vinum at all, so it looks as if some of the new code is walking over disk labels. That will be fun to fix. In the meantime, finally committed my remote debugging macros, which I've been working on for years. And almost immediately somebody piped up that the location where I had put them (/usr/src/sys/debugscripts, the same place as in NetBSD) was wrong and that they should go in /usr/src/tools/debugscripts. sigh.

Saturday, 7 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Internet abuse is climbing all the time. I've been intrigued for some time just how much of my network traffic is abuse. Today I looked at the firewall on my Linux satellite receiver box and found:

Chain FORWARD (policy ACCEPT 522 packets, 211331 bytes)
    pkts      bytes target     prot opt in     out     source               destination
  513267 75030538            all  --  sm200d *       0.0.0.0/0            0.0.0.0/0
  352363 17493623 DROP       tcp  --  sm200d *       0.0.0.0/0            0.0.0.0/0          tcp dpt:80
     106    17396 DROP       all  --  *      *       0.0.0.0/0            192.109.197.169
  160798 57519519            all  --  sm200d *       0.0.0.0/0            0.0.0.0/0

In other words, I'm blocking all incoming HTTP traffic (dpt:80). The first line counts all data and packets, the last line counts what is left after dropping the HTTP traffic and all traffic in either direction for the address my laptop uses on the rare occasions it runs Microsoft. Nevertheless, in the time I did these measurements (a few days), I had 352363 connection attempts totalling 17493623 bytes, or 69% of the packets and 23% of the data. It's really time that people with broken computer systems should be made responsible for their actions.

The other thing on the rise is spam. Over the last few days I've been rethinking how to block spam. For several years now I've been refusing mail from systems without reverse DNS. The problem there is that you can't just refuse point blank with a 5xx code, because it could be a transient error; instead you have to send a retry code (4xx) code. The result is that the server will try again later. The recommended time is between 15 and 30 minutes, and to continue trying for 5 days. Unfortunately, some spammers retry far faster. Today I saw one who retried every 10 seconds. That's 8640 tries a day.

Other spammers have reverse lookup. I've been firewalling them, and collected about 700 class C nets. They retry even more often, of course. The firewall statistics above show that it's not as bad as HTTP traffic, but it's still significant.

As a result, I've taken down the mail firewall. Instead I reject mail from specific sites. It seems that Spamassassin does a reasonable job of rejecting most spam, so it's not as much work as I feared. My postfixerrors file is now growing much more slowly. When I have the configuration ready, I'll put it on the web.

One of the spams that came through was from Fleetwash, trying to sell me car wheel washers. Well, we have an Acceptable Use Policy about that. Where I can, I'll take $5,000 in payment for spam. Only idiots stupid enough to leave their address can play, of course. We'll see what happens with this one.

My web site hits have really gone up on the SCO complaint rebuttal page; over 100 hits, which is high for my web site. I must have had a link from somewhere. Unfortunately, I discovered I hadn't finished writing it, so spent some time doing that. The more I look at this mess, the less sense it makes. I see that Robert Cringely has also come up with the suspicion that SCO might have included Linux code in UnixWare, though he's found a less obnoxious reason.

Somehow ducks were the theme of the day. Apart from a silly joke that Yvonne and Chris have been laughing about lately, the ducks on the pond have come back in force and spent a lot of the day walking and flying around. Then Yvonne was in town and saw a couple of white ducklings for sale, and bought them. Finally, we had a duck to cook for dinner this evening. Tried a Chinese stewed duck recipe, which didn't work out as well as I had hoped.

Sunday, 8 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Quiet day today. Spent some time working on debugging tools and found Yet Another panic in Vinum. I suspect that there's really something wrong with zaphod: I'm getting random crashes which show things that seem only to be explicable by hardware problems.

Monday, 9 June 2003	Echunga	Images for 9 June 2003
		Top of page
		previous day
		next day
		last day

Sad news coming into the office this morning: a commit message for revision 1.98 of src/usr.bin/calendar/calendars calendar.freebsd, removing a line. Alan Eldridge died on Friday, not of natural causes.

Yesterday Yana discovered that the front license plate of the Magna was missing. Last month we had heard from the police that somebody with our license plates had done a drive-off in Henley Beach Road. I didn't check whether the rego plates on the car were still there afterwards, which in retrospect looks rather silly. Made a report to the police, with the annoying consequence that we need to get new rego plates for the car.

Apart from that, another day with little to report. Spent some time working on my paper for AUUG 2003, and also more work on Vinum. I really need to fix my hardware: zaphod won't even make it through a make world, which also makes the significance of debug errors very suspect.

Tuesday, 10 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

More work on articles and code today. Finally gave up on zaphod and turned to the Duron machine which for some time has been the NetBSD machine daikon, but had also variously been a FreeBSD test box under names like daemon and monorchid. I had somewhat given up on this box a while back because it kept panicking on me running the same software that worked fine elsewhere. Now I recognized the panic: it was the Vinum bug that I solved last month. Reinstalled 5.0 from scratch and then brought it up to date, which kept me busy all day. It's interesting how faster machines encourage more expensive builds. I once tried to cross build 2.11BSD on a FreeBSD machine. I didn't get it working as I wanted, though the compiler worked fine. I estimate that nowadays a complete build would take about 10 seconds and probably run completely in L2 cache.

Apart from that, more documentation. This SCO story is getting more stupid by the day. Now they have shown common code to two non-programmers who confirmed that the code looked very similar, right down to the comments. /usr/include/sys/errno.h perhaps?

Wednesday, 11 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

AUUG SA Chapter meeting tonight, and I was to talk on digital audio under FreeBSD, essentially the stuff that I did at the end of last month. Spent most of the day writing a presentation for that, leaving little time for other work.

Into town in the evening, gave the talk, and came home again.

Thursday, 12 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Today discovered that a FreeBSD developer had made some changes and not tested them properly, with the result that Vinum no longer compiled on SPARC 64. Spent a lot of time investigating that: it was all in the kludges that I had to make to get Vinum work with the disk label code that I found five years ago. The recent changes are definitely an improvement, and I was able to remove the code, but I wish people would test things properly, and not expect me to help them out when they break things.

Testing the code on zaphod caused repeated panics, which made no sense at all in view of the fact that the code was no longer there. I think the machine has had it. It worked fine on monorchid. In the process, also did some changes to the debug macros. They now get installed automatically in the kernel build directory when you build a debug kernel.

Friday, 13 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Finally got an up-to-date kernel installed on monorchid and was able to verify that yes, indeed, there's some strange glitch in vinum_scandisk() and friends which causes it to occasionally try to open a drive without a name. Couldn't face that, so left it in the kernel debugger and went out to Google looking for documentation on how to tune a piano: somewhat to my surprise, Yana has spent some time learning to play the piano, and it badly needs tuning.

Didn't find very much. I'm left with the impression that piano tuners don't want other people messing around in their livelihood. Found a book about repairing pianos, but it specifically omitted tuning. Finally found an old free software disk with a text written for display on the screen, and instead of doing anything else useful, spent the rest of the day marking it up in groff and finally printing it. It clarified one of my problems, anyway: I had had this impression that the three strings of the treble notes had to be tuned a few cents apart, but that's wrong. They should all be the same. I suppose I've just heard too many pianos where it wasn't the case.

Saturday, 14 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Spent the morning tidying up the How to tune a piano book, including adding the appropriate references to the author, Ron Van Hoose, who hasn't been heard of on google for nearly 8 years. I wonder if he's still around.

After that, returned to the SCO nonsense. I'm sure there's a more articulate word than "nonsense", but I really think it describes my attitude. Today their deadline to IBM expired without us hearing anything about it. It looks as if they will now have to get an injunction to stop IBM from shipping AIX. The general feeling is that that will be difficult to achieve, which is probably why IBM hasn't said anything yet. Spent some time writing that up, which included porting sccs so that I could read the CSRG commit logs: one of the accusations that many people (myself included) have made is that USL deliberately removed the famous BSD copyright from BSD code. In fact, what I see is (taking /usr/src/usr.bin/syslogd/syslogd.c as an example):

D 5.41  90/06/01 19:01:17       bostic  54      53      00001/00011/01116
new copyright notice

Prior to that, the license was much simpler.

Finally got back to my Vinum bug. I wonder what I was smoking when I wrote that code. I think it still needs rewriting (it's far too i386 centric), but at least it works now.

Sunday, 15 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

In the morning committed my Vinum patch and wrote a little more about SCO, but I didn't feel like doing any real work, so apart from a quick NetBSD installation for the sake of AUUGN, decided to play around with recording data from (vinyl) records.

Before CDs appeared on the scene I already had quite a number of records. Since then I have acquired about 800 audio CDs, but to my surprise I had about 30 records with music on them that I didn't have on CD. My previous attempt had been thwarted by the fact that the driver for the Maestro 2B on kondoparinga didn't support inputs, but since then I had moved an old Technics amplifier (built about 1979) into my office, which enabled me to connect its preamplifier output to wantadilla and record like that.

Or so I thought. We really desperately need a multimedia guide for FreeBSD. I finally found a couple of programs (sox and GramoFile) which did the recording. I chose GramoFile because it promised to separate .wavs into individual tracks, but it didn't work. Everything seemed to work OK, but nothing got recorded.

After a lot of messing around, discovered:

You need to set the rec mixer value to be able to record. It's not enough to set line or mic. The latter will let you hear the music, but it doesn't get recorded (by default rec is set to 0).
The rec mixer source needs to be set to the corresponding input (at least). mic and line refer to the right hand and middle jack on the back of an ATX motherboard (usually marked with unrecognizable symbols).

With a bit of help on IRC, got things working. The nice thing about having the source is that you can fix problems like this, so from the next release on mixer will show the recording inputs with the default output:

$ mixer
Mixer vol      is currently set to 100:100
Mixer pcm      is currently set to 100:100
Mixer speaker  is currently set to  75:75
Mixer line     is currently set to 100:100
Mixer mic      is currently set to   0:0
Mixer cd       is currently set to  75:75
Mixer rec      is currently set to 100:100
Mixer line1    is currently set to  75:75
Mixer phin     is currently set to   0:0
Mixer phout    is currently set to   0:0
Mixer video    is currently set to  75:75
Mixer monitor  is currently set to  75:75
Recording source: line

After setting the inputs correctly with mixer's somewhat bizarre syntax, was able to record things. GramoFile's track separation logic is not as refined as you might expect: it split Beethoven's Choral Fantasy into 11 tracks, not quite what I was looking for. I'm obviously now going to have to find a good audio editor.

Monday, 16 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Wasn't feeling too bright today, and spent most of the time sitting around and reading. Did go to the trouble to research documentation on PC audio hardware. When you consider that just about every PC has audio hardware on the motherboard today, it's really difficult to find good documentation. Finally found some at MIT, a report on the AC97 standard, which at least gave some help.

There seems to be something wrong with my turntable: I can't adjust the tracking properly. As a result, the quality of the playback wasn't what it should be, and I considered using MP3 encoding. Experimenting with MP3 compression ratios was interesting: 128 kb/s seems to be good enough for the quality of the recording I got. Went all the way down to 8 kb/s, which was amusingly bad. Beethoven's Choral Fantasy sounded like a dogfight.

Tuesday, 17 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Another day with nothing much done. Discovered more about my turntable, and I am now able to make quite acceptable copies of the records. It's a can of worms, though: now I need to find good ways to index the software, and having MP3s on line means that I need to establish an audio connection with the HiFi system. And that again probably means a small laptop to use as a remote control.

Yvonne finally let our ducks out into the pond today. They really showed the meaning of the expression “like a duck to water”. They had never been in a pond before in their lives, yet they were immediately in their element. Here are some photos.

Wednesday, 18 June 2003	Echunga	Images for 18 June 2003
		Top of page
		previous day
		next day
		last day

Well, SCO have gone and done it. They've announced that they have terminated IBM's license to use UNIX. What a lot of idiots. It's interesting that IBM don't mention the fact on their web site. Somehow I can't see IBM suffering greatly. In addition, Byte published an interview with Chris Sontag, Senior VP of SCO. If the stuff we've seen before looks stupid, this interview puts it all in the shade. The only uncertainty is how accurate the reporting it. Spent a lot of the day writing up an analysis of the interview.

Apart from that, catching up on my mail; didn't do much else.

Thursday, 19 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

FreeBSD is ten years old today! How time flies.

Spent most of the day working on various documents. The SCO nonsense continues to amaze me. Also discovered that they're reprinting The Complete FreeBSD already, after less than two months. Looks like it's selling well, or at least a lot better than they expected. That gives me a chance to add some corrections.

While working on the SCO documentation, looked at the front of Peter Salus' Quarter Century of UNIX. The cover shows a number of buttons and stickers, including one at the bottom that says "System V does everything Unix does, only not as well". That would have been written by a Berkeley UNIX person, of course.

More playing around with cooking in the evening, and cooked a large quantity of Ayam lemak to eat with Nasi lemak.

Friday, 20 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Spam is a serious issue, and I spend a lot of time trying to combat it. But I'm coming to the conclusion that the RBLs (Realtime Black hole lists) are worse. Today I answered a message to the misc@OpenBSD.org mailing list and got a message from a French organization claiming (without any attempt at justification) that it was spam. What kind of idiots run these things? Sent them a reply asking them to explain. I wonder if they will.

This SCO affair is hotting up. Spent a lot of time writing up what's happened in the last few days. No wonder I don't get much else done.

Apart from that, also worked on the article for acm queue. It's really difficult to know what to write, but I think the best thing is to try and position it in the modern history of UNIX. The more we look at this thing, the more it looks like they're trying to get license fees for the use of Linux.

Saturday, 21 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

When I came into my office this morning, I discovered that I had had nearly 12,000 hits so far that day (US EST) on my analysis of the interview with Chris Sontag in Byte: I had been slashdotted. The number finally rose to 13,244.

Spent some time reading various other articles on the subject of source code theft and ended up writing another article on the subject. One of the people allowed to see the code did not sign the NDA (he was in Germany; the incident confirms my prejudices against the German legal profession), and was thus able to describe what he saw. It's interesting that the code was in the scheduler: Rusty went through the scheduler in his Linux.conf.au talk last year, and I'm pretty sure that the Linux scheduler owes nothing to UNIX. Also, the Linux code he was shown came from the Linux Kernel Mailing List, not the kernel. Both of those suggest to me that the code is more likely to have wandered in the opposite direction.

Didn't do much else during the day. I need to get this article finished.

Sunday, 22 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Another quiet day. Broke the back of the article about SCO for acm queue, with which I'm pretty happy now. It will turn out to be completely different from the web pages I'm keeping. Also spent some time rearranging the pages and putting them in their own directory, though unfortunately I'm going to have to keep the old places too, since there are so many links.

In the afternoon, off to the nursery at Belair National Park to buy some grevilleas, and also finally found some seeds for miniature acacias, both of which we want to plant near the pond. It's only the second time this month that I have left our property.

Monday, 23 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Still more work on the SCO affair today. I'm glad it's slowing down a bit; it's surprising how much time it takes up. Also a bit of work on the update to the book, but I'll need to do a lot more this week.

Tuesday, 24 June 2003	Echunga	Images for 24 June 2003
		Top of page
		previous day
		next day
		last day

It's amazing how much work writing coherent thoughts is. Today I finally finished the SCO article for acm queue and sent that off, then on to revise The Complete FreeBSD for the reprint, which went surprisingly smoothly. I don't have that much work piled up after all.

Wednesday, 25 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

I don't know what gave me the idea yesterday that I had finished the work I on hand. I only had a little work on the book, so went straight at it first thing in the morning. That kept me going most of the day: just about any change affects the layout, and one of the issues is that the reprint will only change specific pages. I really need a utility which will compare PostScript files and extract page pairs. Finally ended up with about 38 pages (19 pairs) which I put up on the web and sent to the reviewers.

Also more work on the SCO article. Fortunately there wasn't much to do there. It seems that they like it.

Thursday, 26 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Mail message this morning from Christos Zoulas: I've been awarded a NetBSD commit bit. That'll keep the boredom away.

Just when I thought I had the SCO article finished, got some mail from Kirk McKusick pointing out errors in my depiction of the USL wars. Spent some time on that; I didn't agree with all his points, but it's difficult to argue with him when I hadn't been there.

Off after that to Internode, who are having performance problems with FreeBSD 5.1. Strange indeed: their dual 2.4 GHz Xeons appear to perform less well than the Athlon 2200XP+ that I built for Chris Yeardley earlier this month. Spent some time with that without coming to any particular conclusion. On shutting down my machine, managed to shut down wantadilla as well. Grrr. I should fix it so that shutdown will only work from a local tty, the way it does on System V.

Mail from Kirk on my return: he agrees with my changes. That's a relief.

Friday, 27 June 2003	Echunga	Images for 27 June 2003
		Top of page
		previous day
		next day
		last day

More pressure in the morning to get things out. The review feedback for my book modifications was surprisingly negative in the sense that I had made a lot of typos. I need to be more careful. Also spent some time taking some pretty terrible photos of myself for the AUUG 2003 brochure, which is due to go to the printers today, along with a welcome message that didn't want to come out right.

After that into town for the second day in a row for the ADUUG lunch. Including myself, only four people had signed up for it, and one of them (Mark Prior, the organizer) didn't show up. Nevertheless, it was pretty active: we had had no less than 5 gatecrashers, including Dan Shearer, whom I haven't seen for a while, Mike Bruins, with whom I was working yesterday, and David Newall. David, Dan and myself had time on our hands, so stuck around until 3 pm. Then Dan and I set off to Dan's office in Hindley St. Funny place to work, behind a computer gaming shop. It's beginning to look as if computer gaming is really coming into its own, though.

Saturday, 28 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Plenty to catch up on today after the last couple of days out of the office. Spent a lot of the day catching up on my mail, which had swollen to nearly 3,000 messages, and was only partially successful.

After that looked at more audio stuff: I had bought a couple of connector cables yesterday, so was able to connect up the output of the sound card to the amplifier in my office. The sound was terrible, and I was left wondering what I had done wrong. In the process discovered that I had already installed mpg123 on sat-gw, the Linux box in the HiFi cupboard, and was able to play them to the HiFi system with no trouble and with adequate quality once I found the name of the Linux mixer programs (I ended up using aumix, though I'm told that there is another one called gmix). After a lot of messing around, including carrying kondoparinga, the laptop that I used to call sydney from one room and amplifier to another, decided that the problem must be in the analogue parts of my old amplifier. More investigation necessary.

Sunday, 29 June 2003	Echunga	Images for 29 June 2003
		Top of page
		previous day
		next day
		last day

More work on the book upgrade today, which took a lot longer than I had intended. There must be an easier way; the more trivial a change is, the longer it seems to take, and all attempts at automatically locating the correct pages in the resultant PostScript output failed miserably.

Took some time off to try an alternative approach to building the book which I had used up to the second edition. I had a shell script generate an enormous list of rules, one per chapter and depending on the previous chapter. It was useful for building individual chapters, which wasn't what I really wanted then, but it is what I want now. Last time I ran it, some time in late 1997, it was slow as a dog; now it runs without a perceptible delay. Even more surprising for such a crufty script was that it still worked. It needed tweaking, of course, and while doing that I thought of a better way, so back to proof read my corrections, and found more stuff to change. This is a never-ending job.

Wandered around the house with the mini-loudspeakers that I had taken off the old all-in-one music centre that I used to have in the office. It looks like they're the problem: I must have inadvertently blown the cones with too much volume at some point. I wonder where you find cheap loudspeakers by themselves nowadays. Did some looking on eBay, but the stuff they had there is too expensive.

Monday, 30 June 2003	Echunga
		Top of page
		previous day
		next day
		last day

Still more proof-reading the updates to the book today, which kept me pretty busy. It's days like today that make me wonder if I'm achieving anything at all. To make up for it, spent a bit of time tidying up my web pages. Once I had the updated documents completed, put together a new errata page and sent a message about it out to the FreeBSD-questions mailing list. That may have been a mistake; got a number of comments back about that too, and now I have to decide whether I want to incorporate them or not.