Greg's diary--July 2002

Monday, 1 July 2002	Echunga
		Top of page
		next day
		last day

Quiet day. After all the excitement of the last week, I couldn't decide which of the remaining things I had to do was most important, so didn't do any of them and went off riding in the forest again.

Tuesday, 2 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Yet another of those days at the end of which I had the feeling that I hadn't done anything, even though I was rotating all day. Arranging the AUUG 2002 conference has certainly kept me pretty busy, and other AUUG things are an issue as well. How do we get our message to people?

Wednesday, 3 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Still more work on the AUUG 2002 conference. This could be a full-time occupation. Also did some more work on getting the latest version of FreeBSD-CURRENT working, which now contains KSE. Slow work.

Thursday, 4 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally got past most of the AUUG 2002 organizational work, and turned to the next item... my paper for the conference, due by the weekend. It's interesting looking back at the last ten years, and how much has changed since then.

Friday, 5 July 2002	Echunga	Images for 5 July 2002
		Top of page
		previous day
		next day
		last day

Still more AUUG work today. I hope it's not going to stay like this. Somehow I need to find time for my real job. I haven't done any real work in over a week now. Made fair progress on my paper, which is due in tomorrow.

I did find some time to clean out nearly all the remaining boxes in the shed, though. They've been there for almost exactly 5 years, and still I found things I needed, like the oscilloscope probes that Geoffrey Bennett and I were looking for a year ago. I suppose it would be fitting to get Michael to come in and start tidying up the remaining mess on 17 July, the fifth anniversary of moving in.

Saturday, 6 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

More work on the AUUG paper today, which kept me going for quite a while. It's funny that I don't seem to be getting anywhere, but it's tiring nonetheless.

To Mark and Essey Deayton's in the evening for dinner. Essey went to quite a bit of trouble to make an authentic Mexican dinner. Spent most of the time after that playing with Birdie, their pet Galah.

Sunday, 7 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Spent most of the day working on my AUUG paper, and got it just about complete. It's really difficult trying to keep years of development in perspective.

With Yvonne's help, spent some time rearranging my bookshelves. After five years, it's beginning to look almost tidy. Found a number of redbacks in the magazines we brought in from the shed. I hope we got them all.

Monday, 8 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Still more AUUG stuff today, mainly getting my paper ready. Thank God that's over and done with, and only two days late.

Also spent some time looking at Gordon Hubbard's redesign of the AUUG home page, a great advantage over the previous design. Now if we could only decide whether “Open Systems” is an obsolete term which should be removed, or whether it adds value to AUUG's mission statement. I fear that might take a while.

The deadline I had set Davey had expired on Friday, so today I sent them another message threatening the involvement of the Office of Business and Consumer Affairs. As in the past, that got a pretty sudden reaction. In the evening I got an Email from them, unfortunately missing the point in nearly every instance. He did, however, state that the pumps have no protection against any pressure beyond their operating range. Looks like we're going to have fun. Interesting to note that Davey's mail server identifies itself as (AIX4.2/UCB 8.7/8.7).

Tuesday, 9 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Yet more AUUG stuff today. At least I got it finished, but found I couldn't think of anything else to do, so ended up reformatting the AUUG Membership Handbook, which was written in Quark XPress (or some name like that), converted to Microsoft Word and last updated 5 years ago. It was surprisingly easy to convert to troff.

Phone call from Davey to arrange an inspection on Friday. The name of the Office of Business and Consumer Affairs really works wonders.

Wednesday, 10 July 2002	Echunga	Images for 10 July 2002
		Top of page
		previous day
		next day
		last day

Despite all good intentions, found myself tied up with more AUUG stuff today. Still, it gave me a bit of time to catch up on the ports and the chapter on burning CDs for my book. Since I wrote the draft chapter about 9 months ago, they've changed the port, and the new one is called sysutils/cdrtools. It includes both the SCSI cdrecord program and mkisofs, but not the corresponding program for IDE CDs. Discovered that cdrtools doesn't install mkisofs unless you set the MKISOFS variable while running make. That's not documented anywhere except in the Makefile, which I think is totally inadequate.

In the evening to the AUUG SA meeting, which today included the elections. Before that had a chat with David Bullock, the new Secretary on the AUUG Board of Directors, and did a sort of handover, somewhat hampered by the fact that I had forgotten the power supply to my laptop Yet Again, and now both batteries of this Dell Inspiron 7500 are stone dead after fractionally over 2 years, despite careful treatment, so the machine was useless. Grrr.

The meeting itself was a little unusual. David Lloyd brought in an NCD “thin client”, which I discovered is another name for “X Terminal”, and demonstrated something, but suffered somewhat from lack of preparation. He did, however, lend me an IDE CD-ROM drive, for which he shall be remembered in the CFBSDIV acknowledgments.

Thursday, 11 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Wayne Smallacombe from Davey came along this morning, along with a colleague, and inspected the plumbing. The results were quite interesting:

There's nothing obviously wrong with the installation, though they would have preferred to see flexible connections from the pump. This is one of the issues that I mentioned to Robert Young on Boxing Day, and which he said didn't make any difference.
It's very unlikely that the pumps burst because of back pressure from the water heater. We discovered a second non-return valve on the supply line to the heater.
They suspect that water hammer might have played a role; there's a small indication of that, but I think it only started when we installed the new pump. Anyway, it goes along with what I suspected at the time.
A pressure cell would have been a good idea to reduce the shock on the pump housing. I had suggested one when I first bought the pumps, but they said I wouldn't need it.
These pumps don't run at constant pressure. They run all the time as long as any water is being taken off from them.

Wayne's going to send me a pressure gauge with maximum indicator, which will show whether we really do have any back pressure. I'd be surprised, but it's certainly worth checking. The new pump should be less susceptible to any problems. He left me data sheets-the first time ever that I have seen them-which state interesting things such as that the current pump has a maximum casing pressure of 1 MPa, and that the old pumps had a non-adjustable cut-in pressure of 140 kPa. And I was sold these pumps to run a machine which had a minimum pressure requirement of 250 kPa. Clearly somebody has done something wrong. But who? Certainly if Davey had responded to my initial request to find the cause of the problem, we wouldn't have had the follow-up problems.

The dropouts in the new pump appear to be something that they know about, though Wayne was cagey about saying so in those words. Anyway, it's not what Paul Johnson said:

Pump: HS60-08
Symptoms: Does not always start, intermittent fault.  Switch power off and on to
restart pump.  The HS60-08 is not starting under pressure, this can be the
result of either a low voltage situation (brown out) and / or an over compressed
mechanical seal.  In a low voltage situation the motor does not have enough
starting torque to overcome any resistance and will not start.  The other
scenario I mentioned is the mechanical seal is over compressed and this also
creates excess resistance and once again the motor does not have enough starting
torque to overcome the resistance and does not start.The excess compression of
the mechanical seal is caused by a pressure increase in the downstream pipe work
back pressuring the mechanical seal.  When you turn on a tap the pressure in the
downstream pipe work lowers and reduces the force in the front housing and
around the mechanical seal and the pump is able to start.To overcome this
situation we recommend upgrading the mechanical seal to a low friction Bergann
type 21.

He had obviously written this without reading what I said, since the problems happened with lack of pressure, and there were no power issues. What is interesting, however, is that it's possible that this sort of thing would happen when it's very cold. I recall it was very cold the first time round, and it's quite possible that was the case the second time as well. Wayne described a scenario where a switch can stick in cold weather. So they changed the control unit, and they'll send the old one back for testing-at normal temperatures. No wonder their test department never finds fault with the things returned.

Playing around with CD-R burners, but not very much. Finding your way through the Ports Collection is non-trivial.

Friday, 12 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally time to do a bit of other stuff today, and decided it would be a good idea to work on my Ports collection article for next month (which is why the link won't work until then).

The link (http://ezine.daemonnews.org/200208/ports.html) doesn't work at all any more. Daemon News still exists, but the Daeomon's Advocate pages seem to have disappeared.

A number of new web browsers have sprung up while I've been doing other things, and some of them seem better than Netscape (which isn't hard to do). Started installing some and writing an article.

Also did some work on the book, in particular the burning CD-ROM chapter. I had borrowed an ATA burner from David Lloyd, but it didn't seem to work too well. Made my first-ever coaster. OK, it wasn't quite a coaster, but it had a large number of I/O errors, so it's basically useless. Still, the drive served its purpose, and I now have a section in the book on ATA burners.

Saturday, 13 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

More work on the web browser article today. It's surprising how complicated this can be; hopefully the article will be really useful.

Also spent more time cleaning out the shed. I thought I had removed everything a week or so ago, but today I found about 5 more cartons and moved most of them into the house. Looks like my office will still be full despite all the new bookcases.

Also updated echunga to the latest -STABLE version. For once, things went smoothly, though I still can't make friends with mergemaster. We need to improve on the situation.

Sunday, 14 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Got in this morning and discovered that I had received almost no mail. On further investigation it appeared that echunga was suffering from the dreaded microuptime() went backwards syndrome. The kernel I had built yesterday was GENERIC, and we still have this bug in there which bites Athlon processors. It's been there at least since I got the system which is now echunga, a good two years. Built a new kernel without apm, and all went well, but it's high time to find and fix the bug.

Apart from that, a quiet day. Spent some time tidying up my office, which was sorely in need of it, and managed to get all the connecting cables behind my desk off the floor, so now the cleaners can clean it. Only five years since we moved in...

Monday, 15 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Some days I seem to get nowhere, just administrative overhead. Today I did manage to find time to upgrade the LEMIS web site, which posed some interesting problems, noticeably the bars on the left and right, which should remain the same across a number of pages. Decided against frames, and server-side includes seemed a bit dicey, so in the end decided on using soelim to include them in the install phase, using make. As an advantage, things even (roughly) work under lynx.

In the evening, the old Pioneer air conditioner failed, just stopped heating. I suppose it might just be a loss of pressure, but based on the past experience with this air conditioner, we're expecting a rough ride until it's repaired. Spent some time connecting up electric heaters, most of which were malfunctioning, and which blew fuses and threatened the satellite receiver, and finally got the heating working again. Not a happy evening.

Tuesday, 16 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

As if the failure of the air conditioner last night wasn't enough, when I got into the office this morning I discovered that our web site, hosted by Safeport, was off the net. Did a bit of investigation, and discovered that all of Safeport was off the net, which suggests a network outage rather than a system failure. Spent some time investigating the problem, but given the time on the US East Coast, it seemed unlikely that the problem would be solved before their morning, so moved the web site back to echunga, which went pretty easily. I suppose it's worth shortening the TTL on the DNS records for www.lemis.com.

Still more boring work. I really must get round to doing more interesting stuff.

Phil from Barker Air Conditioning came in this evening to look at the air conditioner. Found a burnt-out relay caused by incorrect repairs years ago: the relay had a rating of 24 A across three gangs, but one had obviously burnt out in the disasters of early 1999, and so the 5 HP compressor had been connected by a single gang (presumably rated at 8A). I'm surprised it held out that long. We need to look after Phil; other air conditioning people seem to have a lead time of up to 6 weeks, and Airtech themselves didn't even want to look at it.

Wednesday, 17 July 2002	Echunga	Images for 17 July 2002
		Top of page
		previous day
		next day
		last day

Five years today since we moved to Echunga! In honour of the day started cleaning out the shed, which has been accumulating stuff since then, and got about halfway to where we wanted to be. Here are some photos of before and after.

www.lemis.com was back when I got up. As I suspected, the systems hadn't gone down:

=== grog@newharbor (/dev/ttyp0) ~/ 9 -> uptime
11:33PM  up 155 days,  3:53, 1 user, load averages: 0.05, 0.02, 0.00

The HTTP access logs show a different story, though:

203.173.160.126 - - [15/Jul/2002:15:04:58 -0400] "GET /grog/slashdot/topicsecurity.gif HTTP/1.1" 200 3034
203.173.160.126 - - [15/Jul/2002:15:04:58 -0400] "GET /grog/slashdot/ HTTP/1.1" 200 39311
203.173.160.126 - - [15/Jul/2002:15:05:00 -0400] "GET /grog/slashdot/slc.gif HTTP/1.1" 200 139
66.77.73.204 - - [16/Jul/2002:10:47:28 -0400] "GET /~grog HTTP/1.0" 301 301
66.77.73.204 - - [16/Jul/2002:10:48:29 -0400] "GET /vinum.html HTTP/1.0" 200 387
66.77.73.204 - - [16/Jul/2002:10:49:34 -0400] "GET /handbook/what-to-download.html HTTP/1.0" 404 294

Apart from that, spent some time installing OpenBSD on an old prototype Dell laptop, gateway. For some reason I wasn't able to get the machine to boot from hard disk, but I did manage to boot via a floppy. There are still a number of demons in that area.

Thursday, 18 July 2002	Echunga	Images for 18 July 2002
		Top of page
		previous day
		next day
		last day

Trying to work on my book again today—it's high time. Somehow there are just too many interruptions. Maybe I should designate a couple of days a week where I don't do any normal communication, and just put my head down and work.

Did get some other work done today, mainly on FreeBSD-CURRENT. Unfortunately, most of it was wasted on chasing down a recently introduced bug.

Friday, 19 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Spent some time looking at the panics I had had on zaphod yesterday, and building a new world. With the WITNESS options, the system runs at a snail's pace: even building a kernel (with modules) took 80 minutes. It looks as if the panics could conceivably be a yet-undiscovered bug in Vinum, but it's so glacially slow to find out.

Finally got some time to work on my book. The last edition had almost nothing about ssh, and after scouring google, came to the conclusion that there's a serious lack of tutorial introductions to the more esoteric ssh functionality, such as tunneling. Hopefully this chapter will fill the gap.

google has an annoying habit: they steadfastly insist that I'm in Germany and reconnect me to www.google.de. The web page comes up in German, but at least they accept when I change the language to English.

Saturday, 20 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

As if things weren't dragging enough already, today my almost new (well, 16 month old) sources drive died on me, a 40 GB Western Digital drive that I bought in Singapore on 20 March 2001. Fortunately it chose a good time, for a change: I just did a complete backup of the drive yesterday, before it failed. Tried various things, including finding the source of a new drive, but being Saturday didn't get far. It did manage to stop me from doing any useful work done, though.

Sunday, 21 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

The problems with the failed disk drive wouldn't leave me alone today, and I decided to build a new machine with 14 4.2 GB drives from my Storage Array. That was a lot more complicated than I thought. I was able to install FreeBSD 4.5 pretty easily (the ISO copy for 4.6 was on the dead disk), but setting up Vinum was not. I got multiple panics when trying to configure, something that others have reported to me but which I haven't previously been able to reproduce. Things weren't made any easier by the fact that I was using a GENERIC kernel off the CD-ROM, which didn't have debug symbols, so I couldn't analyse the dumps. Also, not having the Ports Collection on line, I didn't have my normal tools available. Spent most of the day investigating that and trying to get enough infrastructure to test the bugs. It seems that there are at least three issues:

The CPU fan is flaky, and it stopped at least once, causing the CPU to overheat and panic the system.
I seem to be getting random rejections of disk labels, and I wonder if there are race conditions in the sym driver code. I have a total of four Symbios host adaptors in the machine. I know there was a problem in the older ncr for the same driver, but it was more obvious: it paniced as soon as more than one host adaptor was in the system.
There's probably still a bug in Vinum: a number of the panics happened after a configuration error. I need to find those ones.

Monday, 22 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

If yesterday was a bad day, it paled by comparison with today. It started when I looked at the tape which had been ejected from echunga's tape drive. It was the tape onto which I had managed to back up my source drive on Saturday. A quick check revealed that I had managed to overwrite it with last night's backup.

Further checks revealed that the previous backup was illegible. In fact, it turned out that all backups were illegible: my Sony DDS-4 drive had failed in a way which made it return read errors after successfully reading. Spent the rest of the day looking for ways to solve the issue, discovering in the process that I could read some data from the failed disk, and that it was still under warranty, despite having been bought in Singapore. Also got confirmation that nobody knew a way to position a DDS tape past the double tape mark at the end. Somehow didn't get much else done.

Tuesday, 23 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Enough of tape backups

Topic: technology

Link here

Finally managed to get my hardware orders in today, and by late morning Yvonne returned with a total of a quarter of a terabyte of disk storage: a 20 GB “Fujistu” 2½" drive for an unrelated purpose, and three Western Digital WD800BB-18CAA0 drives, ostensibly 80 GB, but dmesg tells me they're really 76293 MB. It seems that 80 GB is currently the lowest price per byte. One disk is destined to replace the old 40 GB drive, and the other two are to replace the unreliable and slow tape drives.

Got the stuff installed pretty quickly, and I was quite amazed to see the speeds at which I could perform backups, up to 10 MB a second of compressed data, and that across the network. Spent some time working out some dump scripts, and finished the day relatively happy with the results.

Wednesday, 24 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Up early this morning with some new ideas about backups. One of the most obvious things about making backups onto file systems is that the average file size is much larger. Normal files have an average size of less than 4 kB; the files on my backup disks had an average size of about 1 GB. That means that you don't need nearly as many inodes, and it makes sense to have a larger block size.

Accordingly copied my backups from one disk to another, via NFS. To my astonishment I got approximately 10 MB/s transfer rate, coming close to what the (100 Mb/s) network can handle. Then did a newfs on the disk, using 32 kB blocks and 4 kB frags, and with one inode per 32 MB instead of the standard one inode per 8 kB. The results were significant:

Filesystem    1048576-blocks     Used    Avail Capacity  Mounted on
echunga:/dump          75100    18714    50377    27%    /dump
/dev/ad2h              76285        0    70182     0%    /dumpa

That may not look like much, but the difference is 1.185 GB, more than all the storage on any system I had used before 1985.

At the same time set soft updates, though it's unlikely that this will make any difference. After that, spent some time tuning the backup scripts. It's surprising that there aren't more of them out there, but maybe that's the reason for the popularity of things like amanda.

Later installed the 20 GB laptop disk for Sue Blake, and put a surprising amount of data onto it. Also investigated how to get Java to work with Mozilla, not the sort of thing that most people would want to get involved with. To start with, I had to download about 70 MB of sources, which Sun insisted I do manually (so that I could accept their license, which didn't display anything of use), and building took for ever. At the end, I discovered I wasn't done yet and needed to download Javascript in order to display the pages I was looking at. This is definitely not the way to do it.

Thursday, 25 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

To the ADUUG lunch in Adelaide today. We discussed how Simon Hackett has been able to gain the attention of the press (there was another large photo of him in The Australian's IT section on Tuesday). Seems that he has a method that we can use too. Also spoke with Phil Kernick, who runs the SAGE-AU SA chapter, and came up with a potential new source of speakers who would be suitable for the AUUG-SA meetings as well.

After that off to have coffee with Dan Shearer. That bloke is always full of ideas, some of them bright. Left with a lot to think about.

Friday, 26 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally got around to doing some work on my articles today, and not too early. Though I started on the article on web browsers nearly two weeks ago, it has become a can of worms. The bottom line is probably that Mozilla isn't worth the trouble yet.

Also did some work on the instant-workstation port. Rasterman recently gave an interview in which he gave the following answer to the question “Where do you think the future lies for desktop Linux?”:

At the time I specified a URL, http://www.linuxandmain.com/modules.php?name=News&file=article&sid=141, but of course that's gone, along with its domain. This appears to be a copy.

R: Not on the desktop. Not on the PC. Not on anything that resembles what you call the desktop. Windows has won. Face it. The market is not driven by a technically superior kernel, or an OS that avoids its crashes a few times a day. Users don't (mostly) care. They just reboot and get on with it. They want apps. If the apps they want and like aren't there, it's a lose-lose. Windows has the apps. Linux does not. Its life on the desktop is limited to nice [sic] areas (video production, though Mac is very strong and with a UNIX core now will probably end up ruling the roost).

I don't agree with Raster on that statement, but he has a point that it's just too complicated for the average user to get a usable desktop on UNIX. instant-workstation was supposed to address that issue, but it has some problems which I need to iron out, and since I had this disk I need to send to Sue Blake, I decided to do a virgin install of XFree86-4 and instant-workstation on it. One thing's clear: instant is a misnomer if you want to build things from scratch. It took most of the day to compile on a 600 MHz laptop, but by the evening it was done, though I fear that some of the stuff that got installed is broken.

While talking to Liz on the phone in the afternoon, watched idly while a little pied cormorant, a relatively frequent visitor, plunged into the pond outside my office and came up with what looked like a big carrot in his mouth. On closer observation, it was a goldfish, one of 6 that we had put out in the pond last summer, and the first confirmation that we still had real goldfish in there (I saw a silvery one some time ago, but it died anyway). The cormorant could hardly get it in its beak, but somehow managed to fly off without landing flat on its beak while Yvonne and I called obscenities after it.

A couple of hours later Yvonne discovered that the cormorant had not been able to do much with the fish, and had left it lying on the ground about 20 metres further on. It looked pretty dead, but still twitched its tail when I picked it up, so I put it back into the pond and noticed with surprise how it gradually came to life over the space of a couple of hours.

Saturday, 27 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

More work on the web browsers today. I think I've bitten off more than I can chew. Mozilla really isn't ready for prime time yet, and both it and Opera have a number of strangenesses which make them unsuitable for real work.

Also spent more time working on the instant-workstation, which really doesn't install as well as it should. For some obscure reason a make clean and reinstall don't work cleanly. It's not directly a problem with instant-workstation, which really doesn't do very much. The Ports are just getting too complicated, and the documentation leaves a lot to be desired.

Finally got sound working on echunga, and started looking at ripping CDs. Installed grip, which greatly surprised me by showing the names of the tracks on the CD. They must be stored there, something I didn't know before. That makes it quite a useful CD player.

Unfortunately, ripping didn't work nearly as well: grip doesn't rip itself, it just interfaces to a RIP program, and by default it doesn't install one. There's no information anywhere I can find about how to do so, and the port doesn't even install all of what documentation there is. Spent a couple of hours trying to run dagrab, which didn't work at all, and tosha, which only worked when not in conjunction with grip. Couldn't find cdda2wav, which turned out to be hidden in cdrtools. We really need to do a lot to make this stuff work well.

Sunday, 28 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

End of the month is always a busy time, since I have at least one article to submit. Today it was two, so I ignored both of them and carried on trying to get my audio stuff working instead.

There's no doubt about it, grip is buggy. I installed exactly according to the documentation, and it was unable to rip correctly with either dagrab, tosha or cdda2wav. Then I recalled that it had a potted configuration for cdda2wav, and that worked. Spent the rest of the afternoon installing mpeg players, which are at least as much of a can of worms.

Writing things in this diary is useful. I got a mail message from Michael Pearson describing a misassumption I had made about grip yesterday:

> Finally got sound working on echunga, and started looking at ripping
> CDs.  Installed grip, which greatly surprised me by showing the names of
> the tracks on the CD.  They must be stored there, something I didn't know
> before.  That makes it quite a useful CD player.

Nope: it generates a CDDB id and looks the record up at freedb.freedb.org.
The information returned saves a lot of time, but is often poorly typed,
categorised wrongly, etc.  It's a good idea to go over what it is
displayed before ripping the CD.  Also, grip doesn't behave well if the CDDB
lookup can't be performed, or the CD is ejected mid-lookup, etc.

Well, of course I had disconnected the Internet downlink and tried again before I wrote that, but I had forgotten that this was on echunga, the machine which runs the PPP link, so it went via that link and not the satellite link. Another test with the PPP link disabled confirmed what Michael said. I'm surprised, though, that so far I haven't found a single CD which isn't in the data base, neither the very new ones nor the very old ones.

Monday, 29 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally got down to finishing off my articles today, and managed to finish the article on web browsers, sort of. It's a real can of worms, and at the end I was left wondering if I was correct in what I said.

After that, working on the other article, the Dæmon's Advocate. Rasterman recently gave an interview in which he stated that Linux on the desktop was dead. I disagree, and that's the subject of my article. Obviously you can read “BSD” instead of “Linux” and there will be no basic difference. In the process, decided to install Microsoft Windows 98% on one of my machines with an 8 GB IDE disk. Yes, I know it's not the latest, but it was forced on me on one of my laptops, and I have no intention of buying Microsoft. The results were underwhelming:

Boot from CD-ROM. You get the choice to (continue to) boot from CD-ROM or hard disk. Obviously you need the CD. It offers to install a Microsoft partition on the disk. After that, it reboots.
Boot from CD-ROM, the second. Again you get the choice to (continue) to boot from CD-ROM or hard disk. It suggests hard disk. If you choose that, it hangs.
At this point a FreeBSD boot CD indicates that so far the install program has created a single primary partition over the entire disk.
Boot from CD-ROM, the third. The install program goes off and “formats” the disk, really only creating a file system. On a disk of this size, FreeBSD would take about 40 seconds. Microsoft takes 15 minutes. After that, it runs scandisk in case errors have occurred since the format, runs a “setup wizard” and copies data for 25 minutes at speeds which give the CD-ROM drive time to spin down between accesses. Then it reboots.
Boot from CD-ROM, the fourth. This time I decided to boot from hard disk, and it proved to be the correct choice. Finally it asked for the “Product Key”. I typed it in, and it was rejected because I had entered the hyphens which are embedded in the key. OK, it tells you not to do that, but then it should just not accept them at all. It proved impossible to blank the key field in one action: you have to mark each of the five fields individually and then delete them. Finally I had the key entered, and it rejected it again, though it looked right.
The problem is, I have two Windows CD-ROMs, and I have two manuals with a Product Key written in a font where 8 looks like B. Which CD-ROM belongs to which manual? There's nothing on them to indicate which, but they know. What a pain. I entered the other Product Key, and it worked fine, but it wasted 5 minutes. If I had lost the other manual, I might have ended up, after nearly an hour, with a disk I couldn't do anything with.

Then it started doing something else, and I left it for a minute or so. When I came back, it was in the process of rebooting.
Boot from CD-ROM, the fifth. It carried on doing stuff by itself, congratulating itself all the time about what a good operating system it was. I left it for a while again. When I came back, I think it was just finishing the process of rebooting.
Boot from CD-ROM, the sixth. This time, it found “new hardware”, the monitor I had been using all along. It performed some kind of setup, but at the end of it all it stayed at a 640x480 resolution. Then it decided it was finished, after 1 hour and 8 minutes.
I then tried to change the monitor resolution and “install” the Ethernet card. Neither was successful: it couldn't find the card, and it was convinced that I had a generic VGA. At this point I gave up.

By comparison, installing FreeBSD on this box takes about 20 minutes. I need to manually set up the X configuration, but that's only a couple of lines. It reboots exactly once in order to run from disk, and it recognizes all the hardware. Clearly the acceptance problems with FreeBSD have nothing to do with the installation process: the real problem is just getting a useful “desktop” up and running.

Tuesday, 30 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

More work on my article today, which of course required some work in order to confirm what I want to do. It would be really nice to get the system to a point where it's as usable as a correctly installed Microsoft system, but we currently have a large number of bogons in the Ports Collection.

In addition, things went crazy on the AUUG front. Today we exchanged 44 mail messages. In all of September 2000, it was 38. It's nice that we're finally getting things done, but I hope we don't have this much activity all the time.

In addition, we had a lot of work on the web site, which is beginning to show its age: it's a 166 MHz Pentium with 32 MB of memory, admirably suited to demonstrate how bloated python and mailman are: with some mailing list distributions we were swapping for all the disk could manage. It's beginning to look like we need a bigger machine.

Wednesday, 31 July 2002	Echunga
		Top of page
		previous day
		next day
		last day

And I had hoped that things would cool down yesterday; instead, they got worse:

On 31 July 2002 you received 1530 mail messages and sent 75 mail messages.

It's the sent messages that count, of course. I've seldom sent so much mail.

As if the workload wasn't enough, I'm planning to go to Melbourne on Friday for the inaugural meeting of the new AUUG Board of Directors, and Qantas staff decided that it would be a good idea to strike on that day. Spent quite a bit of time trying to arrange things, not helped by the fact that Qantas decided to cancel my flight and suggested that I should fly tomorrow. They weren't prepared to pay for my overnight accommodation, though, nor even refund my ticket if the flight were cancelled and I had to drive. A pretty poor show.