Friday, 1 June 2001	Echunga
		Top of page
		next day
		last day

Winter's here already! That wasn't the worst thing about the day, however; in fact, the weather was quite nice.

The tilers came in this morning as threatened and took up the entire floor of the dining room. We're locked out of half the house until Tuesday or Wednesday.

Into Meadows, Yana driving, to look at a trailer Yvonne thought might be worth having. A complete rust heap. Took Yana driving a bit on the way back; she still needs a lot of practice.

In the afternoon restarted some work on Vinum for root file system support, which I had working a year ago, but people hadn't liked the way I did it. Didn't get very far today.

Wanted to take a photo of the floor of the dining room once it had been cleaned up. My Nikon CoolPix 880 digital camera has died, only 7 months old. “System Error”, apparently because it no longer knows how to extend the lens. Called up the agents and discovered that they would not honour the guarantee. I find that remarkably unprofessional of a company like Nikon. People don't just live in one place, and the quality of their products shouldn't depend on the fact. Grrr.

For what it's worth, the camera wasn't dead, just pining. It was finally retired, still functional but out of date, on 5 November 2004. See the entry for tomorrow.

Saturday, 2 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Nobody loves Orlando

Topic: general, animals

Link here

Somehow mail is overwhelming me again. Maybe it's just all the things we have to do. I had intended to go riding today, but I woke up feeling too tired. Orlando is back with us again, and he doesn't seem to get on with any of the horses any more. We tried putting Darah into a stall next to him, but she reacted so violently we thought she might hurt herself. Poor Orlando's all by himself while we decide what to do.

No new furniture

Topic: general

Link here

Into town in the afternoon looking for lounge room furniture. Yvonne knew just what she wanted, but by the end of the effort we were completely indecisive. What a waste of an afternoon.

Camera not dead, just pining

Topic: photography

Link here

I've been doing a lot of thinking (and grumbling) about the failure of my Nikon CoolPix 880 digital camera. The instructions, and the repair people, were pretty clear: “It's dead, Jim”. But there was one thing that got me: I had accidentally turned the power on (far too easy on this camera; an accidental nudge will do it), and a bug in the power management causes it not to work when the lens cap is on; instead, it displays “please remove lens cap” (after which you must also power cycle the camera), and seems that it keeps trying to extend the lens anyway. There's obviously a bug in the microcode in this area. Anyway, I considered looking at doing a complete reset to factory conditions to see if that would help, but I didn't need to go so far: I replaced the battery (after about 24 hours), and it worked. My best bet is that the system gave up trying to extend the lens and marked it as defective. After a lot of power down, that information might have got lost.

There's something basically wrong here. I'm glad the camera is working again, but this problem and its solution should have been more obvious.

Sunday, 3 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Quiet day today. Round lunch took Yana driving again, and we managed to do about 60 km, going round Kuitpo Forest, Strathalbyn and Mount Barker. She should have plenty of experience by the time she's ready for her P plates.

Later in the afternoon decided to go riding with Darah and La Tigre. Yvonne wanted to clean them on arrival in Kuitpo Forest, but I preferred to do so before leaving. Good thing too: La Tigre had got a kick on the hock, and was lame. We had to take Pebbles instead. That was enough fun in itself: she's not used to holding back. Still, Yvonne got her under control, and Darah behaved very well. I need to ride more often.

Monday, 4 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Somehow the work doesn't slow down, and I'm still getting nothing done. It doesn't help that I have to accompany Yana to and from school; Yvonne is scared to do it. I suppose I'll have to carry on for at least the rest of this week.

Vet in today to look at La Tigre's hock. It could be serious, possibly a hairline fracture. Yvonne had to take her off to Greg Rodda's clinic in Prospect Hill for X-rays. Yvonne's on tenterhooks, of course.

Tuesday, 5 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

A quiet day. Finally the fires have stopped burning, and I'm trying to find out which of the less urgent things I should do next.

La Tigre doesn't have any broken bones, but she does have a torn tendon. They're keeping her at the clinic for a few days, and it looks like she won't be rideable for at least a month. Looks like Pebbles will get more action.

Wednesday, 6 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Somehow I have run out of steam, and domestic affairs are taking over. Spent some time today shopping for a new TV (the old one didn't look too happy any more, and it's over 10 years old). Decided to buy one with an 80cm tube, conventional analogue: the digitals and 16x9 are still pretty rare round here, and the prices are corresponding.

The decision was between Sharp and LG devices. On the face of it, the LG TV looked more interesting, until I saw the instruction manual. The description didn't even match the hardware: for example, it described a remote control (shown only as a diagram) which looked nothing like the one supplied. It also didn't include technical specs, so we couldn't find out whether it supported SECAM-L (the Sharp does, at least according to the manual). So we bought the Sharp.

After getting the TV home and putting it in place—mercifully, the tilers have finished—discovered that the Sharp had its problems too. The manual describes a procedure to adjust the deflection to adapt to the effects of terrestrial magnetism, but the button that is supposed to do that (with the likely name “Preset”) didn't cause any effect at all. It did on an older one that they had in the showroom. I am suspicious of the fact that this appears to be the only function of the button, yet it required to be pushed four times to select the correct page. I suspect that they have changed the firmware and that this function (like many before it) is no longer necessary, but they didn't update the manual. Anyway, they're sending somebody along, and there's a good chance I'll get a new TV. I'll be interested to see whether the button is defective, or whether it's a documentation bug.

Thursday, 7 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

More running around today. Somehow this week has been unproductive. Maybe it has something to do with taking Yana to school; that's two hours a day. Today also had to take Yvonne to Bridgewater to have the Commodore fitted with an electric aerial: people keep breaking off manual ones.

Finally got round to doing my flight bookings to the USA. Takes a bit of time to get used to.

Friday, 8 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally Yvonne is prepared to go driving to school with Yana, so I should have got more work done today. Did spend a bit of time reading about EVMS, which is gradually fitting together.

The TV repairman showed up to look at the non-functional geography correction. He had never seen the feature before, but was able to find the problem: the “Preset” button only works when in reception mode (tuned to a channel), not in monitor mode (external AV input). That makes a bit more sense when you discover that the other functions of this button relate to channel tuning. But they overloaded it with geometry correction, when it should have been an on-screen menu option, and they didn't document the restriction. I wonder how many service calls that has created.

Saturday, 9 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Every 10 days the connection to www.lemis.com seems to die somewhere in Palo Alto. It happened on 20 May, 30 May and again today. Spent most of the afternoon tracerouting. They removed the problem, but we still don't know what's going on. It's particularly surprising that the problem affects 192.109.197.0/25 differently from 192.109.197.128/25. I wonder what's causing it, and whether it'll happen again on 19 June.

Sunday, 10 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Gradually I'm getting to do things that are not directly work-related, probably a good thing to do on a weekend. The problem is, of course, that I don't really dare leave my mail alone for a day, or I'll never catch up again. Apart from that, though, I did manage to take a look outside the house and watch the rain. Since building the extension the garden has been excessively soggy in the winter, since the extension blocks the natural slope over which the water used to drain. Dug a drain on the East side of the extension, which seems to have helped a bit.

In the evening working with Yana on her web pages. It's not immediately apparent that renaming large files here is a bad thing to do, since they get rsynced from echunga to www.lemis.com, and rsync doesn't understand renaming. Spent quite a bit of time doing it manually.

Somehow the entries for 11 and 12 June have got lost. I didn't discover until much later. If you have a copy, please let me know.

Wednesday, 13 June 2001	Echunga	Images for 13 June 2001
		Top of page
		previous day
		next day
		last day

Another day with seemingly little achieved. Finally got my flight to Canberra booked, not before time considering that I had to go tomorrow. Despite multiple reminders never got an itinerary, and had to call Ansett to confirm that I really had a booking. Also tried to join the Ansett Golden Wing Club, but didn't get any confirmation about that either.

Thursday, 14 June 2001	Echunga –> Canberra
		Top of page
		previous day
		next day
		last day

Up in the middle of the night to fly to Canberra. Bumped into Alan Modra immediately on arrival at the airport, where we were informed that we were delayed by what proved to be 50 minutes. At least we got into the Golden Wing Club, despite the fact that my application hadn't been processed.

In Canberra to the office, which bears a certain resemblance to the Linuxcare office: big and empty. They're still waiting for furniture. Anton showed us his Nighthawk, a big 16-way PPC machine with an unmistakable IBM identity. Inside it looked rather like a more modern Helix or S4000.

Met Greg Rodgers, who is leaving today. Spent some time with him trying to get Linux to boot on a Sphinx, which is the code name for the 44p /270 that I am getting. 4 SCSI busses, one of them differential, which will be great for the disk trays I have at home. For reference, the top SCSI connector is scsi@11, and that's probably what I should be booting from.

Pretty quickly to lunch with just about the entire OzLabs mob, also Rasterman and Horms, with whom I had intended to talk, but didn't get time.

Back at the office, unpacked my new laptop, a T22 Thinkpad with 1000 MHz processor, 256 MB memory, 32 GB disk and a 1400x1050 display; nothing to complain about. Spent some time installing FreeBSD 4.2 on it and discovered that the chipset wasn't supported. Was still in the process of downloading XFree86 4.1.0 when we had to go to dinner, a going-away dinner for Greg R. Good food, just round the corner from our hotel.

Friday, 15 June 2001	Canberra
		Top of page
		previous day
		next day
		last day

Walked to work with Alan, who was mightily impressed by the frost and my barefootedness despite. Finished installing XFree86, but there seem to be a couple of issues which we haven't been able to fix yet. In particular, it seems that the chipset is capable of driving two different displays with different contents, which is really nice for a laptop, but we haven't been able to establish how it works. Also spent a bit of time trying to find out various things about being an IBMer, and also tried to install some useful software on the Thinkpad so that I can access the internal network.

In the evening to dinner with Warren Toomey and discovered that the hotel is just round the corner from where we had the Christmas Party on 1 December 2000. Found another nice restaurant, the Santa Lucia, just a couple of doors from last night's place. Ate a risotto con caglie, one of the best quail preparations I've ever had. Early to bed.

Saturday, 16 June 2001	Canberra
		Top of page
		previous day
		next day
		last day

Up fairly briskly to the ANU for the AOSS, and got there fairly easily despite the taxi driver who wanted to drop me somewhere else. Warren Toomey was already there, Stephen Rothwell arrived soon after, and we got things started with relatively little difficulty.

The symposium went well; we got more attendees than we expected (though less than in Adelaide), I had less occasion to truncate people's talks than last time, and again most of the papers were by people directly involved in the project. I was left feeling “let's do this again”.

In the evening to dinner with everybody, but I was pretty tired and didn't last long.

Sunday, 17 June 2001	Canberra –> Echunga	Images for 17 June 2001
		Top of page
		previous day
		next day
		last day

Up and to the airport with no particular problems this morning, spending most of the time in the lounge with Alan trying to get wireless networking going. It didn't work. We were able to set up the config files to reflect the ID strings: the IBM card has different strings from the Orinoco, but I one of each, and each has the same FCC ID label on the back. The problem was that Alan's fresh Debian installation didn't include the module needed. I'm really puzzled by how often this seems to happen in Debian, even when you ask it to install everything.

Back home without too much trouble. Flying via Melbourne is more trouble than it is worth. I hope we can find better routing next time. Spent the flight talking to Alan about binutils cross-build environments.

In the afternoon, playing with my new toys and trying to get my 2700 mail messages under control. It's fun having new hardware.

Monday, 18 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Still in catch-up mode today. This week's going to be particularly problematic, since I'll be leaving for the US on Sunday. Did find enough time to take another crack at the binutils port, and got as far as requiring the Linux header files. More tomorrow maybe.

Tuesday, 19 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Did a lot of different things today: spent some time trying to unpack old packing cases from the shed, played a little more with binutils, tried with a remarkable lack of success to build an up-to-date FreeBSD system on my new laptop, canberra.

Why canberra? Well, the last machine, a Dell Inspiron 7500, was called sydney because that was the project name at Dell. I don't know the IBM project name for the Thinkpad T22, but I got it in Canberra, and the names seem to match. All the better if they don't rhyme.

The FreeBSD 5-CURRENT kernel died during the boot probes, so spent a lot of time building debug kernels, and found the problem (data freed then used, easy to fix) after a while. Also discovered that devfs apparently doesn't support APM at the moment: the devices don't get made, and there's currently no way to create a device node in devfs. Noted also that the Aviator cards didn't work.

Trying to build a 4-STABLE world was at least as frustrating: the build kept dying in a multitude of ways. Spent most of the rest of the day doing that, also preparing for my trip to the USA.

Wednesday, 20 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally got a 4-STABLE build done on canberra. And the ray driver didn't work; well, the card was recognized, but the driver allocation failed. We need more descriptive error messages, but I don't have time to play with this at the moment, so left it.

Spent a lot of time planning the CDs for the next AUUGN. We're planning to put in OpenBSD 2.9, and we'll have to make our own CDs. More work than I bargained for.

Thursday, 21 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

More preparations today. I wish I had an overview of just how much time each activity takes. I don't like organization, so I suspect that I try to play down the amount of time it takes, but somehow I was so busy that I didn't even get through my mail, which I normally complete early in the day. Well, anyway, it looks like I'll be ready by the time I leave.

Friday, 22 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

Catching up with 2500 mail messages today kept me busy for longer than I wanted. Spent the rest of the day catching up with a surprising number of loose ends. The good news is that it looks as if all will be well by the time I leave, which for reasons which weren't explained was put forward an hour, from 9:25 am to 8:20 am. Grrr.

Saturday, 23 June 2001	Echunga
		Top of page
		previous day
		next day
		last day

More travel preparations. I usually take it easy before a long journey, but this was ridiculous. Made up for it by submitting a paper for the AUUG annual conference along with a whole lot of mini-articles for AUUGN.

I need to get a life. Yvonne and friends went out riding today, but I didn't have time.

Sunday, 24 June 2001	Echunga –> Sydney –> Los Angeles –> Boston
		Top of page
		previous day
		next day
		last day

Up at 6 am, but it seemed like the middle of the night, and I was still yawning when I got to the airport. There, for once, I was able to authenticate with SkynetGlobal, but since Yvonne was there, didn't do much.

In Sydney, that was another matter, and managed to reply to a lot of mail, unfortunately hampered by the fact that SkynetGlobal have broken their DNS again, and neither the NAT proxy nor their mail server (smtp.skynetglobal.com) had reverse lookup, so I wasn't able to send half my mail.

Off to Los Angeles, where I got my connection to the MobileStar network in the American Airlines lounge, but that was as far as it went: authentication didn't work. I called up their rather reluctant support desk (“we'll call you back tomorrow”), who were also very well informed (“you can't open the gate?”. “No, I can't authenticate”). She finally put me through to somebody who did understand the concepts, but he said that SkynetGlobal wanted to handle their own support calls. At the time, of course, they were not available, since it was the middle of the night. Growling, wrote a demand for improvement, and spent an hour watching the air space and investigating the net. One other person came on line just before I had to leave. Obviously not a heavily used service, even this side of the Pacific.

On to Boston with no mishap, and got to bed just before midnight. Anybody who asks for 36 hour days has probably not experienced them.

Monday, 25 June 2001	Boston
		Top of page
		previous day
		next day
		last day

What a lousy night! Normally I don't have too much trouble sleeping after a long journey, but today was an exception, and I only got about 3 hours of troubled sleep. Up accordingly drowsily and to the first tutorial, about mobile networking. Quite interesting, but not as detailed as I had hoped.

Wireless networking is the way of life at USENIX. Got some quite nice connectivity, but we're still suffering from lack of reverse DNS. Ended up having to download all my rejected mail messages to Yana's machine firefly.lemis.com, the only machine we have which is still running sendmail, and running the queue from there.

A number of Europeans have showed up already, including Nik Clayton, Brian Somers, Paul Richards, Guido van Rooij, Arjan de Vet, and also Michael Wu. I don't know what they're doing, since they're not participating in the tutorials. Out to dinner with them in the evening before hitting the sack really early.

Tuesday, 26 June 2001	Boston
		Top of page
		previous day
		next day
		last day

Slept a whole lot better this night, and up in pretty good shape. Down to try to change my tutorial, but it had to be with one on the same day. Went to the network programming course, particularly with a view to having it held again at the AUUG annual conference. I've decided that USENIX tutorials aren't a good match for me any more. I didn't learn too much, and left early.

It was Brian Somers' 34th birthday today, and he had the idea of drinking one drink for each year of his life. He didn't succeed, but he made a valiant attempt. To an Indian restaurant for dinner; I've had some bad Indian food in the last year or so, but this one took the cake.

Wednesday, 27 June 2001	Boston
		Top of page
		previous day
		next day
		last day

Up bright and early to have breakfast with Henry Hall and Rasmus Lerdorf, and had a lot of trouble picking up my hired car (the pickup point didn't exist). Through dense traffic and building sites to Henry's place, arriving late, and discovered they had already had breakfast. Off to find my own, also had a haircut, then back and and spent a leisurely morning comparing notes.

In the afternoon, off cross-country to Framingham aided only by this emetic Hertz map, to buy a cheap Palm Pilot for Yvonne at BestBuy. Turned out they didn't have them any cheaper than elsewhere, in fact they didn't even have the model I was looking for (m100). Bought a Tivo instead, and then found a Palm Pilot on the other side of the road. Back to Sitara to talk to people, and discovered a crisis looming, so off to buy a book at Softpro. Back, found that the crisis was still in full swing, and so off back to the hotel, where everybody had left for dinner, so had dinner in the hotel and early to bed.

Thursday, 28 June 2001	Boston
		Top of page
		previous day
		next day
		last day

First day of the conference, started off by Dan Frye talking about IBM's position to Linux. On the whole not a bad presentation, though other people's reactions were interesting. Met Dan later on and had a brief talk.

Heard an interesting paper on a multi-queue replacement algorithm for multi-level disk caches, unfortunately marred by a lack of consideration for the relative speeds of the caches; it's possible it could even perform worse than simpler systems for some cache configurations. Also Yet Another RAID-5 paper which solves the wrong problems by spreading all accesses across as many spindles as possible. I think I need to write a paper about this.

In the evening to the BoFs, which were a little disappointing, especially as I missed the NetBSD BoF. Geoff Steckel and John Hood came along later. I need to get them more integrated in the FreeBSD community.

Friday, 29 June 2001	Boston
		Top of page
		previous day
		next day
		last day

Somehow this conference feels deflated. We were expecting the attendance to be down, but they tell us that it was only down by 7%. Subjectively it feels like a whole lot more. I suppose I'm looking at things with a different viewpoint now I'm helping to organize the AUUG annual conference.

SkynetGlobal still haven't replied to my mail messages about authentication in the USA, modulo a couple of contentless voice mail messages from their support manager, so called up MobileStar as threatened in order to sign up for their services. The bloke on the other end of the line told me what the SkynetGlobal support people didn't do in a week: to log in as grog (that's me) from SkynetGlobal, you need to type in the user ID grog@skg. They didn't tell me that when I asked specifically for instructions, and they have spent a week not replying to my messages. What a load of losers.

My paper this morning; it went across well, much better than the talk about Vinum 2 years ago. A matter of practice, I suppose.

In the afternoon to a press interview with Dr. Dobbs. I've been talking to Robert Watson about co-authoring the O'Reilly book, and just by chance bumped into him and Andy Oram in the press room. We need to follow up after Robert's wedding.

Did a bit of looking around for people for the AUUG conference, and found Evi Nemeth, who was coming anyway, with intent to get a tutorial out of her. Looks like we can do that. Rob Kolstad is also coming to Sydney, but I haven't seen him here.

The only other talk of any interest was about Linux Volume Managers, unfortunately so superficial as to be completely uninteresting.

In the evening to the conference reception. After experiencing this for the third time, I'm very happy that we (AUUG) have opted for table service. I couldn't get at the food for at least 20 minutes, and in the end Michael Wu went and got some for me. Strangely, the drinks had no queue at all. That wouldn't have happened in Oz.

Saturday, 30 June 2001	Boston	Images for 30 June 2001
		Top of page
		previous day
		next day
		last day

Up a little later today, since there wasn't much going on early. The FreeBSD people decided to have a “kernel summit” meeting today, which rather upset me, since there are other reasons to come to USENIX. Went to the beginning, just enough to mention a word or two about PPC, and then off to a talk about an NFS-based toolkit for userland-based file systems. I fear it won't be good enough to host Martin Pool's NFSv4 implementation.

Back to the meeting, found Jordan hand-holding a video camera, and went to get a tripod. By the time that was sorted out, I had to go to the next talk, about aspects of compressed file systems. Seems reasonable, but I don't know when I would want to use it.

Back to the meeting in time to go to lunch, where at least I got a chance to talk to David Greenman and get agreement on some aspects of SMPng. By the time we got back to the conference, there were more papers, for once interesting ones. The first, “Scalable Linux Scheduling”, should better have been titled “adapting BSD scheduling to Linux”, but it did point out some interesting ideas that Linux had in order to maintain processor affinity, one of the points that David and I had been talking about over lunch.

The next paper, “A Universal Dynamic Trace for Linux and other operating systems”, was very interesting. It could provide a flexible replacement for ktrace. It was developed for OS/2 by Richard Moore, now at the LTC, and it looks like it could be very useful both for FreeBSD and for both Linux and FreeBSD on the PPC64.

Back to the meeting, where we finally got round to a few words on the SMP implementation before going to the closing session. Sam Leffler was there, and he expressed basically my concerns that we need somebody with a big stick to tell people what to do, and that we need to instrument our system to find out just what it's doing.

Rob Kolstad is here. He must have arrived last night, and he closed the conference off with another quiz session, amusing as usual. After that with the rest of the core team to have dinner together, with the intention of talking, but the weather had turned bad, and Mike Smith has a leg in plaster and is thus not waterproof. To Chilli's, where the noise level was too high for most of the discussion.

After that to the party, and found Rob Kolstad and more or less extracted a commitment from him to do a tutorial while he's in Sydney. Left relatively early; it's been a tiring week.

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.