I've been having wrist problems recently, so I decided to try a track ball. I've bought a Logitech Marble trackball. It has a track ball and four buttons — two main ones, and two smaller ones which can be bound to different things. The secondary buttons did strange things by default (operating back and forward history in Firefox, and Mouse 4 and 5 in Emacs). So I ended up installing the Logitech drivers to rebind these. Very annoying. As well as mouse drivers it insisted on installing Music jukebox and a desktop E-Bay shortcut! This has to be the most irrelevat co-install ever. The drivers are also annoying; the GUI removes the "pointer trails" options which I generally use and always binds a click on both the main buttons to something, rather than just letting in through to the underlying system, thereby blocking cygwin's middle mouse emulation, for example. I managed to recover this situation with a bit of judicious registray hacking. Pointer trails can be turned on here, and the allowable gap for recoginition of a keychord can be turned down. But, really, should we have to go to these lengths?

Still. my intial experiences with the track ball are good. My wrist already feels more comfortable; it's largely static during mouse use. My accuracy still leaves something to be desired, but it not far behind mouse use already. Hopefully this will improve in time which will mean I can turn up the speed somewhat as well. For those with stiff wrists (oh, er, Missus) a trackball is recommended.

I spent the weekend back in Worcester for a sad occasion: the funeral of my Uncle Viv. A funeral can be a maudalin experience, but this wasn't. It was a great opportunity to reflect on my Uncles life. I remember him in his house describing events from his life. Most of these involved his work — he was a train driver — and a lot of them involved incredible feats of alcholic excess. Often at the same time. But he was a much more than this; in his time he was heavily involved in the trade union movement, making the life of other workers better and, crucially, safer. It didn't take long in his presence to appreciate his humour and the ease of his personality; it's perhaps only now, after his death, that I've realised how much his compassion defined his life.

One of this other major characteristics was his ability to talk the hind legs of a donkey; this is something that most of my father's family — including myself — share. An opportunity to see and talk with them is always good; the stand out moment was my aunt telling the most tasteless gag that I am ever likely to hear at a funeral. Excellent!

I've borrowed an old IBM laptop to be going on with. It's fine. Nice big screen, good hard drive. Slightly broken keyboard and a wireless card that only seems to work at 11M. I'm going to get one of the small Sony 11in laptops in the long term though. This machine is something like 3kg which is way to heavy.

Yep, shortly after getting broadband at home, and a funky new NAS box, by laptop died. It had been getting increasingly slow. I had decided that this was probably because it was having a Microsoft Moment, although there was possibility that physical trauma was the cause. So, I tried a full reinstall. This went okay, but didn't solve the problem.

On Saturday, while drunk, I discovered that it was only registering half a gig of memory, rather than the 1G it should have. So I took the memory off, blew the contacts clean with compressed air and reinserted. And now it won't boot. Worse, I have now discovered it should only have half a gig of memory.

Never, ever fiddle with hardware when drunk. It's only going to end in tears.

After weeks of not much of interest happening, there was a flury of activity today on the Semantic Web for Life Sciences mailing list. This was largely the fault of Alan Ruttenberg who used the two words which on their own are most likely to cause an argument between bioinformaticians — "identifier" and "standard".

How depressing it is that we are still having these discussions after so much has been achieved. Bioinformatics will use a standard when it suits them; people have been active in using GO or MGED. Identifiers, however, still remain a problem.

I've seen Billy Bragg before, but never at a gig that I've had to pay for. He gives value for money, it has to be said, and played a good long set; he still can't sing, and it still doesn't matter that much.

The stand out performance, though, was the support which was Seth Lakeman; I've rarely heard such an intensely rhythmic band, and certainly not one playing on acoustics. Be looking forward to seeing them live at a smaller venue.

Finally got my broadband connection. An interesting experience; spent 20 minutes fiddling with it, and failing to get a connection before I gave up and phoned the support. They were alright — the modem drivers needed re-installing and the windows config needed doing manually. Two days later, I got a Linksys ADSL Modem/Wireless router. Ironically, I managed to get up and working in five minutes.

My file splitting scripts didn't work initially. The problem is that Unison uses temporary file names while copying and this includes directories. So if you transfer a single directory containing 2G, for example, Unison will use a temporary directory till it has the whole lot. So I tried rsync instead; this worked well up to a point — about 2.4G as it happens, where a bug causes it to block.

Finally, I worked out how to get Unison to work — firstly, I run it with

ignore = Name {*.*}

This matches all the files (and not directories by and large). So the directory structure gets transferred. Then, I run Unison again; now it transfers the files so you get a restart with the granularity of a file. For really large files, I can still use my split file scripts.

Having said this, I am getting very variable download rates from the ADSL — 1.5Mbs and more at night, but during the day rarely more than 500K. To be expected, I supposed.