It's all about the news

Now that the syndication system is all up and running, I've finally got my new subscriptions all in one place.

I'm now subscribed to:
bbctech, bbcukfrontpage, bluesnews, wired_news, saloncom, plasticdotcom, theregister, kuro5hin and slashdot.

Oh, you can only do this if you're a paid member, by the way. Just one of the perks. Apparently 356 other people subscribe to slashdot, but only I like the BBC's tech coverage. Anyone can add a feed, but they cost 'points' to subscribe to, and those points go down as more people join a group. I have just over a point left, so I can either grab a whole new feed nobody else reads, or I can join 3 or 4 that lots of other people subscribe to.

Or I can hoard it all to myself. Bwahahahaha.

Oh, and the syndication page is here. Of course, there's no link off of the front page or anything, that'd be far too easy for people to find.

Windows is dead, long live Windows

Here's the windows support timeline. You'll notice that Windows 95 is long gone, unless you've got paid extended support, 98 and 98SE went a couple of months ago. ME will drag the carcass of DOS on for another 14 months (26 if you pay for support) and then that vanishes too, and it's all NT based from there on out.

Speaking of NT based - you'll see that unless you've got the Security Roll-out Package, NT4 isn't supported, and even with that NT4 dies this Christmas. Is your company dependent on NT4? Has anyone told them that their support is going away?

Oh, and if you're not on IE6, then pretty soon you're going to have to patch holes all on your lonesome.

I can think of a few places that are going to have to run an update or two...

Rant Du Jour

I've been handed a set of reports to do which make up a subsection of the Intensive Care Unit's report for the year 2001. I was handed 2000's report and told to do it for 2001, so my first step was to try and duplicate 2000's data, so that I could check my methodology. When I've got 2000 figures looking right, I'll then just change a single variable from 2000 to 2001 and the right figures will appear as if by magic.

The problem being that no amount of jiggery-pokery is making the right figures appear. There's a lot of possibilities involved, and the different people involved sometimes tell me different things about what the figures actually mean. For instance, the figures apply to a set of babies for 2000. Whether the figures are supposed to be those babies treated in 2000, born in 2000 or admitted in 2000 has a fairly large affect on the totals. Additionally, if a baby was admitted in 1999, went elsewhere for a bit and then was readmitted in 2001, is that an admission or not?

Also, the figures came from 3 different sources, each of which give conflicting totals, because the reporting systems are entirely separate and work in different ways. This year they plan to use one reporting source (me!), but this makes checking I'm conforming to the normal standards almost impossible.

Aaaagghh!!

(no subject)

The only thing worse than working on old horrible code that's so kludgey that you want to kill the person who wrote it is...
working on old horrible kludgey code that you wrote yourself.

I'm fairly sure that this code was written to be fairly aesthetic originally, but was done before I had a couple of nifty ideas and simplified a lot of the system, and has now been encrusted by revisions because I was making whole rafts of changes that needed to be done quickly.

The result is a form that's going to be faster to rip out and replace than it will be to puzzle through and rekludge.

Ewww!!!!