It’s been a busy week. I’ve spent a lot of my time at the office, trying to get the replication model for Bovini working – causing much stress as it failed time and time again. For those of you without a grounding in computer science theory, replication is the art of making data be indentical (and editable) in several places at once without the fundamental problems that this goes on to cause, such as data identity conflicts.
In this particular case, we have two master copies of a database, and five smaller copies of a particular one-fifth of the data each (plus a little shared data), split around seven UK sites, and who’s computers can only be made to talk to one another between the hours of midnight and 4am each weekday. So: not only does the program I’ve been writing (and sweating on, crying over, and shouting at, this week) have to pull all the data back together and spread it out, it also has to detect whether two users at different sites edit the same piece of data during the same day, work out who’s most likely to be ‘right’, and ‘fix’ the data accordingly. Or, if it’s not sure, know who to ask for assistance. It’s a clever program.
And now it seems to be done. And working. Great!
Unfortunatley, working like a dog on this little project has only taken time (and energy) away from my preferred software project – Three Rings – a program I’m writing for free for National Nightline. I’m likely to have a busy weekend catching up!
Regardless, tonight… will be a night for relaxing – Bryn, Claire, Paul, Kit and I are going to spend the evening in the Ship & Castle, drinking Real Ale and playing Chez Geek. A perfect way to end a week.