I've done an entry for this, and I'll be posting my monotonous blatther later on after I've tweeked it a bit to remove the bits where I had to re-read sentences that I fucked up.
I've also got a picture to post in the PFC, although I've got to find the cable to get it off my camera and onto my computer.
I was going to write an entry for it, but I got a bit put off by some of the comments in the initial diary and couldn't be bothered in the end. I didn't even end up voting. I read about half the entries, didn't get around to reading the rest and then the voting was over.
So, I screwed my bike up a while ago. I took it out in the rain, it got very muddy and I just put it away again without cleaning it off. Then I got it out of the shed again and it was fucked. The chain was rusty, the front derailleur was seazed up. There was also a distressing bulge in the front tyre.
I gave it a good clean and lube. The chain is OK now I've given it a scrub, the front derailleur is almost back working properly, although it's got a bit of stickage still. I had to get a new tyre for the front and that's OK now. Unfortunately I lost a small strip of rubber from my rear mud guard so it doesn't fit around my seat post anymore :(
I was sure I had lots of little strips of rubber around, but I can't find any now.
I've only gone and bought some proper cycling clothes as well. That was no easy task for a fat bastard like me. I've got a top, and it's great. Much better than the t-shirts I'm used to wearing.
Next: Cycling shorts.
It really does perform very poorly. I'm fairly sure that if some effort was made to correctly optimize the code it would work a lot better on smaller hardware.
For example. There is a similar bit of code at work that built up a data tree in much the same manner as the scoop comment code. I reduced run time from several minutes to several seconds by changing the way data was fetched and processed. Of course there was a trade off between speed and memory size, but I doubt it was that large. Modern computers & languages work very well with allocating huge chunks of memory at a time, and quite poorly with multiple allocations of lots of small bits of memory. The change I did at work was to read every "comment" into memory at once using a single database call, then sort the tree out in another pass through memory. The original code did many recursive calls into the database, just like scoop, and it was very slow.
Part of the problem with changing scoop is there is a lot of code there that I'm not sure what it does but must be there for a reason so I don't want to change it.
I do wonder if it would be better in the long run to just do the changes without bothering about being backward compatible with other installations. It's not like a lot of code makes it back into the main code base anymore anyway.
Not bad, not good. Just sort of ticking along.
Got one more person at work, who's got off to a good start. Seems to be picking up the job quickly, which is always a bonus.
|< Just one time in your life. . . | so far, so good >|