>Again, if you aren't playing with it as a hobby, it's a bad idea to try to stretch out the life of a system, given the rate that hardware prices and performance change. I think I have enough of a clue to handle my own hardware, and the same advice applies to my mother as to my clients; the availability of support is more important than proving how much testosterone you push through your hardware. I don't want to have the same machine 2 years from now, and I do work with hardware as a hobby. IOW, it's bad advice for the developer who doesn't want to deal with the headache of rolling his own and guessing wrong.


This is exactly my attitude. I don't consider myself to be a hardware type. I've had very good reliability with my computers because I tend to purchase brand name (Compaq, HP, Dell, IBM) and then LEAVE THEM ALONE. The one time I changed this approach, the after-market WD drive I had installed crashed hard. I found out that my Cheyenne backup disks were not good and I lost weeks of work.

I changed to backing up zipped files to a ZIP disk. I can backup over 300MB of development work this way (3 to 1 compression) and it has never failed. Until I get a tape drive I plan to use the same basic approach with my CD-RW drive. That will enable me to permanently archive over 2 gigs of work which should hold me for a while.

When a computer is older (2-3 years), you can give it away or sell it cheap and buy a new one which is inexpensive because you are not trying to allow for a lot of future expansion. Then there are two computers in the world, one current one and one usable one. If you upgade your old computer, you get (at best) a less old but not current computer. And no one else gets anything.

At worst, when you upgrade, you get a flakey computer that gives you intermittent trouble or fails catastrophically.

IMHO.

Peter