From: "alex" <[EMAIL PROTECTED]>
>  My first name is actually Christopher, but handily my parents changed
> their minds after registering my birth and decided to call me by my middle
[Oddly enough, same here. I'm Chris Paul ... It's an absolute pain in the
arse. Note to parents: don't do this.]
Without wanting to sound too real-world and pragmatic, why not upgrade the
server when you actually *need* to, i.e. learn how to monitor performance and
when it starts to suck, *then* buy new stuff. Otherwise you'll end up in the
situation it is now: a load of kit that's not being used & is depreciating
128MB RAM and a K6 is quite enough to run a decently hammered mod_perl site.
You only need more memory if you end up using a large database or doing
something rash like install Oracle. Assuming you're not on an OC-12 backbone
and you're not doing finite element analysis of an F15 jet per form
submission, your IO bottleneck will be the net.
Building reliability is probably your best aim: does it have a UPS? does it
have a RAID 1/0 config? Dual PSUs? Tape drive & backup policy? Those things
are way more important than a faster chip or RAM.