On Thu, Jul 10, 2003 at 11:24:37AM -0400, Timothy E. Miller wrote: > Well, since that accidentally went public, I'll add that "most bang > for least currency" should also take into account people/management > resource costs as well as simple hardware costs. > > Saving US$10k on hardware only to require having to hire an extra > full-time person to continually manage the poor quality gear is *NOT* > a bargain by any stretch of the imagination. It can be a tough compromise to get right, with unknown variables in the form of staff value and tradeoffs even in the gear quality (how much extra mileage can you get out of top quality stuff, would it be cheaper to get lower quality gear and upgrade sooner), in addition to the hardware/software choices, server room / infrastructure choices etc. It's not always particularly easy work making choices. Most people probably go with what they know? Leading to commercially targetable groups in different areas. For example some people may trust/know their staff and their admin abilities, and may save on admin software - the hardware enthusiasts may do that the other way around. In a slightly different way, some groups would be less inclined to 'try something new' with such sizeable amounts of money to spend, even though a different architecture with a clever programmer could be much more beneficial to them. Or just generally 'stick with what the other biologists/physicists are buying' etc. Interesting whether 'management resource costs' would extend to buying a subscription to a magazine for their employees, in the context of Douglas's posted survey. Kp -- Karl Podesta + National Institute for Cellular Biotechnology, Ireland + School of Computing, Dublin City University, Ireland