On Wed, 3 Mar 1999 10:09:27 +0100, Cornelius Krasel
<krasel at wpxx02.toxi.uni-wuerzburg.de> wrote:
>Andreas Milton <milton at san.rr.com> wrote:
>> Actually I have experience with many OSes, but I am trying to make a decision
>> that impacts a large group of people. I want to accommodate all the
>> essential applications used in Molecular Biology. The problem I'm having is
>> that this field tends to be Mac and unix based. However, my gut feeling is
>> to go with Microsoft products because of the global acceptance of the windows
>>Microsoft does not offer products for the Molbio market.
>True, but lots of companies are now producing molbio
applications for Windows and many are phasing out their
Mac products. In general the NT versions are faster
and far more stable than the Mac versions.. We wouldn't
touch Win95/98 workstations for obvious reasons, so I cannot
comment on their performance.
>IMO, the choice is fairly simple:
>>If your user group is mostly GUI-based, buy Macs. If your user group
>consists mostly of "power users", buy Unix.
>You would regret moving to Macs within a few years.
>Other points you might want to consider if you will have to do the
>administration of the machines:
>>Macs are incredibly easy to network. Just plug them into the wall, give
>them an IP number and you are ready to go. On the other hand, Macs are
>a pain to administer because they are not intended to be multi-user
>machines. That is, every user can mess around with the system files,
>file sharing is fairly slow, etc. On the other hand, this is one of
>the advantages of Unix machines: you can share the file systems via
>NFS and you can do the administration of a whole farm of workstations
>from a single machine. Plus, the machines are running much more stable
>than any Mac or Windows machine I've seen so far.
>Unix servers tend to mess up their storage of Mac files and
often ruin them for use by Macs afterwards. NT on the other
hand was designed from the ground up as a file server for
the Mac and does a much better job of it. Far, far easier
to configure too. Correctly configured NT servers are just
as stable as any Unix server, but beware of NT administered
by Unix specialists.
>If I had to equip a lab with computers today, I would probably try to
>buy a host of cheap PCs running Linux and some standard software for
>word processing and the like (e.g. Applix). To that I'd add one "real
>workstation" running GCG for the more complicated molbio problems.
>I would use NT workstations on the desktop with a few Linux
servers to run Unix applications remotely and perhaps a Sun
for the things that haven't got a Linux version yet. For
file serving an NT Server with a big RAID system. For
running stuff like ABI sequencers you will need Macs, but
the ABIs come with Macs anyway.
Helgi Briem | Bioinformatics dept.| deCODE Genetics Inc.
Lynghals 1 | helgi at decode.is | Tel:354-5701805
Reykjavik | IS -110 Iceland | GSM:354-8988261