🚀 go-pugleaf

> The difference between a thin-client and a fat-client in my mind is one of
> hardware upgrade paths. The question is rather, should the firm buy a
> powerful server (the thin client route) or upgrade the workstations (the
fat
> client route). Every other man and his dog seems to be developing a linux
> thin-client solution. Unfortunately the client is tied to MS Excel for
> certain mission critical parts of their workflow (OpenOffice and
StarOffice
> do not support VB due to their opensource and Java allegiances) - so
> whatever solution is chosen, be it, CrossOver Office, Lindows, Win4Lin, or
> Citrix - it will need to be able to run MS Excel satisfactory.

As much as I'd like to see more linux/open office out there, this doesn't
sound like the place for it (at least initially).

Transitioning to another office suite can be fairly simple if file formats
are the only issue, but a custom workflow system using VBA could be
prohibitive to migrate. I'm not a bit fan of building up applications inside
office suites for that reason. This is where the traditional unix way of
keeping the data and the tools seperate avoids lock-in, but you'd need
fairly savvy users and/or a lot of admin/developer effort to make it work.

It sounds like fat windows clients with a Linux/Samba server or MS SBS (I
hate supporting SBS by the way - but that's just me) would be the way to go.
SBS 2000 (or 2003?) would need a lot more hardware than a samba server.

I'm not sure I'd trust relying on Windows 'emulation' for critical systems.

The only economic justification for using Citrix/Terminal Services would be
if you'd have to upgrade all the workstations otherwise. P2s are more than
capable of running W2K Pro, if they have at least 128MB (pref 256MB) of
memory.

It sounds like some strategic workstation ram and OS upgrades (get all the
workstations running the same software versions), plus installing a
linux/samba server would be the cheapest option.

Future plans could then revolve around making a cross platform version of
that workflow system (eg Python/PHP/Java etc), then replacing the office
suite, then last and optinally replacing the OS. Users don't tend to like
too much change at once (and it's also very risky), let them get used to
each step first.

Cheers
Anton