Getting Speakup working on a server Linux OS

[Garry Turkington]

Hi all,

Apologies if a duplicate of this appears, I sent it Sunday but it's
not hit my inbox or the archives.

I've been using Speakup on a single Linux machine for years, using
CentOS 4.x and a Dectalk Express.  This last means I've remained
reasonably oblivious to the software speech machinery.

In a recent international move however I've had a bunch of equipment
die, including my main server and the aforementioned Dectalk among
other items.  So this gives me the opportunity to do some
rationalization.  Basically I want to Speakup-enable a Linux box which
will have as a main part of its role to be a VMware Server host.
Consequently I'm looking for a relatively stable OS, ideally one of
the server variants out there.

With only hardware synths to worry about this would be reasonably
trivial as Speakup is my only dependency.  But if I need to use
software speech -- and especially with my preference for some
commercial voices -- I need get speech-dispatcher and speechd-up
working.

This is where the server variants get tricky as they tend not to have
any of this stuff in the main repositories, or indeed many of the
dependencies.  I just installed CentOS 5 in aVM to play with and it
looked like this was going to turn into a major self-build activity.
Ubuntu Server comes out of the box with no audio and I'm having a bear
of a time getting that to work.

So, anyone had success with either of the above or got other
recommendations?  I've got Debian 5 installing as I type and am musing
on just using that booted to runlevel 3 as an interim solution at
least.  Basically I want a host OS where the upgrade cycle on
dependent packages and kernels is relatively slow, with the server
hosting many VMs extended uptime is important.

Thoughts?

Thanks,
Garry