Re: getting the ball rolling

[Karl Trygve Kalleberg <karltk prosalg no>]

On Thu, 20 Jun 2002, Jeff Waugh wrote:

> 3) Corporate users, who have very little control over what's installed on
>    their machines, and their administrators, who want complete control over
>    the machines and how they operate.

This is definitely the most interesting target group, IMHO, as this is
where the cost savings of properly written free software shines (forget
about licensing the cost; it is the maintenance that's expensive). 

The de facto practice of reinstalling each client machine's operating
system with all productivity apps once a month because of anal bugs is not
alluring. Add to that the fact that none of the client boxes in a regular
network can at present day be remotely administered, then the regular
complement of malicious email viruses, and you should quickly realise that
this is a money-consuming vortex.

But, in all fairness, to compete with the established solutions, a
heap of work on the usability level must be done. Requiring the
administrator to edit the configuration files by hand for trivial updates
will simply not be acceptable. The simple things must be simple, the
difficult things be easy enough, and the impossible must remain eminently
doable :)

The three different scenarios proposed pose quite different problems. On
the corporate desktop, you do not, in general, want a C compiler at all,
for instance.

Going down _one_ of the routes will prove more than enough work for a
small and focused team. Going down all the routes, during a reasonable
timeframe (12 months?) is hardly feasible, as I see it.

Therefore, we should narrow our focus so that we may be fairly confident
that we have a shot at releasing something worthwhile within a year's
time. (That does not imply rock-steady, ready-for-primetime. It implies
at-least-ready-for-preliminary-deployment).


Kind regards,

Karl T