[plug] Hack, triplej and open source

Craig Ringer craig at postnewspapers.com.au
Wed Jun 30 22:32:40 WST 2004


Russell Steicke wrote:

> He sounded quite well informed, but I was a bit surprised at one thing.
> He mentioned that companies could save money because the boss realised
> that "I've been supplying my people with a Rolls Royce on the desktop,
> when they only need a Mini".

It's IMHO very true, and an important area where open source desktops 
are useful despite what I see as some serious issues with them. Users 
with basic needs - email, word processing, and 'net access, plus maybe a 
custom in house client/server app or two - don't need WinXP. Giving them 
a full PC with WinXP or whatever imposes a significant management cost 
as well as a significant licensing cost and a hardware cost that's 
probably also larger than it needs to be.

That's one of the reasons why so many companies have stuck with Win98 
and even Win95 for so long.

(note: I don't have enough time to edit the following down, so sorry 
about the verbosity):

I see a real role for Linux in this "basic-needs user" environment. Not 
so much "full" Linux desktops, which I'm not convinced of the benefits 
of at this point, but Linux thin clients. It's often possible to re-use 
your existing hardware (don't tell me you can do that nicely with a 
current desktop distro running locally!) while providing a simplified 
environment to the users that lets them focus on getting their job done 
with little need for support and only minimal client management.

I'm using LTSP with XFCE4, Mozilla, OpenOffice, Xterm (for a SCO 
OpenServer console app) and ROX-Filer for exactly that purpose. None of 
it is particularly spectacular and it has some serious deficiencies, 
especially in the user interface department (#include 
<dialog-box-rant>), but most is so simple that even our computerphobes 
have managed to grasp it better than they ever understood Win95. It's as 
close to a computing appliance as I can make quickly with commonly 
available tools, and that's what these people need. A side effect of 
this has been that the users appear to be more effectively using what 
tools _are_ put in plain sight. I even have a couple of users who've 
forged ahead and learned to use the GIMP (!!) so they can make montage 
desktop backgrounds ;-) . Directly produtive - no. Useful indirectly in 
terms of improved understanding and computer use ability - yes.

The users in Sales here would've cost $1500 to $2000 to provide with 
"conventional" desktops (new Celery, 15" LCD or CRT, XP-Pro, Office) and 
perhaps $1000 to fit out with WinTerms (new dedicated terminal, Office, 
TS CAL, re-use existing monitor) - with a larger server and more 
expensive server software, too. Instead I built the clients for $130 - 
of which $60 was a new keyboard and mouse - and no license fees. To top 
it off, I hardly have to spend _any_ time in supporting them - it's if 
anything slightly quieter than with the well-worn-in 486 Win9x boxes 
they used to use. Of course, the server cost megabucks, but it would've 
done so with the 2k / 2k3 TS approach too, and would NEVER have been as 
versatile as this one has proved to be.

If there was a central configuration scheme for Linux user-level apps 
and environments - not eight different and incompatible ones and many 
apps that just don't have _any_ - then it'd be even nicer. /etc/skel 
just doesn't cut it, and many apps have such broken config schemes you 
can't even use that (Mozilla, for example). I'd _dearly_ love to see 
something functionally similar to Group Policy for user-level Linux 
applications, especially if it was implemented in a way that was easier 
to get started with. Being able to integrate _all_ applications into 
something like KDE's configuration scheme and Kiosk lockdown system 
would solve 90% of my needs with only about 10% of the complexity of 
Group Policy, so that would in many ways be preferable. But that would 
require CONSISTENCY and STANDARDISATION so I figure it'll be at least 
five years before anything like that gets rolling well enough to make 
sure that OO.o and Mozilla support it. I suspect this need is shared by 
a lot of "early adopter" organisations running Linux desktops, 
especially thin clients. Of course, as I don't have the time and 
resources to do it myself I guess I'll just have to wait.

So - I think the guy actually made a really valid point. Even if you 
think the Linux desktop is very poor in a lot of areas, the chances are 
you'll be able to make good use of it for users who only have basic 
needs. In my experience, Windows just doesn't cater very well to users 
who only need simple 'appliance-like' access to computers. You can 
potentially save quite a bit of cash in admin time and licensing costs, 
and even more in hardware if you don't need to replace their existing 
machines.

Hooray for the diskless P100/32MB smoothly running KDE 3.2 on a 19" 
1280x960 monitor - WITHOUT having to pay TS licenses, CALs, or per-seat 
app license fees. Hooray also for the P100/32MB running IceWM or XFCE4 
and Mozilla for basic users.

Phew. Congratulations if you made it this far ;-)

--
Craig Ringer




More information about the plug mailing list