[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: corporate debian desktop



Jerome Warnier wrote:

Le mercredi 05 avril 2006 à 01:53 +0100, Luis Matos a écrit :
Jerome Warnier wrote:
[..]
i gave up running linux on older machines like that. easely you can have pxe and bootp mount / trough nfs or have X clients and a server.
Those machines are enough, believe me. 128MB RAM is a little too short
to run OOo, though, especially when using other (huge) applications like
Evolution at the same time. 256MB make it far better. Still, it is
usable.
Thin clients are another thing. You would then need a powerful PC as server.
Both solutions share stuff, but they both show different issues and
solutions, and the goal of specializing is to get more specialized...
So, if you want to have a full-blown desktop suite (with all bells and
whistles), keep on working on that, first, and then see what you can
apply to thin clients.
BTW, we also have a fair experience in thin clients, being also the
authors of Plume (http://plume.bxlug.be) a while ago, when LTSP wasn't
the greatest stuff on Earth (still not? ;-)). Plume is just the
infrastructure needed to have thin clients booting, and booting really
fast, and working great. Then, we had to tweak and personalize a lot
more stuff to have it ready blow (even on clients based on 486SX
processors with 8MB RAM). Work on it is only realized when I get a paid
customer for a customization on it nowadays.
i have an example for you: 7 workstation client + 4 on other store. have a nice xeon server (some previous linux company based on suse put it there).

11 workstations cost in here, minimum 450 euros that makes 5000 euros +/-.
11 thin clients (regarding that we have a powerfull server, costs 200 euros and the last for ... many, many years.
all 1500 euros.

if a powerful server costs 1500 euros, added to previous 1500, it reaches 3000, so even without the server we stay 2000 euros short. simple questions, in a simple client.


Also you have java and C# full support for develloping applications. Maily my objective with that is to create somehow a base desktop for companies to build some software on it.
Well, the question here is "why don't you use Ubuntu, then?" :-D.
Ubuntu was judged to power-hungry for us, that's it (and the current
version when we started was Hoary, so not as nice as today's Breezy
going to Dapper, but still too hungry).
I like ubuntu, and dislike ubuntu. Debian, for me, represents stability, versatility, and losts of quality packages.
Ubuntu, for me, represents a good home desktop.
6 months release ycle i think is not enought to have stability like debian has. And installing software that is "not supported" by the ditribution, i think is out of case.

I think, sooner or later, for me, ubuntu will be the way to go when they can make me think that have a solid server and desktop solution, wich they don't have right now.

Maily, in portugal you have linux with some support, in servers is becoming very used, but in desktops, simply there is no end costumer applications, such as local crm, stock management, accounting. So, having a stable desktop, is easier to make people to create applications.
Release cycles for Debian are too long for desktop usage, that's for
sure (while for servers, it is probably adequate). I personally think a
1.5 years release cycle would be perfect for an enterprise desktop.
I think not. Imagine you install ubuntu. will you upgrade the systems every 6 months? No. (ok, if you hav nothing more to do) In my university you have gnome 2.6 (on some mandrake machines) ... and it is all there. If you install sarge now, you have a good gnome 2.8 desktop, red hat uses it in rhel workstation. Off course it does not matches with 2.14 ... but that is called evolution and, ofcourse, as you know this, you would like to stick with upstream version, the you have ubuntu. I prefer stability, so i choose debian.

In the example i told you, we had ubuntu there, and it simply had bugs, mainly in openoffice. (using kubuntu)

Also, using an lsb compatible distribution (or using dcc as base, witch have kernel improvement and is lsb directed) you make it even easier.
I'm not sure DCC is actually leading to something, though. We'll see
with practise.
i have been seeing dcc. It is a much better way to install sarge that the debian installer, mainly because you have a new kernel in the installer witch makes easier to install.

Also, dcc advantages will be seen when some companies start to make software to linux.

i saw your cvs, and what you did i basically my intentions, but without modifying the installer.
We did modify some things related to it. But again, our goal is to stay
as near to Debian as possible, and we intensively use the BTS to submit
bugs, improvements and patches so that Debian benefits immediately from
our work. We are really serious about this.
nice than ... i have the same objective.

Maybe provide tasksel support, or to announce a sugestion for preseeding.
I don't understand this sentence.
instead of making a new installer, we could make a patch to tasksel and then the installer would have the option to install the librassoc's, or others' pckages, and by that mean you could suggest example preseeds to those who want to use them.

Best regards

Luis Matos



Reply to: