SEUL: FW: Re: COAS ??

[Cyberdyn <cyberdyn@seul.org>]

Here's a bit more on Jack's opinion of COAS.


-----FW: <Pine.LNX.3.95.980204123320.26140H-100000@486dos.solucorp.qc.ca>-----

On Wed, 4 Feb 1998, Rick wrote:

> 
> On 04-Feb-98 Jacques Gelinas wrote:
> -> Nothing happened. Mostly, COAS people were not interested in linuxconf.
> -> They basically wanted me to fold whatever I had into their architecture.
> -> Oddly enough, they wanted to fold a 80,000 line project into a newly
> -> borned project. And they could not tell me what was wrong in linuxconf.
> 
> So they wanted you to hand over your work to them to become theirs?  Sounds to
> me like that kind of move would pretty much put an end to Linuxconf and hand
> over your current users to them.

This is not exactly this and in a sens, this is exactly this. COAS is gpl,
so the code does not become their code. But they wanted me (they did not
ask that much, but see below) to work in their framework and basically
forget linuxconf.

Mostly, their strategy/attitude is: There is this Jacques Gelinas who is
interested in system administration. He already has done "some" work on
the subject. We will show him COAS and its great architecture and he will
join us in the effort to bring a nice system administration to linux.

During this presentation (I was invited in San-Diego to preview COAS), I
was trying to figure out where linuxconf was fitting there. I mean, they
were surely aware that linuxconf had already a sophisticated architecture

        -modules
        -multiple UI, including remote GUI operation
        -fully translatable
        -user privileges
        -and already so many service and package supported

So I was waiting for them to tell me how they envision a merge since I
don't believe (nor they do) that the linux market need two admin systems.
This is of prime importance that one must be available on all distribution
so they look the same in this area. This will be the primary argument used
by M$ and other anti-linux: THe market is fragmented and they are all
different version, incompatible and so on.

So I was waiting and they simply never talked about linuxconf. I think
they were convinced that they have a super architecture and that linuxconf
was simply a toy, an attempt.

After a while, I started asking them about various stuff linuxconf already
had without refering to linuxconf itself. It was clear that linuxconf was
not part of the project. So far what I had seen was a big "hello world" 
with a user interface much less sophiticated than linuxconf, but I was
thinking that there was something else. So I started asking, without
talking about linuxconf, if COAS was handling

        -configuration change activation and booting

                No!

        -user privileges

                No!

        -And few others, such as configuration file management
         (changing paths here and there)

always no! And I was convinced that those features, especially the first
was the key to a more reliable linux and also the key (as proven later) to
system profile versioning and multiple machine management

        (If you hope to publish some configuration changes to 100 machines
         you better have a reliable way to make the changes effective,
         something that linuxconf is able to do for a long time now, but
         something that is very complicated also)

In the end, a little desperate, I asked them if they had experience with
linuxconf. Another no!

My understanding is that the prime goal of COAS is to have a C in front of
its name. It must come from Caldera and be visible, even if it is free. Do
you imagine if COAS is the standard sys admin tool for all distributions ?
What a benefit for Caldera.

I think it explain why linuxconf is so wrong that

        they can't explain me what is wrong

Basically, they will tell you that I was not ready to discuss in an open
manner. The reality is that the case was closed well before I was
introduced to the project. Anytime I was asking about a feature linuxconf
does, the answer was that

        "we can add that to COAS also"

Well, linuxconf does it now, what is the point. What is so wrong that
can't be fixed in linuxconf.

Today, I care less about Caldera since linuxconf will be officially part
of some next release of redhat and they are very serious about it. We
exchange several mail a week and I have a precise list of things to do so
this becomes reality. They are also sponsoring linuxconf development

Once linuxconf is part of redhat, the visibility will be unavoidable. And
I will make sure it works on Caldera too (as it is currently doing).

The effect of that agreement (with redhat) is already starting to show
with some contributors appearing. I suspect that in a short while,
linuxconf will turned into a major show. It is already very sophisticated,
but expect to see more and more modules.

--------------------------------------------------------
Jacques Gelinas (jacques@solucorp.qc.ca)
Linuxconf: The ultimate administration system for Linux.
see http://www.solucorp.qc.ca/linuxconf
new developments: remote GUI admin, multiple machines admin, wu-ftpd

--------------End of forwarded message-------------------------

---
E-Mail: Cyberdyn@seul.org       Date: 04-Feb-98        Time: 13:24:30
Take everything in stride.  Trample anyone who gets in your way.