the viability of free software (Re: Debian needs you, and "the free , software community")

[Magossa'nyi A'rpa'd <mag@bunuel.tii.matav.hu>]

On Sun, 2 Nov 1997, Alex Yukhimets wrote:

> > I also would like to demontrate to Alex and others that their
> > "apparently reasonable" request will break that model and Debian
> > therefore. No religion nor dogmas here. DFSG wasn't a "decision" taken
> > after personal feelings, but a technical requirement to let the model
> > works.
> 
> Great! That's exactly what I would like to hear. DFSG is a *technical*
> requirement, not a bible. And as any technical requirement it must have
> some underlying goal to attain. But as environment changes, some
> requirements appear to be too restrictive, others too loose. So
> we should feel free to modify them and not rule out _forever_ anything not 
> complying with _current_ requirement.
> 
The current requirements are the free software base and the bazaar modell.
I will try to prove here that the first requirement will be with us for a
long time (I think at least for another 30 years), and the second is the
best practice we know. I also have a third idea, which I want to share with
you. It could serve to resolve the "contradiction" between the "free" and
"commercial" software market (I hope).

It is also pointed out, that the freeness of the software is a very
important point. This requirement is crucial because the quality is
achieveable only if we have and can modify the source. As soon as someone
can work out another modell which serves this goal, this requirement will be
a dogmatic tie. I doubt however that there would be a better one.

The bazaar modell also undoubtedly better than the cathedral one. The
uncounted times referred "cathedral vs bazaar" article can prove it better
than me, the crucial part is that more developer and tester can bring a
software into a good state, than few developers and little testing. It has
its problems, as we could see with the 2.0.3x kernels, but still the best I
know. I see that the important point is to strictly insist the double
(stable/instable) versioning scheme, and doing code freeze at the milestone
points. It is possible that someone can come out with a better versioning
scheme, or even a better modell. I vote we should stick to it until there is
something which is proved to be better.

There are problems with free software however:
Free software is superior in quality, have better support if you know where
to look for it, etc. But still only barely accepted among the manager type
of people (and they will make the decision, which is a pity:). The main
factors are:
-Marketing
-Support and warranty understandable and acceptable to a necktie.
The first one is a very hard question, we all do what we can (have you put
that DataPro survey to your boss's deck also?), but I don't know of a
marketing scheme which can be carried by small businesses and can compete
with the full scale brainwash our beloved software monopoly does.

I think I have however an answer for the second problem. You've heard the
argument a lot of times: it is free software, and there is no support or
warranty. You might also read some support contract offered by the big
players: they are talking about big heaps of money (the heap they get for
the support is especially big), but the technical details have nothing of
the type a ten-people company with good net access couldn't easily overbid.
There are those small companies doing linux-related business: I think
especially of installation, system integration and the like. Those companies
typically have an order of magnitude brighter staff and an order of
magnitude more modern and scaleable and etc. solution than the big ones.
They have a big disadvantage tough: they are much smaller, and just can't
have the monetary mass which is necessary for the neckties even to play with
the thought to choose their solution. 

The solution would be a new type of insurance which could enable the small
company to be a representative of a much greater monetary mass. It could be
made by an old-style insurance company also, but there are some techniques
which would make it more effective and efficient. The insurance company
could have a "backup team" for each work it warrants (the backup team could
be among the companies it warrants, as well), could promote free software in
general, and it could act on behalf of the small companies where it is
necessary. It could also organize the small companies if a big work is
coming along. The small linux companies I know also use some of these
techniques, but the supporter company in most cases only one magnitude
greater than the one which does the real work. I can think of at least one
case where a linux-based solution would be the most appropriate, and the
ones who could provide that severely considered to take the opportunity, but
there wasn't enough monetary mass, and there wasn't a clear way they could
act as one entity.

I am courious what others think about the above: is it just daydreaming or
have some useable cue?
---
GNU GPL: csak tiszta forrásból


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-private-request@lists.debian.org .
Trouble?  e-mail to templin@bucknell.edu .