samba development

[Chris Odgers]

I've been subscribed to this mailing list for the best part of six months, but
this is the first submission I've made.  (end disclaimer)

What I can see happening here, and what would certainly not be in the best
interests of either the Samba team, the end users, or anybody who has any
time,
effort or stress invested in the project, is a netbsd/openbsd style forking of
the
code due to different ideologies.  In the past, it has been common to see Luke
and others arguing on one side for their approach, and people on the other side
of
some imaginary fence complaining that Luke et cie are trying to make Samba
into
their own little version of NT.

I don't have any good suggestions for a resolution to this, but false
dichotomies and imaginary lines in the virtual sand are what screwed up Theo De
Raadt and
made him fork NetBSD into OpenBSD way back when.  It would be truly saddening
to see the same thing happen here, with Luke and co taking what has been
referred
to as a 'reference implementation', but which is obviously far more than this,
forking it, and calling it something other than Samba; and then having the
'official' samba project set back by many months as they try to retro-fit the
ideas
'documented' by this 'reference implementation' to the 'official' version of
Samba, months from now.

Always there will be differing opinions about ideas, both political and
technical, but it would be a shame to see everybody's work set back to a large
degree
by idealogical issues.  But then, if Luke wants to fork code which he has had
a
very large time and effort stake in, then it's his perogative.

Good luck, you're going to need it.

Chris Odgers
System Administrator, 
Sausage Software Ltd


On Thu, 17 Aug 2000 12:09:06 Luke Kenneth Casson Leighton wrote:
> > In other words -- from R&D to production.
> 
> that's been happening for two to three years.  the problem comes when it
> is "assessed" that those people responsible for the production releases do
> not accept the development of ideas, despite proof-of-concept bloody well
> staring at them in the face.
> 
> i admit that i have not outlined _all_ of the aims behind the samba
> dce/rpc development: i am basically aiming for a portable [that means no
> threads] ms-compatible dce/rpc development environment.
> 
> to that end, various "short-cuts" that have been proposed, such as the one
> by andrew tridgell yesterday on the samba-technical mailing list do not
> pull any weight.  arguments such as, "it is unlikely that" and "not
> frequently used" combined with "too complex a concept" to conclude that
> "the idea is therefore not justifiable" just do not pull any weight when
> aiming to provide the sort of functionality that a dce/rpc environment
> requires.
> 
> it is somewhat unfortunate that samba is in fact "just a file and print
> server".  i have been trying to break out the dce/rpc services for some
> time.
> 
> perhaps the suggestion i had of breaking out an independent source fork
> has merit.