Re: [docbook] An apology and some minor revelations

[Daniel Veillard <veillard@redhat.com>]

On Sun, Jul 13, 2003 at 04:35:04PM +0200, Tobias Reif wrote:
> Any independent entity (eg a group of develppers, OS/distro vendors, or 
> any commercial company) could offer a toolchain setup ("system") or even 
> a one-click installer plus GUI interface, or a full-blown WYSIWYG editor 
> offer everything any tool collection could offer.

  Okay, who does except some commercial entities at the moment ?

> > Integration can be done to some extend
> > by the people creating the Operating System but to be really integrated
> > and getting simple for the end-user this would require a more synchronization
> > at the DocBook group level,
> 
> 
> I don't agree.
> 
> > which so far defines it only as a language and
> > not as a system (I'm not saying it the way thing should be done, but it
> > would be one way to make it simpler).
> 
> 
> I firmly believe that the DocBook TC should concentrate exclusively on 
> developing the language (specs, schemas), and leave the tasks of 
> implementing the language (creating tools, toolchain setups and 
> installations, the "system") to others.

then this can be another group. Take most of the successful OSS developments
they need centralized release and usually the definition of a platform in
order to get reliable across distributions/OS deployment to end-users.
As long as your target group are technician, the current unsynchronized
deployement are okay, but as soon as your want to target a less technical
group, I don't think it is doable.

> For me, the "system" is the loose network of independent entities (TC, 
> developers, users, vendors, etc). Only if the workload is distributed 
> redundantly across separate and independent entities can DocBook as a 
> whole continue to flourish, IMHO. I prefer an open democratic community 
> even with all it's complex challenges over a single-entity "simple" 
> "dictatorship".

Well well, you prefer it, depends who are you trying to serve. I will note
that projects like Gnome, are targetting end-users, have come with large
software base deployement, are using an open democratic process but are still
able to define a platform and a system. This is doable !

> Having the TC take care of everything from language to the complete 
> toolchain and installation would mean less competition (which generally 
> is know to be *worse* for end-users), and a single potential point of 
> failure.

You neglect the fact that defining a platform doesn't stop developping
concurrent tools which may get better in the end and be in the platform
at the next release. If there is little interest, then I agree there
would be little compatition, but having at least a suggested system 
will also bring more people to DocBook. You seem to believe that defining
a system can only be detrimental, I can see a number of open-source
developments groups where this was the basis for their success...
In the extreme why do you want to even standardize the grammar then, see
there are good reasons to try to define a common ground, and this applies
to software as well as to syntax.

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard@redhat.com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/