Re: [xmlroff] What needs to be done to make xmlroff a success?

[Stefan S. <se...@sy...>]

Tony Graham wrote:

> It's been a while since I've had this sort of conversation with myself, let
> alone with anybody else, and I surprised myself by not mentioning C or GNOME.

Heh ! :-)

> Given the above, a successful outcome for the project would be:
> 
>    A multilingual XSL formatter with conformantly implements the formatting
>    objects and properties necessary for formatting most technical and business
>    documents.
> 
> Nobody is going to tell you that the current xmlroff represents that
> successful outcome.
> 
> Opinions on the purpose, principles, outcome, etc., of the project are
> welcome, but if the definition of a successul outcome is sufficiently accurate
> for your purposes, my question is:
> 
>    What needs to be done to make xmlroff a success?
> 
> To brainstorm for a moment, ideas that come to mind include:
> 
>  - Implement region-before, region-after, etc.
> 
>  - Implement markers

I'm not qualified to comment on these two items, as I'm 'only' a user.
However, I think a good measure for success and adoptation in the FOSS
community is xmlroff's ability to process docbook into pdf (i.e. notably
the fo generated by Norm Walsh's xsl stylesheets).

>  - Provide RPMs as well as SRPMs and .tar.gz downloads

These are pure convenience. As soon as you have a tarball and a simple
build system that conforms to current standard practice, people will be
able to write packagers (e.g. rpm spec files).

>  - Provide a Windows port

That would be nice.

>  - Improve developer documentation

That would be nice, too, though I'm not sure in what sense it is lacking.
You already seem to have good embedded API documentation. Is that already
extracted and processed into an (online) reference manual ?

>  - Improve end-user documentation

I'm not sure what is lacking here, either. The most important information
for users, at this time, is the conformance matrix that informs potential
users about the current status.

>  - Rewrite in a more OO language

I'm not sure that would be such a good idea, at least not if by 'rewrite'
you mean 'start over from scratch'. Migrating would be a different matter...

Being a C++ developer myself, I would of course replace the GObject-based
type system by a true compiler-supported type system. :-)
But I'm not sure how practical that is, and whether there is really any
gain at this point. (It would be wonderful if xmlroff could become
GNOME-independent, though I doubt this is possible due to its dependence
on pango, right ?)

Regards,
		Stefan