Re: [docbook] SGML support

[Frans Englich <frans.englich@telia.com>]

On Monday 27 December 2004 20:22, Lauri Watts wrote:
> On Tuesday 21 December 2004 21:55, Frans Englich wrote:

[...]
> We are inherently conservative about moving DTD's.  We must retain
> backwards compatibility for third party vendors, and providing a stable DTD
> for the documentation is just as important as maintaining binary
> compatibility for the code.
>
> Dropping SGML support would not IMHO be an issue for KDE, and even if it
> turns out it is, the existing versions won't cease to work the second a new
> one is released.  I do expect we will continue to use a DTD rather than
> schema, we simply have far more expertise available, but I see that a
> default DTD will still be provided, or we could generate one.  So nothing
> to see here folks, move along, and keep up the good work.

This is apparently something we have different perceptions of and will have to 
be discussed in the KDE community.

A major version is the most suitable period to do such move, because it among 
other things explictly means interface changes(although it of course should 
be avoided). Moving towards a more standard setup, say a standard Docbook 5 
setup e.g. less customized, would also be a gain for 3rd parties in the large 
picture.

Regarding expertise; imagine where KDE or any other project would be if 
progress had to move inside the boundaries of what already is known. What is 
required for moving towards being more XML oriented is neither that 
astonishing: it's a schema language for those few who are to work with the 
internals; for documentation writers it means the XML namespaces concept. But 
if it nevertheless is -- then we can learn.

The vague ideas of a new incarnation of the KHelpCenter, as dicussed on 
kde-cvs and kde-usability, also assumes fundamentals such as XML namespaces. 
While those ideas are currently vaporware, this isn't: any development and 
new XML applications will be introduced via namespaces.

I think there is wide, general advantages for moving towards Docbook 
5/XML-orientation, and that we're approaching the perfect moment for such an 
transition. Where the consensus will steer during KDE 4 development is 
obviously another matter, let's see :)


Cheers,

		Frans