Re: debiandoc-sgml vs. docbook
In article <Pine.GSO.3.95.981129003755.6503Nfirstname.lastname@example.org>, Havoc Pennington <email@example.com> writes:
> On 29 Nov 1998, Adam Di Carlo wrote:
>> Docbook *doesn't* meet our needs. It does meet the need for
>> tables (CALS model) and figures and all that, more or less.
>> . docbook is way to scary for newbies.
>> . docbook doesn't support texinfo or ASCII or NROFF output.
>> . even then we'd have a lot of work ahead of us to extend docbook,
>> i.e., the <package> tag, etc., etc. Volunteer?
> If we feel these things are important, we can't ever use DocBook for
> debian docs, even if it does have some features DebianDoc lacks.
> At least, from the perspective of the tutorial: I need the texinfo
> output, other people will demand ASCII, and frankly DocBook seems
> awfully scary to me even though I've written some stuff in it. (It's
> just plain annoying to type <emphasis> instead of <em> for example,
> and the tags are very API-docs centric, not book-centric).
> There's no point avoiding one set of disadvantages only to get
I don't think you've really grasped what a Formal Architecture would
buy us. We could *translate* from Docbook to debiandoc.dtd in a
(theoretically) lossless format. We would be DTD agnostic. Since the
DTDs (at this particular point int time) are bound to a formatting
engine (debiandoc.dtd with Perl/sasp, docbook with DSSSL/jade), we
could exploit the benefits of both.
Of course it's all vapor until someone does it, but, please, don't
knock it until you at least understand it.
.....Adam Di Carlo....adam@onShore.com.....<URL:http://www.onShore.com/>