discuss: DocBook Tools
Subject:
Re: DocBook Tools
From:
David Lawyer ####@####.####
Date:
16 Jul 2001 21:32:49 -0000
Message-Id: <20010716142257.C720@lafn.org>
On Mon, Jul 16, 2001 at 07:23:03PM +0200, jdd wrote:
> Le Lundi 16 Juillet 2001 02:44, Poet/Joshua Drake a écrit :
>
> > Command Prompt (My company) has written a tool that will take ANY html and
> > convert it to valid DocBook. Why am I saying this?
>
> this is very interesting
[snip]
>
> I think this can be a very useful thing. Really anybody can write
> html. It is very easy to give some guidelines/template to write
> acceptable html code.
It turns out that HTML doesn't contain the info (in the tags) that
sgml does. This goes for both docbook and linuxdoc. For example,
HTML has no sections but a parser may be able to guess what the
sections were from HTML tags for links (that look like they were meant
to be a "table of contents")
sgml tags define the content type while html tags are mostly to define
presentation. If one takes a sgml doc, does sgml2html and then
html2sgml (the actual command lines may be different), I would expect
to see a lot of differences and losses (except possibly for a doc that
was short and/or used few tags).
One can write a doc in html without any section-subsection
organization and that shouldn't be acceptable to LDP. Not only that,
but HTML is a lot more complicated than linuxdoc. So suggesting that
one write in HTML is not such a great idea. If you believe that most
of the docbook tags are useful, then you will likely find few of them
(or the equivalents) in HTML. So the loss will be even greater if one
compares writing in HTML as compared to docbook. (Not that I'm
necessarily advocating writing in docbook :-)
There's the possibility of writing in HTML, converting to sgml and
then checking to see if meets LDP requirements. This could be
automated. The alternative of linuxdoc is so easy to learn and the
problems with HTML are so severe that unless we can find a volunteer
to look into it more, I don't think we should plan on accepting
submissions in HTML.
David Lawyer
