Re: Where, what and how - The future of DocBook

[Michael Smith <smith@xml-doc.org>]

Alan, you wrote:

> Mike, I was reacting against those who implied DocBook was so
> difficult that you absolutely required sophisticated tools to deal
> with it.

Yes, sorry if I came across a too emphatically. I definitely concur
with you there -- no sophisticated (commercial?) tools are required.
Using DocBook is never more nor less complex than your needs -- that
is, what you need to encode in your document instances.

And regardless of complexity, my response to those (not you of course)
who question using DocBook for hardware/software documentation is,
-Please show me an alternative-.

I guess someone who wanted to could re-invent the wheel and come up
with a kind of subset of DocBook. But Norm Walsh and the others
guiding DocBook have already provided DocBook with a customization
layer designed to facilitate subsetting and/or enhancements. Sure,
it's not trivial to design/configure a DocBook subset -- but it's a
hell of a lot easier than with any other DTD I've ever used.

> Also, you have, IMO a weak argument about validation since it
> actually doesn't take very long to do it on a modern PC
> independently of the editing process if you do editing in one window
> and validation in another.

OK, yes maybe I overstated that. But to put things in perspective: you
mentioned that you were working on a hundred-page document.

Not to downplay the work I'm sure you put into marking up that
document, but for a lot of the doc folks I know, a hundred page
document is a walk in the park -- these people (not me, thank god
almighty) are working on things like huge "butterfly" manuals that run
to so many pages no one bothers to count them. Awful stuff.

I would not want to work on documents of that length without a
vailidating editor. Actually, wouldn't want to work on that kind of
stuff at all myself -- but I guess somebody's got to do it.

> [...] I rarely (about 4 times in the whole project) had a validation
> error. So things went really fast for the simple method I chose.
> That is an important point I hope we can agree on; DocBook ain't
> difficult and doesn't absolutely *require* sophisticated tools.

You're right of course.

> That said, you have been a good advocate for emacs/psgml (and so was
> my partner in the project) so I will probably try it next time

Very glad to hear that.

If after using psgml, you or anyone on the list has ideas on improving
psgml, I want to encourage you to help make it better -- because
Lennart Staflin has, yahoo!, recently moved the psgml source to
Sourceforge, opened it to collaborative development and set up mailing
lists for discussions along those lines. Take a look:

  http://sourceforge.net/projects/psgml

> but I don't regret for a moment doing my first DocBook project
> this simple and straightforward way. I seem to learn the best if I
> solve problems with low-end tools to start, and then move on up to
> more sophisticated tools later on. YMMV, of course.

Again, you're right -- no one can take issue with that philosophy. 

> This has been an interesting discussion about methods and perhaps
> even valuable as well, but I sure wish I could get as quick a
> response on this list to the practical problem of Cygnus pdfjadetex
> not working in the latest release.

Well, I honestly wish I could provide insight on that specific
problem, but can't. Unfortunately, in my experience with Internet
discussions, it seems that instead of practical solutions to acute
problems, you instead often get just a bunch of opinionated
pronouncements from someone like me.

   --Mike Smith

-- 
Michael Smith          mailto:smith@xml-doc.org
XML-DOC                http://www.xml-doc.org/