Re: Mangling other people's code
"Michael" == Michael Alan Dorman <mdorman-debian-policy@debian.org> writes:
> As to putting something in policy, I'm sceptical---what real effect
> will it have? The people need to be told are least likely to pay
> attention to it.
I agree. W.R.T. debiandoc-sgml, is there any evidence anywhere that
Ardo has done less than a sterling job there? So why inhibit
innovation on software that you've moved on from....?
> This does bring up one thing, though---can we look at transitioning
> away from debiandoc-sgml to something more standard, like docbook?
> The backends in the latest sgmltools are supposed to be much more
> sophisticated and produce better output than the old linuxdoc-sgml
> stuff, and as I remember it, that was your primary reason for
> creating debiandoc-sgml.
I could consider it, but you'd have to give some actual reasons why it
would be a good idea. As I read it, you only gave one:
* using a standard DTD is better than using a non-std one
[ to which I agree, but see below ]
Here's my reasons against it:
* Docbook is about 10 times more complex than debiandoc-sgml. It will
raise the learning curve for new maintainers *considerably* (it's
already pretty high with CVS + SGML at all)
* debiandoc-sgml has our own set of tags (like <package) which we can
play with and modify and use without worrying about it. Granted, I
suppose we *could* add tags to docbook
* the limited set of tags presented by debiandoc-sgml is a *benefit*
in that we have a greaters structural and presentational consistency
I just don't see why we should be using docbook. However, I would
like to see a DSSSL stylesheet to translate into docbook, and anther
DSSSL for print output as an alternative to the current (perlsasp or
whatever) mechanism.
I would like to see an XML flavor of debiandoc's DTD also; I don't
know what structural changes we might have to make for that....
--
.....Adam Di Carlo....adam@onShore.com.....<URL:http://www.onShore.com/>
Reply to: