Re: (sgml vs XML)+indexing

[Dan Mueth <d-mueth uchicago edu>]

Hi all,

I just had a nice chat with Laszlo about this, and thought I should put
some of what we discussed on this list.

First, I think where Laszlo says XML he means "DocBook/XML" and where he
says DocBook, he means "DocBook/SGML". Sun uses SolBook which is a
_subset_ of DocBook.  This means that the GDP should be able to run any
SolBook doc through jade just fine and use it.  It means that Sun cannot
use any GDP document as-is unless they convert it (ie. rip out unsupported
tags) or extend their tools.

If I may summarize Laszlo's questions:
1) Why don't we use DocBook/XML?
2) Would we want to take up Sun on its generous offer to let us use their
Java-based SolBook/XML indexer?

As for (1), Dave Mason is really the guru here.  The short answer is the
free tools which are available to us are not yet ready to our
satisfaction.  Parts of the solution are ready, and parts aren't.  We
can't really move over until we have a complete solution.  I'll have to
defer the details to Dave who knows the status of each part.  But I spent
about an hour trying to convince him to move to XML and he very skillfully
explained why we should not, even though I can't remember all the details
now.

As for (2), I see this as a technical point which should be addressed by
people like Jonathan, Maciej, and Elliot.  People have discussed using
medusa for indexing in the past, but I am not qualified to make a
comparison.  A couple things which stick out are (a) it would probably
have to be extended from a SolBook indexer to a DocBook indexer, and
(b) it may be a problem to have it Java-based since Linux does not have
very good Java support.  I will have to defer this issue to the hackers
though.  (Whatever they decide, we appreciate Sun's offer:)

Jacek - Welcome to the GDP.  

Dan

On Wed, 25 Oct 2000, Laszlo Kovacs wrote:

> Hello,
> 
> I talked to Jacek Ambroziak today who works for Sun on the StarOffice
> (OpenOffice) Help System. They work on an index based search engine on
> XML files (implemented with JAVA XML parser) as XML is the OpenOffice
> documentation format (I hope I got it right Jacek). He asked me why
> Gnome still uses DocBook as XML is more superior to that. He mentioned
> that conversion from DocBook to XML should be easy. He also said that
> the JAVA based search engine will be open sourced soon in which case if
> the Gnome docs would be XML we would get a complete search engine for
> the Nautilus Help Browser. 
> 
> Indexing and search was scheduled for Gnome 2.0 as I remember Dan's
> proposal. Dan's proposal seems to be much wider though as it has
> application name, doc title, index based search and body search in it. I
> think that indexing the most important of them though.
> 
> I don't know much about XML vs DocBook issues so I couldnt answer
> Jacek's main question (why still DocBook?).
> 
> Jacek is not on the mailing list so please keep him in the loop in any
> reply as probably he is the one eho would answer most of the questions
> related to this issue.
> 
> Thanks,
> 
> Laszlo