On Mon, Jun 21, 2004 at 12:45:32PM -0400, Francois Lachance wrote:
> What I am suggesting is splitting the content modelling activity into two
> parts: first, housing the information that will be required; second,
> designing the "rules".  I think that the two arise from distinct
> questions: what needs to be collected, how does it need to be processed.

I just posted another email hinting at further complexities. As I said
in that email, none of these problems are unsolvable, but they do
require careful thinking and planning.

But I absolutely agree with your suggestion above that rules should be
divided into two distinct questions. My goal is to create an XML
vocabulary that contains all the relevent information about a reference
that one would need for any type of work. The xslt stylesheet will deal
with the complexities of formatting this infomration.

For example, the data for a title should be entered as:

<title level="m"><leadingSmallWord>A</leadingSmallWord> Time to

Now, in some bibligography systems, this should appear as:

Time to Harvest, A

And in other (such as MLA), it will appear as

A Time to Harvest.

The BibTeX system doesn't take the approach of seperating data and how
that data should be processed. That's why I don't think it is a
particulary strong system.

With the right XML markup, an xslt stylsheet could really create
complicatd references and bibliographies.

If I find enough interest, I will create a project on sourceforge to
tackle the bibliography challenge. But I don't think I have enough time
or expertise to create a through, workable system that I can submit to
the TEI working group before August 30.



*Paul Tremblay         *
*[log in to unmask]*