Hi all,

I'm cross-posting this to both SW and XML dev lists, since I'm not really clear how to separate the issues out. They deal with both how SW might work, and about the XML details underneath.

I'm the co-project lead for the OpenOffice bibliographic project. We really need feedback on the following. I apologize if this is a little long, but the issues are complex.

We want to dramatically improve the bibliographic support in OOo to not just match commercial Word-oriented competition, but to exceed them.

Priorities:

1)  Better metadata.

Current metadata model is based on BibTeX, which is not sufficient for scholars in whole swaths of the humanities, social sciences and law. Better metadata = new GUI.

2)  Better formatting.

We need more functionality (styles like APA demand a lot more than the current system can handle), and we need that functionality to be dynamic. The test of this is whether a user can author a paper in an in-text citation style (e.g. Doe, 1999), and later change that to a radically different style like a footnote-based one with a change of a menu item ... and then change it back; all without modifying document source.

If we do this, OOo instantly becomes REALLY attractive to higher ed.

There are some other goals (like moving to a plug-in API), but let's just focus on these.

As a first step, I and Daniel Vogelheim co-authored a proposal to revamp citation coding in OpenDocument. That proposal was accepted by the TC, but has not yet been implemented.

In this model, citations just become pointers to separate bibliographic records. Internally, the result would look like:

<cite:citation>
   <cite:citation-source>
      <cite:biblioref cite:key="doe99a"/>
      <cite:biblioref cite:key="smith01"/>
   </cite:citation-source>
   <cite:citation-body>
      (Doe, 1999; Smith, 2001)
   </cite:citation-body>
</cite:citation>

The principle is the same as BibTeX, basically.

The content of the body element is dynamically generally from the bib metadata, and can be any doc content, including, for example, footnotes. This opens the door to solving the automatically footnoted issue mentioned above.

OK, so a great first step.

Now, about the rest:

We want to move the bibliographic metadata out of the content.xml file and embed it elsewhere in the file wrapper. Given the recent discussions about RDF in OD, we would like that metadata to be some flavor of RDF, either embedded perhaps in a dedicated element in the meta.xml file, or perhaps in its own dedicate file: let's call it bibliograhy.xml.

Advantages: easy to work with for developers, and easy to extract the metadata (publishers could be really interested in this, but it's convenient for user collaboration too).

The second bit is the formatting. I wrote a complete framework to handle citation formatting in XML, called CiteProc:

http://xbiblio.sourceforge.net/citeproc.html

It has it's own XML style language:

http://xbiblio.sourceforge.net/csl.html

Basically, I want to know how we are going to incorporate this into OOo, and perhaps OD.

There are two issues: styling and processing.

Styling:

The advantage of CSL is precisely its document format independence. I imagine users being able to use the same style files in any document format: DocBook (as I do now), LaTeX, OD ... even WordML.

But that raises the question of exactly how we might integrate that into OOo/OD, which has its own styling system.

The options I see are:

1) Just embed the CSL whole, and don't worry about any integration, or standardization at the level of OD

2) Give up the idea of interoperability across document formats and just figure out how to add some of the principles of CSL to OD (if the TC is willing to consider it of course)

3) a hybrid approach that imports/exports CSL but otherwise uses native styling support

I would emphasize that standardization and document-independence are important gaols, because part of the problem of the bib software landscape is precisely that it is fragmented (and in fact has some of the characteristics of the suite market as of a few years ago; monopoly ownership, proprietary binary formats, tie-in with specific document formats like Word, etc.). I don't want to make that mistake.

Processing:

How could we best integrate a CiteProc-like approach into OOo? What would need to change in SW, in OOo, in OD?

Current code is XSLT 2.0, and was designed first to handle formats like DocBook (e.g. not GUI apps).

I want to know how to move forward on this, as I keep getting conflicting information about who is responsible for what piece of this puzzle.

Bruce

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to