On Thursday, March 20, 2003, at 05:18 AM, Andrew Savory wrote:


I don't know if there are issues/dependencies related to changing the
DTDs though (maybe related to Forrest?).

Just taken a look at Forrest DTDs, and I can't see any deprecated element
or attribute ... the closest is "warning". So I guess we'd need to add it
to both cocoon and forrest ...

Thoughts:

1. Please let's not add anything new capabilities to our currently used document-v10 dtds. Last I checked, all our v10- docs translate to doc-v11 with existing transition tools. Sorry to sound like a broken record, but we need to either transition to Forrest first or add this capability to the transition trial files currently located in the Forrest cvs. That is, build a prototype of a "transitioned" Cocoon docs (which will reveal a few outstanding problems which you are welcome to fix), copy over your suggested changes, and go/report/advise from there.

2. It's my impression that supplementing the v11 dtds is an ongoing/unresolved issue at Forrest. You might get more information if you discuss it on the Forrest list. Seems to me there are different approaches:
a. Ssk forrest to change a particular dtd -- or add it to the dtd changes wishlist, (sorry, can't find the link)
b. Make our own more expressive dtd which then gets transformed in an intermediate step to doc v11 or xhtml2 before the final "presentation" processing step


3. In some ways, marking content with deprecation tags feels like the wrong approach. Technically, these docs are reference pages which could be generated in a uniform way, with deprecation status etc. pulled from the source code, in this case, src/java/org/apache/cocoon/components/language/markup/xsp/java/request.xsl.
We discussed this previously for java source files, not sure how it would work in the above case with the logicsheets.
I'm not against this type of approach, or course, for docs like user manuals, how-tos, etc. Still, I wish we could implement the above need by looking up the components/code referred to by the content, checking for deprecation status there, instead of manually inserting this info in docs, but that probably would probably kill doc-generation performance.


Thanks.

Diana



Reply via email to