Still need details about how we'll handle versions

[Klortho]

This is really a follow-on to #9.

I wrote up and added a couple of detailed procedures to the workflows wiki page:

• Detailed procedure for new topics
• Detailed procedure for releasing new recommendations

Here's a survey of the places where versions show up:

• In the URL of our schema files. For example, "http://jats4r.org/schema/0.1/jats4r.sch". See JATS4R schema processing instruction
• In the corresponding location for the generated XSLTs. The validator has to know which version of our schema to use
• In the bin/process-schematron.sh script, which has to know where to get its input, and where to write its output.
• Tags in our GitHub repository. This should be the canonical way to identify a version of recommendations/schema. IOW, whatever is in the repository at the commit pointed to by the tag is, by definition, that version.
• The recommendations pages on our website should have a version number attached to them. Our website should have a complete set of the recommendations at each version, with an easy way of navigating between them, and also a set without a version number in the URL, that corresponds to the "current" or latest version. Question: what to do with a particular page (e.g. math) if it doesn't change at all when a new version of the recs is released?
• In the validator: it has to be able to read the processing instruction, and select the appropriate XSLT files to run. See Parameterize the version number #51 .

In addition to the "Detailed procedure for releasing new recommendations", I think we also need a way to do a "bug fix" release, where there's a minor change or amendment, that doesn't affect the official version number. We need to make sure these are backward compatible.

[Klortho]

I'd like to do a "dry run" of these procedures, and put the changes in place for handling versions, using as a dry run the draft recommendation about the JATS4R schema processing instruction, and related issues (#51 and #52).

[Klortho]

I moved this to the validator issues list, here, since it's too detailed for most folks to be interested in.