P4 to P5

[Syd Bauman]

> I agree with Sebastian, both in not really grocking the
> significant difference between a URN and a URL, and in preferring
> a simple numeric identifier to one which doubles up as a date.
> For one thing, at least 3 digits of the date are going to be
> redundant.

I still prefer the date-like system, but am willing to concede. But
we do need to address what, exactly, does and does not cause an
increment of the various components of the version number. (E.g., if
a change is made to an ODD file that has an effect only on the
Guidelines output, but not the schemas, is the version number
incremented? How about if only the schema, but not the Guidelines
gets updated? What if no ODD files were changed at all, but only the
stylesheets that generate the output(s) were changed?)