RE: [office] Action required: OpenDocument v1.0 Second EditionDraft 28 (revised)

["Mason, James David (MXM)" <masonjd@y12.doe.gov>]

Rob says:

If we do this, I'd like us to also write up and adopt a formal deprecation policy for the TC -- what does it mean, how is it decided, is support ever removed entirely, etc.  Deprecation is a powerful thing and useful for maintaining the health of the specification.  But like forest fires, we need to consider carefully how and when we use them.

As Patrick knows, I'm very cautious about things that impact backwards compatibility. It's necessary, particularly when you have a new standard, and you know there might be bugs in it, not to break things for early adopters. But our experience with SGML was that backwards compatibility became a stumbling block to updates, so that the developers of XML picked up and moved to the W3C rather than stay in ISO and constantly run into a brick wall. (Of course some of them then moved from the W3C to OASIS and then brought their work back to ISO in the case of RELAX NG.)

 

I've argued in favor of backwards compatibility to XTM 1.0 in XTM 2.0, but I don't think I'd argue for such things in SGML any more (I just sent Patrick the source of an ISO Technical Report that was marked up in GML, and I don't think anybody would argue in favor of supporting some of the constructs there today).

 

Jim Mason