- From: C. M. Sperberg-McQueen <cmsmcq@blackmesatech.com>
- Date: Wed, 06 Apr 2022 16:45:17 -0600
- To: public-ixml@w3.org
The CG's decision to abandon our work on pragmas leaves those of us working on a Balisage paper presenting the pragmas proposal with a slight dilemma. Either (a) the paper becomes a post mortem report describing was proposed but not adopted by the group, explaining the fine points of a design in which the fine points no longer matter, or else (b) the paper becomes a description of an add-on to ixml: not part of standard ixml, so not guaranteed to be supported by all conforming ixml processors, but an additional facility which may or may not be supported by processors, just as the functions in the EXSLT namespace are not part of the XSLT spec but are supported by some XSLT 1.0 (and 2.0 and 3.0) processors. This is to place on record that I think (b) is the preferable option, and to invite any members of the CG who agree that (b) is preferable to collaborate with me to refine the pragmas proposal as an extension layered over standard ixml. To ensure that pragmas are not rejected by standard processors, it will be necessary to change the pragma delimiters to include "{" + something and something + "}"; I regard that as unfortunate, but since I don't particularly want to fork the ixml effort it's a price I am willing to pay. In the proposal as I expect to describe it, pragmas will thus be structured comments, just as in other languages where structured comments are an unfortunate and ugly hack made necessary by the failure of the initial specification of the language to provide a distinct syntax for pragmas. The one thing we can do differently in the ixml ecosystem is for those who see the need for pragmas to agree at the outset on common implementation-independent rules for handling pragmas in the form of structured comments. The CG has not been able to achieve consesus on such rules, so I propose that we use the Balisage paper on the pragmas proposal to sketch out a proposal. If a significant percentage of the initial crop of ixml processors implement pragmas as defined in the Balisage paper, then we may be able to set expectations among ixml users that implementations should not just roll their own syntax for structured comments, but should use an implementation-neutral syntax designed to avoid collisions. That won't be binding on future implementors, but it may encourage them to do the right thing. CG members interested in working with Tom and me to reformulate the pragmas proposal as a separate spec (an ixml profile, in effect) layered over ixml are invited to let me know by email. Michael -- C. M. Sperberg-McQueen Black Mesa Technologies LLC http://blackmesatech.com
Received on Wednesday, 6 April 2022 22:45:37 UTC