Re: Thoughts on pragmas and iXML

Steven Pemberton <steven.pemberton@cwi.nl> writes:
> That's why I think the XML PI solution is the best: maximum freedom for the software, minimal hassle for us to try and guess what their needs will be.

I’m pleased that you think XML PIs are a good solution because I think they way that PIs are defined and the way the requirements we have been discussing help to frame the design of pragmas for iXML are closely and sympathetically aligned.

I am a little confused, though, by what you appear to believe are the implications of modeling pragmas on PIs. It seems like you think XML PIs are somehow unconstrained in ways that they aren’t. XML PIs are absolutely part of XML. There are specific places where they can occur, and places where they cannot. Their location in the XML tree is clearly part of the specification and is unambiguious. What’s more PIs have a name, the form of which is explicitly prescribed by the XML specification, and they have pragma data, which is again, prescribed (but is, naturally largely unconstrained).

Very much like what I hope we arrive at for pragmas in iXML.

In fact, if the bare minimum we could achieve was “pragmas like XML PIs”, I could probably live with that. I think we could easily do a little better in the context of iXML and we should strive to do so.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica

Received on Monday, 17 February 2025 16:25:09 UTC