Look, if I say something I can see it being ignored, but when Dorathea who has both a better knowledge of all this and A-List status says it, and people ignore it also, it’s just silly. So, what do we think will happen if other people start with the same sentiments?
Well, they’ll be ignored still. Please don’t bastardize the XML forms. I mean, make it RDF if you want to, make it it’s own vocab and DTD if that is what you feel, but please, please for the sake of my sanity as a coder, don’t allow people to randomly mix up any vocabularies they think they might like! No piece of software can nor should have to parse any random form of XML, be it past, present or future. The ability to add cool things is all very well, and I can even live with “Hang all backwards compatibility”, but the current temptation to allow anything in any documents means that in order to process any single document I’m either going to have to work with some kind of Uberparser which takes in XML to it’s wide gaping maw to spend a week digesting and being shat out as something I can actually understand; or I’m simply going to have to ignore any elements I don’t understand (Which could be categories, authors, extras or equally the content itself).
I have a complex RSS2 feed, it contains data from RSS 0.9* all the way up to the latest and greatest in RSS modula, from the number of comments to the trackback URL. If there is a reader out there which understands even 90% of the tags in it, I’ll be shocked and stunned.
I’m in favour of loosly defined data formats for internal use, for passing between known systems. But for a defined internet specification to rival SMTP, HTTP, NNTP et. al (and that’s what Atom/PIE/WOX could be if it works) you simply can’t give it this wide a canvas, it’s simply too much for parsing it on a major scale to be economical.