Fake Non-Live Coverage of XML 2005

Inspired by
>Kurt Cagle's live blogging of XML 2005 in Atlanta or perhaps San Diego. Except I wasn't there, didn't see it, and am just looking over some slides! Here are my random virtual thoughts, without the benefit of jet lag, and without the distraction of hearing the speaker actually explain their bullet points.

>Sam Ruby's talk is pretty daft early on. He seems to feel that standards should be made so that people can implement them by telepathy: if a hacker needs to actually pay attention to what a standard says (e.g. read the thing) it is a bad standard.

>For example, someone's XML Schemas tool does not implement XML Schema's boolean correctly and converts "1" to a Java "false": Sam apparantly thinks that is XML Schema's fault. Similarly, whenever a parser has something implemented wrong, it is XML's fault...Many times the same pattern comes up...if someone writes non-well-formed XML and the parser complains, that seems to be XML's problem...if someone uses incompatable and wrong encodings, XML's problem not theirs (no mention of application/* versus text/*, hmmm)...that you cannot feed any arbitrary XML to HTML readers seems some kind of shortcoming or gotcha too to Sam...

Later...is Sam pulling a rabbit out of the hat? Ada's definition of erroneous: interesting idea to apply this to the XML stack. Sam's example is XML Schema's datatypes precision requirements. No, he's completely lost me here: it looks like he is discussing XML Schemas, but he actually discussing Java. I really don't get what he on about here: I think he is saying that because different platforms use different precisions (and may truncate incoming data etc), chaos is stalking the land and therefore XML is hard.

One good link in the talk was to Joshua Allen's comment on MS and RFC 3023. It is high time that this was explicit, so that we all can build our systems so that text/html is for friendly HTML that kind of muddles through and application/xml is for rigourously marked-up (and encoding-labelled) XML that attempts to prevent bad data or documents propagating and fails early. In that sense, HTML should be easy and XML should be hard.

Joe Gregorio give astraightforward introduction to Atom.

Good to see the examples use application/atom+xml. Browsing around...apparantly the schema for the Java web applications web.xml don't allow + in a media types, such as such as application/xhtml+xml. This must be another example of the programmer not reading the spec for characters allowed in MIME media types; or is it more evidence for Sam Ruby's case?

I see Josh Lubell from NIST is giving a slide on standardized naming rules, but I cannot find the slides on Google. I suppose this is part of their XML Schema validation Process for CORE.GOV work (also in PDF)

Another paper I would really like to see is Analysis of XML schema usage (Ralf Lämmel and Stan Kitsis and Dave Remy) which relates to schema metrics. I presented on my Document Complexity Metric at Cambridge UK last year and Sydney this year, and have been getting good feedback on other factors that people have found useful. As XML matures, more attention will be paid to metrics.