|
"XML has failed" is obviously over the top as a statement, but at the same time it makes perfect sense -- from a certain point of view...
Pilgrim-san has been told, repeatedly, that writing software that can cope with purported XML entities that are not well-formed is wrong, because XML mandates that things that are not well-formed MUST be rejected. If people are allowed to get away with malformed XML documents, then XML will fail, we are told. All XML generators must be forced to stick to the letter of the published standards, by refusing to let XML consumers consume them.
So he points out that, in fact, many, if not most, XML documents on the WWW are technically malformed, which means that if the anti-Postel arguments were to be taken seriously, XML would indeed have failed. Or, contrapositively, I think his point might be that XML is not the total exception to Postel's law that some people would like to claim it is.
It's all very well to simply respond that people who serve XML as text/xml are foolish and wrong, but this does not address the reasons why people *do* end up with web servers serving XML as text/xml: we were to... by the XML specification. Given that XML processors are supposed to consume byte streams not character data, allowing text/xml in the first place was probably a mistake. :-(
If Mark Pilgrim's outburst reminds a few more web server developers to expunge text/xml from their look-up tables, then this is surely a good thing, yes?
|
Previous Message
