XML Infosets != (angle brackets and UTF-8)

Christian isn’t too enthusiastic about the XML Binary Characterization Working Group. Many others share his opinion and grimace whenever the idea of “binary XML” is floated.

Personally, I’m not opposed to the idea of an optimized binary Infoset representation. Realistically, we already have a binary Infoset representation. It just so happens that the binary representation of XML Infosets happens to be the same as the binary representation of strings. I think there is a more optimal representation out there, and I’m glad there’s a Working Group looking in to the idea.

I think a lot of people have concerns that moving toward binary XML would eliminate the human-readable quality of XML. I don’t think this will be a big issue in the long run. It certainly hasn’t caused problems for strings; after all, Notepad is very good at turning binary data into human-readable strings. No one complains that persisting string data into a standardized binary format decreases readability. Presumably, we’ll create tools that will render a human-readable Infoset with the same degree of transparency and determinacy with which Notepad renders strings. If the standard is solid, I think the tool support will follow and the move to an optimal binary representation won’t affect readability in the long term.

If we do move to a binary XML standard, there will be some pain as the industry adapts to the new standard. Life will be temporarily difficult, but after the transition is complete I think everyone will be a lot happier. The industry has survived transitions like this before (16 -> 32 bit, ASCII -> Unicode), and I think if we go back and look at what made those transitions successful we’ll figure out how to handle the move to binary XML.

Am I being naïvely optimistic?