phil ringnalda

I’m afraid Robert’s right that “silent data loss” (the way aggregators misinterpret things including the less-than character which aren’t intended as markup as being markup, and either strip it or let browsers interpret it rather than display it) is actually worse now than it was before. But, there’s one difference: even though treating everything the same in Atom gives you a one in three chance of being right, where in RSS it gave you a one in two chance, in Atom you have no reason to treat it all the same, no excuse for treating it all the same because it tells you what it is, and when it’s misinterpreted there’s no doubt about who got it wrong. If you don’t treat what your XML parser hands you as being what the type attribute says it is, text, html, or xhtml, then you aren’t pragmatic or reasonable, you are wrong.

So, look out windmills aggregator developers, I’m strapping on the rusty armor, ready to have at ye. From now until the ambiguous time when “I’m satisfied” all my titles will include at least one less-than character, whether they need it or not. Right now, I’m doing possibly the very easiest thing, type="html" in a CDATA section, but if push comes to shove I can always switch to something more difficult.