Semantics
The place is all abuzz this week about semantic markup in plain old HTML, and the benefits and practical applications thereof. Mark Pilgrim makes Daypop yet again with a totally cool trick using the cite tag, and then follows it up (semantically!) with the ins tag. Tantek Celik gives some tips on using the class attribute to give meaning as well as structure. With so much semantic potential in regular HTML, Aaron Swartz knows that XML is bad at everything.
On the other side, we have this kuro5hin writer arguing that a semantic web is just a pipe dream, and the semantic tags in (X)HTML are too arbitrary and limited to merit anything useful. Personally, I'm with the anti-XML folks, and strongly on the side of clever HTML tricks.