RE: Auto-detect and encodings in HTML5

> I'm interested in reducing ambiguity and making web transactions more  reliable


> I also would be opposed to making an incompatible change with actual current behavior.


> Yes, supplying explicit charset is preferable

More: one should move toward making it mandatory in the HTTP header. Anything else should be deprecated, but be live in an imperfect world ... 

> New behavior: IF you see, say, <doctype html5> THEN  assume default charset is UTF8 rather than applying heuristics to guess charset. 

UTF8 should be the last option in a set of rules; e.g.,

 - Get if from the HTTP header
 - If not, get if from META
 - If not, ...
 - if not UTF8



Received on Monday, 1 June 2009 20:34:34 UTC