[whatwg] Drop UTF-32

[Michael Day]

Hi,

Suggestion: drop UTF-32 from the character encoding detection section of 
HTML5, and even better, discourage of forbid user agents from 
implementing support for UTF-32.

Why:

  - It's not widely used. In fact, has UTF-32 ever been used at all, 
outside of test suites?

  - It's not widely implemented. For example, the expat XML parser does 
not support it, and nobody cares.

  - When it is supported, people get it wrong, and the bugs are never 
fixed because no one uses UTF-32 anyway and no one cares.

For an example of this, see html5lib 0.9, which implements the BOM 
detection algorithm, but gets it wrong by checking for UTF-16 before 
checking for UTF-32. Because the UTF-16 BOM (FF FE) is a substring of 
the UTF-32 BOM (FF FE 00 00) the test will always succeed and UTF-32 
will always be misidentified as UTF-16. But no one cares, as no one uses 
UTF-32 anyway.

  - UTF-32 is horrendously inefficient for just about all real world 
text and its use should not be encouraged on the web. Really, UTF-32 
only exists as a tutorial example of how UNICODE can be encoded, not as 
a practical character encoding that people should actually use.

Please, drop UTF-32 and save implementors from worrying about it when no 
one uses it and no one should use it.

Thanks,

Michael

-- 
Print XML with Prince!
http://www.princexml.com