RE: Registration of new charset GB18030 (fwd)

[Francois Yergeau <FYergeau@alis.com>]

Lars Marius Garshol wrote:
> The character encoding identification code in the Opera web browser
> does precisely the same thing (it also ignores underscores). We had
> lots of trouble with people mixing up dashes and underscores, and
> inserting them in unexpected places, and did this to reduce our
> ever-increasing list of aliases.

We did the same in the Tango browser c. 1996.  IIRC, we had a three-step
matching procedure:

1) try an exact (but case-insensitive) match
2) try again, ignoring hyphens, underscores and periods
3) try again, ignoring any "x-" prefix

We never had any complaint.

-- 
François Yergeau