> > I'd suggest that you dont do this... Adding support for=20
> > ISO-8859-1 directly into H::T will set a precedent for other=20
> > encodings (not everyone uses the Latin character set)...
>=20
> Unfortunately, the precedent is that H::T generates broken,
> non-compliant HTML. The spec says that anything that's not
> 7 bit ASCII needs to be encoded. Latin1 is the common
> denominator.
Which spec says that anything other than 7bit ascii needs to be encoded? =
I'm not sure that I understand which spec you are refereing to.
I understand that at minimum the a HTML document should specify the =
document encoding (or at least a META tag with the HEAD of the =
document). If no encoding is specified, then the browser can assume =
that it is encoded in ISO-8859-1.
> > Why not just output the text as UTF8?
>=20
> The infrastructure we're building on supports Latin1, but
> not UTF-8. Fixing H::T to generate compliant bits is simple;
> reworking 20 person years of code to do UTF-8 isn't.
> Nothing in the patch precludes using other encodings.
Thats true, but if we add support for encoding to Latin, should we then =
do that for every other encoding?
regards,
Mathew
|
