Re: IRIs, IDNAbis, and HTTP [i74]

Brian Smith wrote:
 
> <phrase> is not used at all in RFC 2616.
> <text> is not used, there is <TEXT>.

Yet another reason to fix RFC 2616.

> Adding a new requirement for implementations
> to support RFC 2231 in any existing headers
> is not a compatible change. 

We both missed the RFC 2231 Last Call, it says
"updates: 2045, 2047, ...".  With 2231 < 2616
folks could have noted these problems earlier,
maybe there is no problem.

> Atom and HTML 5 (will) allow internationalized
> link relation names, and IRIs as link relation
> names.

Atom and HTML 5 will find protocols to transport
them, if it's not HTTP/1.1 that's their problem.

A relation is what RFC 2068, 2070, etc. say:

| sgml-name = ALPHA *( ALPHA | DIGIT | "." | "-" )

In an older HTML DTD I found a reference to...
http://tools.ietf.org/html/draft-ietf-html-relrev
...and that draft stated:

| The REL and REV attributes are defined as NAMES
| in the SGML DTD for the HyperText Markup Language
| (HTML). As such, the legal values which may be
| assigned to the REL or REV attribute are zero or
| more name tokens. Name tokens are case-insensitive,
| must begin with an alpha character, may include
| digits (0-9), period or hyphen, and may be 
| separated by spaces.

RFC 2068 got it right.  I can't tell why HTML 4.01
says CDATA instead of NMTOKENS.  It could be a bug
inherited from 3.2, after the HTML i18n (RFC 2070)
bug "NAMES" instead of NAMES (I think, sometimes
DTDs still confuse me).

But it clearly never was an URI, let alone an IRI.

 Frank

Received on Saturday, 15 March 2008 17:07:07 UTC