W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1996

Re: (ACCEPT*) Last call on draft text for Accept headers

From: Koen Holtman <koen@win.tue.nl>
Date: Fri, 5 Apr 1996 20:36:52 +0200 (MET DST)
Message-Id: <199604051836.UAA00545@wsooti04.win.tue.nl>
To: greenwd@openmarket.com
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, koen@win.tue.nl
Tim Greenwood:
>
>I propose a minor change to the text in 10.4 and question the multiple
>definitions of matching.

>>From 10.4
>> |  A language-range matches a language-tag if it exactly equals the tag,
>> |  or if it is a prefix of the tag such that the first tag character
>> |  following the prefix is "-".  
>
>Section 3.10 defines a language-tag as
>
>        language-tag  = primary-tag *( "-" subtag )
>
>        primary-tag   = 1*8ALPHA
>        subtag        = 1*8ALPHA
>
>It is preferable to reword the above section in 10.4  to
>
>"A language-range matches a language-tag if it exactly equals the
>tag,  or if it equals the primary-tag (see 3.10)"

I believe your proposal for simplification is based on a misreading of
the syntax definition of language tags.  The current rule needs to be
this complicated because there can be more than one subtag.

The matching rule currently defined will allow the range "i-sami" to
match the tag "i-sami-da".  Your proposed simplification will not
allow this.

I believe all tags currently defined by RFC1766 (Language tags,
proposed standard) have at most one subtag, but a future revision of
RFC1766 may define language tags with three elements:

   i-sami-da  (Harald Alvestrand, the author of RFC1766, is working on
               things like this)

   i-s-bok    (and many other i-s-??? tags, where ??? is a 
               Summer Institute of Linguistics' Ethnologue 3-character
               code for a language)

The rule in the proposed text is supposed to be compatible with such
additions to RFC1766, if they are made.

>-------
>
>The rules for matching language-range with language-tag are currently 
>addressed in three sections of HTTP 1.1 - and they are not the same. 

This is true: the planned edits will resolve the mess.  The matching
text of section 3.10 will be removed, and the matching text of section
12.1 will also be removed, because the complete chapter 12 will be.
This leaves the one matching rule in the proposed section 10.4.

>The rules should be defined in one section only.

>Tim Greenwood        Open Market Inc
>617 679 0320         greenwd@openmarket.com

I hope I was able to address your concerns.  I will assume that the
information above takes away your problems with the current proposed
text, unless I hear from you otherwise.

Koen.
Received on Friday, 5 April 1996 10:43:06 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:50 EDT