[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: models-10 comments



At 02:19 PM 3/14/2004, Hallvard B Furuseth wrote:
>models-10 says:
>
>> 1.4. Common ABNF Productions
>>       ; Any UTF-8 [UTF-8] encoded UCS [ISO10646] character

Are you sure you are looking at -10?  

>There is no [ISO10646] in the References section.  Perhaps this should
>be [Unicode], according to Kurt's suggestion in message 'ISO 10646 v
>Unicode' of 07 Jan 2004.

My copy says Unicode here.

>OTOH, UTF-8 is defined in terms of Iso10646, not Unicode, so [Unicode]
>does seem strange here.

[ISO10646] vs [UNICODE] issues are discussed in [RFC3629].

>[UTF-8] should be [RFC3639].

revision -10 does reference [RFC3629].


>BTW, There are 10646's in [Filters] and [Protocol] too.
>
>> 2.5.1. Attribute Types
>
>>     - a subtype's syntax must be the same, or a refine of, its
>>       supertype's syntax, and
>
>What is a refine of a syntax?

s/refine/refinement/ here.

(I don't think we necessarily need to detail precisely what a
"refinement" of a syntax is.)

>> 4.1. Schema Definitions
>
>>       xstring = "X" HYPHEN 1*( UALPHA / HYPHEN / USCORE )
>
>This is a strange change from Models-09, which said
>
>        xstring = X   HYPHEN 1*( ALPHA  / HYPHEN / USCORE )
>
>The X was uppercase only, now it can be either case.

Yes, to be consistent with RFC 2252 (and our discussion regarding
case insensitive).

>The rest could be either case, now it is uppercase only.
>Why such a swap?

Now it can be either case, "X" is case insensitive.

>Editorial comments:

Thanks, Kurt