[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Registration of new charset GB18030 (fwd)
Because of the number of people that put dashes in strange places in the names, and because no names are distinguished (point to different character conversion mappings) on the basis of dashes, in ICU we switched to a policy of ignoring all dashes (we ignore case also). That turned out to be much simpler, and might be worth considering for the iana registry.
Mark
___
mark.davis@us.ibm.com
IBM, MS 50-2/B11, 5600 Cottle Rd, SJ CA 95193
(408) 256-3148
fax: (408) 256-0799
Martin Duerst <duerst@w3.org>
Hello Harald,
Comments regarding the Li18nux guidelines should be sent to
localenameguide@li18nux.org.
I fully agree with your comment. I seem to remember to have
sent a similar comment in their first round of comments.
Regards, Martin.
At 06:58 02/07/14 +0900, Harald Tveit Alvestrand wrote:
>--On 10. juli 2002 09:52 +0800 Anthony Fok <anthony@thizlinux.com> wrote:
>
>>Hello,
>>
>>First of all, thank you very much for all your help in adding GB18030 and
>>GBK to the registry.
>>
>>After reviewing Li18nux locale naming guidelines, I wonder if it would be
>>desirable to add "GB-18030" as an alias for GB18030,
>>and "GB-K" as an alias for GBK? Many thanks! :-)
>
>In general, I think aliases are a thing to be avoided.
>More names for one thing just means that there is more software that
>understands one name, but not another. And more bugs is not a Good Thing.
>
>>
>> http://www.li18nux.org/localenameguide/
>> http://www.li18nux.org/docs/html/CodesetAliasTable-V10.html
>
>I think you refer to this as the reason:
>
> The standard values for the CODESET field shall consist of multiple
> strings exclusively containing LETTERS or NUMBERS in conjunction
> with the delimiter '-'.
> The syntax of the field in ABNF [RFC 2234] is:
>
> CODESET = STRING1 *( "-" STRING2 )
> STRING1 = 1*LETTERS
> STRING2 = 1*(LETTERS / NUMBERS)
>
> STRING1 shall consist of uppercase LETTERS only.
> STRING2 shall consist of uppercase LETTERS, NUMBERS, or both.
>
>First, I think the li18nux.org people have made a stupid decision.
>They should not be redefining charset names, but using someone else's - if
>that someone is the IETF, they should allow all legal IETF charset names.
>If you can tell me who to say that to, I'll be happy to send them a note
>saying so.
>
>Second, there is no reason to change GBK based on this document; GBK is a
>perfectly good STRING1, and STRING2 is optional.
>
> Harald
>
