Proposal: "compat" keywords (BCP47)

[Steven R. Loomis]

Philippe,
 It is indeed ambiguous. But, so are other relative terms such as
"traditional" or "alternate".

 Versioning could be really great. We have not had consensus around a
locale-based versioning scheme, though. It is not the intent of this
proposal to supply versioning. The results of a "compat" collator are
completely version and implementation dependent.

Thanks,
-s

On 07/29/2014 11:12 PM, Philippe Verdy wrote:
> compatiblity with what ??? the specifier is ambiguous if it's not
> followed by another qualifier (and possibly a version number or
> identifier). Then it would require another registry for those qualifiers.
>
> For me it's simpler to use data from a base URL which is identifiable
> and versionable, containing in some root element the use of a small
> identifier used privately within the domain of the URL (in a way
> similar to XML namespaces, allwing free substitution of the small
> identifier but keeping its definition to its defining URL).
>
> So I would just prefer the like form : ar-u-co-xmlns-xyz,
> where "xmlns" means XML namespage, and "xyz" is resolved as a
> namespace wthin the scope of the base URL of the data containing it
> (possibly with the addition of a namespace definition element in the
> LDML document, or some similar mechanism for other database formats
> providing a mapping of the identifier to a more complete definition
> giving details such as a name, description, version, authors,
> licences, possibly in standard RDF format...).
>
> 2014-07-30 7:56 GMT+02:00 Steven R. Loomis <srl at icu-project.org
> <mailto:srl at icu-project.org>>:
>
>     On 07/29/2014 10:53 PM, Steven R. Loomis wrote:
>     > New keyword "compat"
>     >
>     > Example:
>     >
>     >     ar-u-co-compat
>     >
>     > Meaning: "A compatibility collation tailoring, intended to use a
>     > previous version of the collation rules."
>     >
>     > In ticket http://unicode.org/cldr/trac/ticket/4207  a major
>     change to
>     > Arabic collation is proposed.
>     > So that the previous tailoring can be retrieved, the "compat"
>     keyword is
>     > proposed.
>     >
>     > Please reply to this list if you have comments or concerns.
>     >
>     > The ticket # referenced above will be used to implement this
>     keyword.
>     >
>
>     This is a BCP47 keyword.
>


-- 

IBMer but all opinions are mine.
https://www.ohloh.net/accounts/srl295 // fingerprint @ https://ssl.icu-project.org/trac/wiki/Srl