Re: [VCARDDAV] SORT-STRING [was:vcardrev nits]

[Peter Saint-Andre <stpeter at stpeter.im>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 9/16/09 7:25 PM, Javier Godoy wrote:
> 
> On Wednesday, September 16, 2009 12:14 PM Peter Saint-Andre wrote:
> 
>> On 9/15/09 10:24 PM, Javier Godoy wrote:
>>> Peter Saint-Andre wrote,
> 
>>>> 7.7.5. SORT-STRING
>>>>
>>>>   locale- or national-language- specific sorting
>>>>
>>>> This is hard to parse. I suggest:
>>>>
>>>>   The sort string is used to provide family name or
>>>>   given name text that is to be used in sorting of
>>>>   the formatted name and structured name types in the
>>>>   context of a particular locale or national language.
>>>
>>> I don't agree with the "context of a particular locale or national
>>> language" part, as it seems to suggest that the collation is well-known
>>> and defined somewhere else.
>>> I would modify it by "an implied particular locale or national
>>> language."
>>
>> That's better, yes.
>>
>>> Besides, the current special note "The sort string is used to provide
>>> family name or given name text..." is ambiguous. How could the
>>> sort-string be of use if one cannot tell whether it applies to the given
>>> name or the family name.
>>>
>>> Example 1
>>> FN:Rene van der Harten
>>> N:van der Harten;Rene;J.;Sir;R.D.O.N.
>>> SORT-STRING:Harten
>>>
>>> Example 2
>>> FN:Rene van der Harten
>>> N:van der Harten;Rene;J.;Sir;R.D.O.N.
>>> SORT-STRING:Rene
>>>
>>> In example 2 were specified (which is valid according to the
>>> definition), I would collate "van der Harten" under R.
>>
>> Correct. But that would be stupid. :) We can't save implementers from
>> their own stupidity, but I suppose we can help them understand that the
>> SORT-STRING is used for alphabetical sorting of family names containing
>> multiple words.
> 
> I agree it would be stupid, but it is allowed by the definition!. I
> think there is an implicit agreement on using SORT-STRING for specifying
> a collation key for the family name, despite of the definition allowing
> either "family name or given name". In order to solve this ambiguity,
> the purpose of SORT-STRING should not mention the given name (which is
> not going to be specified there).
> 
> I propose (from the text  you suggested above):
> [[
> Purpose:  To specify the family name text to be used for
> national-language-specific sorting of the FN and N types.
> 
> Special notes:  The sort string is used to provide family name text that
> is to be used in sorting of the formatted name and structured name types
> in the context of an implied locale or national language. Without this
> information, sorting algorithms could incorrectly sort this vCard within
> a sequence of sorted vCards. When this property is present in a vCard,
> then this value, instead of the family name, is used for sorting the vCard.
> ]]

Works for me.

By the way, what is a "national language"? It's not clear to me if that
term has any meaning, because it conflates political jurisdictions and
language communities (e.g., the German language is spoken in Germany,
Austria, Switzerland, parts of Belgium, etc. -- so what does it mean to
say that German is a "national language"?). I think that the phrase
"implied locale or language community" is more accurate, or even
"implied culture or language community" given that onomastics (the study
of proper names) combines aspects of both language and culture but is
not properly speaking a matter of locale.

Peter

- --
Peter Saint-Andre
https://stpeter.im/


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkqxky4ACgkQNL8k5A2w/vxA9wCgp2yne9vA01dT2sU0fGP9d5NV
+qIAoNEHA3/5Jf6GYS8yMaVXpy7N4F2L
=V12F
-----END PGP SIGNATURE-----