[Python-Dev] r83893 - python/branches/release27-maint/Misc/ACKS

[Terry Reedy]

On 8/10/2010 9:13 AM, Benjamin Peterson wrote:
> 2010/8/10 Stephen J. Turnbull<stephen at xemacs.org>:
>> Benjamin Peterson writes:
>>   >  2010/8/9 Nick Coghlan<ncoghlan at gmail.com>:
>>   >  >  On Tue, Aug 10, 2010 at 2:10 AM, alexander.belopolsky
>>   >  >  <python-checkins at python.org>  wrote:
>>   >  >>  +PS: In the standard Python distribution, this file is encoded
>>   >  >>  in UTF-8 +and the list is in rough alphabetical order by last
>>   >  >>  names.
>>   >  >>
>>   >  >>    David Abrahams
>>   >  >>    Jim Ahlstrom
>>   >  >>  @@ -28,6 +29,7 @@
>>   >  >>    Éric Araujo
>>   >  >>    Jason Asbahr
>>   >  >>    David Ascher
>>   >  >>  +Peter Åstrand
>>   >  >   From my recollection of the discussion when Peter was added, the
>>   >  >  >first
>>   >  >  character in his last name actually sorts after Z (despite its
>>   >  >  resemblance to an A).
>>   >  This is correct. Don't think of Å as a kind of "A". It's its own
>>   >  letter, which sorts after Z in Swedish.
>>
>> That's true, but IIRC there are a fairly large number of letters where
>> different languages collate them in different positions.
>>
>> Is it worth actually asking appropriate humans to think about this, or
>> would it be better to use Unicode code point order for simplicity?
>
> I think it's largely a unimportant discussion. If people have an
> opinion of where their name should appear, they can by all means
> change it. However, "rough" is probably as best as it'll ever get.

If I were committing a patch and was checking to see whether a name that 
started with a decorated A (or any other letter) were already in the 
list, I would look in the appropriate place in the A (or other) section, 
not after Z.

Everyone working on the English-based Python distribution knows the 
order of the 26 English letters. Please use that order (including for 
decorated versions and tranliterations) instead of various idiosyncratic 
and possibly conflicting nationality-based rules.

For instance, suppose a 'Jean Charbol' posts a patch? Should we really 
have to ask his/her 'nationality' before adding the name to the list? 
Suppose 'Charbol' was born in Spain but works in France? In Spain, at 
least, 'ch' words are alphabetized in dictionaries between 'c' and 'd' 
words. Did everyone already know that? I an mot ever sure if all 
Spanish-speaking countries still do that.

I am under the impression that either the Irish or Scots have some fussy 
rules for Mc/Mac/O names but I don't know them and don't think we should 
observe them in our list.

Librarians who filed author cards by birth nationality rules made the 
now-obsolete card catalogs less useful for users who not know both birth 
nationality and rule. Lets not repeat that mistake.

-- 
Terry Jan Reedy