Re: Help me improve the 9.1 release draft

I'm still learning about this feature but wanted to put up some information
for discussion.

In the official Pg press release, the feature is still called "Per-Column
Collations." Is this common enough of a term that people will know what
benefit they will derive? As a counterpoint, Oracle has a chapter in their
Globalization (G11n) guide titled "Linguistic Sorting" which goes into
detail on how linguistic sorting is done; however, it only mentions
"collated" once. Is this a valid comparison?

http://download.oracle.com/docs/cd/B10500_01/server.920/a96529/ch4.htm#1008649

The guide talks about both monolingual and multilingual linguistic sorts.
Sorting on Unicode is considered monolingual. Language-specific sorting is
available as both mono- and multilingual. Multilingual sorting (e.g.
FRENCH_M vs FRENCH) uses the ISO 14651 multilingual standard.

From the discussion, my impression is that the Pg feature provides
monolingual, language/locale-specific (non-Unicode) sorting. Is any other
benefit provided?

So two items:

(1) Would the feature be better listed as "Per-Column Collations" or
"Linguistic Sorting"

(2) Is the primary benefit monolingual, language/locale-specific sorting.

John

On Thu, Aug 25, 2011 at 6:53 AM, Michael Banck <michael(dot)banck(at)credativ(dot)de>wrote:

> Am Donnerstag, den 25.08.2011, 16:47 +0300 schrieb Peter Eisentraut:
> > On ons, 2011-08-24 at 17:00 -0700, Josh Berkus wrote:
> > > Here's a considerably expanded and improved 9.1 release:
> > >
> > > http://wiki.postgresql.org/wiki/91releasedraft
> > >
> > > I'd like to get this to the translators in the next 48 hours, so please
> > > send me last-minute corrections!
> >
> > This looks a bit funny:
> >
> > """
> > Linguistic Sorting using Collations: Improve multilingual (M17n)
> > correctness and user-friendliness with proper and consistent linguistic
> > sorting rules. PostgreSQL linguistic sort supports monolingual and
> > multilingual text and is often requested for sorting of diacritics.
> > Sorting can be performed using language/locale-specific rules or using
> > UTF-8 for multilingual sorting. For flexibility, collations can be set
> > on columns, domains, indexes and expressions.
> > """
> >
> > I don't know how one would get from "multilingual" to "M17n". Also, we
> > don't support "multilingual text", which is text that consists of
> > multiple languages(?). Also, you can't sort "using UTF-8".
>
> Heh, I wondered the same and pointed that out to Josh on IRC earlier
> today.
>
> M17n expands to "Multilingualization"; I do not think that is a common
> enough expansion that people won't think "how do they get from
> multilingual to m17n??", so either adding that in parenthesis, or just
> using that instead of "multililngual" (if grammar allows) would be
> better IMO.
>
> Whether or not it is applicable at all is another matter, of course.
>
>
> Michael
>
> --
> Michael Banck
> Tel.: +49 (0) 2161 / 4643-171
>
> credativ GmbH, HRB Mönchengladbach 12080
> Hohenzollernstr. 133, 41061 Mönchengladbach
> Geschäftsführung: Dr. Michael Meskes, Jörg Folz
>
>
> --
> Sent via pgsql-advocacy mailing list (pgsql-advocacy(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-advocacy
>