Dennis Gearon wrote:
>
> I would like to get have a (living?) drawing/document that shows all the
> interfaces inside of a running postgres installation. This would be a
> first step for me to understand how to do a roadmap of implementing
> fully independent encoding down to the column, sorting per index/column,
> whatever it takes to be able to:
>
> Have multiple charsets/languages AT LEAST at the table table level,
> (some people might be happy with one charset/langauge per table, but the
> spec calls for down to the column), have LIKE, REGEX, indexes,
> mulitcolumn indexes all work well under that scenario, natively, w/o
> having to do extra finaggling that would be a gotcha during restores,
> failovers, server farms,etc.
>
> Maybe a bulleted list that would start with the following:
>
> 1/ Set all names in the DB to be in unicode with automatic
> translation between SQL and outputs.
> 2/ Make the initdb only set the DEFAULT sorting for databases.
> 3/ Make indexes automatically use OP CLASSES for unicode.
>
> I don't know, I'm shooting my mouth off here and I know it. I want to
> learn more, but I'd prefer it was from both the top down and bottom up,
> not just bottom up.
Would the following thread help to get you started?
http://archives.postgresql.org/pgsql-hackers/2003-11/msg01299.php
Jochem (not a hacker)
