[wp-hackers] GSoC Proposal: Multiple RDBMS support

[Olexandr Melnyk]

On 3/31/08, Alex Günsche <ag.ml2008 at zirona.com> wrote:
>
> We had this topic a dozen times (please read the archives[1], guys!),
> and I believe that the consensus was that multiple RDBMS support is not
> a good idea, mostly because of the fact that it is  preferable to use
> the advantages of certain MySQL features over portability to DB systems
> that don't meet most of the target audience anyway.


I understand your point, but I don't see how this is related to WordPress. At
this moment, WordPress is supporting MySQL versions starting from 4.0,
what means that even half-advanced features like stored procedures and
triggers can't be (and
thus, aren't) used.

Another fact is that
> it would also requite plugin authors to write cross platform plugins.
> This is not only very hard, because you can't expect them to test on all
> major DB systems, but you would again dramatically limit possibilities
> (like e.g. using FULLTEXT indices) where it's actually not required.


Fulltext indexes
are supported by all major RDBMS'es, but with a different SQL syntax, so a
common wrapper should work in this case.

I'm aware of two
MySQLisms used by WordPress. One are fulltext indexes, which aren't
really MySQL-specific, as I have pointed above, but have
a more varying
syntax for lookups among databases than other common SQL statements.

Another are enum columns, which on other databases can be replaced by
varchars with appropriate domains. Abstracting this shouldn't affect any
code, except for the table creation SQL dump.

-- 
Sincerely yours,
Olexandr Melnyk <><
http://omelnyk.net/