threshar@torgo.978.org (Jeff) writes:
> On Wed, 13 Aug 2003, Josh Berkus wrote:
>> As I have said elsewhere, Informix is probably a poor database to emulate
>> since they are effectively an old dead-end fork of the Ingres/Postgres code,
>> and have already been "mined" for most of the improvements they made.
>>
> With informix 7.0 they rewrote the entire thing from the ground up
> to remove a bunch of limitations and build a multithreaded engine.
> so it isn't so much an old fork anymore.
No, I think you misunderstand the intent...
The pre-7.0 version was based on Informix's B-Tree libraries, and the
file structuring actually bears a marked resemblance to that of MySQL
(that's an observation; neither forcibly a good or a bad thing), where
there's a data file for the table, and then a bunch of index files,
named somewhat after the table.
In the 7.0-and-after era, they added in the "old dead-end fork of the
Ingres/Postgres code" to get the "Universal Data Server."
[This is diverging somewhat from "performance;" let's try to resist
extending discussion...]
--
(reverse (concatenate 'string "ofni.smrytrebil" "@" "enworbbc"))
<http://dev6.int.libertyrms.com/>
Christopher Browne
(416) 646 3304 x124 (land)