[TYPO3-core] Fixing no-brainers?

Dmitry Dulepov dima at spamcop.net
Mon Apr 10 14:55:00 CEST 2006


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi!

Martin Kutschker wrote:
> I was thinking more in terms of layout plus a separation of
> functionality and layout. So we can have a full-blown UI as we have
> it now, but easily a streamlined, scalable UI which is hopfully fully
> accessible. If it's done with AJAX fine, but I don't care as long
> it's fast :-)

Aha. I see many speak about UI changes but seems that there is no common
understanding what it means.

Ajax implementation should be faster. I made several experiments for the
same functionality with traditional approach and using ajax. Second
visually looks faster. This is why I advocating it :)

>> And we can require mysql 5 because it supports views, stored 
>> procedures, etc. It will give major performance increase. But this 
>> should be evaluated from DBAL view also.
> 
> View are fine but does a DB-layer need to support them? Stored
> procedures are not standardized.

Well, if I am not mistaken ADODB can even emulate them for servers that
do not support them.

> I was thinking more about nested-queries, unions and transactions.

For trasactions we need innodb in any case. Btw, I converted several
tables to innodb on my servers and see no locking any more for various
tables (like sessions, index and sys_stat). I think it can be a part of
performance subproject to investigate whether and how we can use innodb
effectively. Not all tables should be innodb.

But nothing prevents us from using "START TRANSACTION" and "COMMIT"
right now. We only cannot use "ROLLBACK" :)

Dmitry.
- --
"It is our choices, that show what we truly are,
far more than our abilities." (A.P.W.B.D.)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (MingW32)

iD8DBQFEOlWkRVcYnKJ8N6kRAuAsAJ4geyFOeQPbLfD1i8gH1Q3/Pjx78gCgmj+d
JuguUsXdH21I+tF7fuXjj90=
=cytW
-----END PGP SIGNATURE-----



More information about the TYPO3-team-core mailing list