Date: Fri, 23 Apr 2004 22:43:09 +0200
From: Alvar Freude <alvar@a-blast.org>
To: Bruce Momjian <pgman@candle.pha.pa.us>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Cc: PostgreSQL advocacy <pgsql-advocacy@postgresql.org>
Subject: Re: [HACKERS] What can we learn from MySQL?
Message-ID: <289880000.1082752989@gnarzelwicht.delirium-arts.de>
In-Reply-To: <200404230409.i3N49jC02890@candle.pha.pa.us>
References: <200404230409.i3N49jC02890@candle.pha.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

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

Hi,

- -- Bruce Momjian <pgman@candle.pha.pa.us> wrote:

> 	o  Are we marketing ourselves properly?

while talking about MySQL, there is the myth, that MySQL is fast; and that
because MyISAM has no transactions, that it is faster.

That is in most cases not true! And for all real live scenarios I know and I
tested, Postgres was faster.

An example: a critical calculation in one of my online projects needs with
MySQL (MyISAM Table Type) about 2.7 to 2.8 seconds (group by on 500000 rows
for some realtime statistics). But on this time, the complete table is write
locked (because MyISAM) :-(

With InnoDB the same needs at least 15 to 20 seconds, but other users can
insert/update.

With PostgreSQL (7.4) it took 1.9 to 2 seconds. Parallel inserts/updates no
problem.


The only reason why I changed the whole stuff to Postgres yet is, that there
are a lot of problems with MySQL special "features" (see the Gotchas:
http://sql-info.de/mysql/gotchas.html)


Other example: Some days ago I had a talk with my project leader; I said,
that for a new application we should *everything* build with transactions,
referential integrity, ... -- his answer: "I want to have a fast
application". AAARRGH! ;-(


So, perhaps it might be a good idea to create a page with feature- and
performance comparison.
I planed to create an independant and RDBMS benchmark suite (as Free Softwa=
re
including the datas for testing), but I'm not sure if this project ever come
true ...


> 	o  Are we focused enough on ease-of-use issues?

I'm not sure about the focus; but the result can be better.

When installing and using any type of software, I want that this is as easy
as possible while it helps me to understand as much of the backgrounds as
possible.

Whats about the initdb, postgresql.conf and startup scripts?


So, It might be good to have a GUI-Tool (!) in the standard package, which
makes an initdb with selectable options and helps the user to set the
required options in the postgresql.conf.=20

I'm a computer freak since the mod 80s and I can edit config files. But to
have a GUI tool with some explaining texts at the buttons etc is much easyer
than hacking a textfile.


Also the other stuff mentioned in this thread are important: auto vacuum,
windows port, better default values etc.


Ease-of-use includes for me localisation and documentation in different
languages. As you can see, my english is junky -- so reading german
documentation is a lot of easyer for me ;-)


> 	o  Are our priorities too technically driven?

AFAIK it is good to have the priorities technically driven -- if nobody
forgets the userfriendlyness ;)


Ciao
  Alvar=20


- --=20
** Alvar C.H. Freude -- http://alvar.a-blast.org/ -- http://odem.org/
** Berufsverbot? http://odem.org/aktuelles/staatsanwalt.de.html
** ODEM.org-Tour: http://tour.odem.org/
**=A05 Jahre Blaster:=A0http://www.a-blast.de/ | http://www.a-blast.de/stat=
istik/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFAiX/eOndlH63J86wRAjzhAJ0f24+yuOSElI7NmFuChZUH3miWiACdFoH+
OLC0iUn7VP/ZIA30vU8M0tg=3D
=3DRVWf
-----END PGP SIGNATURE-----