Re: RE: Why is there so much MySQL bashing???

carl garland wrote:
>
> I think alot of the perceived bashing from the pg community is due
> to the fact historically most of mysql's marketing and pr has
> been deceptive and outright wrong.  After so much crying wolf it is
> hard for the village to believe anything the boy has to say.
> The majority of open source db installs are running on what the
> pg community percieves to be an inferior product.  Granted it is
> *our* perception but if you look at the breakdown of users of mysql
> vs postgres it tends to be professional programmer prefer to use
> pg.  This is due to the fact that have a more concrete understanding
> of the two underlying architectures.  Mysql is a fine product for
> the majority of users. Why because it works, is easy to work with,
> and the momentum is with them, but if you are a looking at a more
> rock solid product with a codebase that will approach the demands
> that mission critical Oracle like users need PG is vastly closer and
> will be the only choice that will most likely perform up to the
> standards that an enterprise solution demands.
> These opinions are my own but the more informed programmers (PHDs and
> such) will probably agree.

Sorry to add to the noise, but I must add that, to the best of my
knowlege, MySQL currently (my last look on their pages, especially the
"crash-me" pages, was about 2 weeks ago) has :
    - No transactions
    - No views
    - No subqueries
and is therefore an excruciating pain in the *ss to use for intricated
problems (databases using a lot of tables whith highly irregular
structures).

Furthermore, various reports show that for simple problems
(simple-structure databases), MySQL has a better performance under light
loads, but that PostgreSQL scales bertter (the degradation of
performance with load is much slower). This information I have
secondhand, so I cannot commit myself on it.

In other words, it seems that MySQL developpers favored high simple-case
performance over competence, thus missing the whole point of an RDBMS
over an indexed file collection : the ability to find answers to
questions not planned at design time.

My two (Euro-)cents,

                    Emmanuel Charpentier

--
Emmanuel Charpentier