Re: What's wrong with SQL Server?

Billy Verreynne <vslabs_at_onwe.co.za> wrote in message news:<bde2ki$c5m$1_at_ctb-nnrp2.saix.net>...
> JEDIDIAH wrote:
>
> >> > From this link:
> >> > In SQL Server, the DBA has no "real" control over sorting and cache
> >> > memory allocation.
> >> <loads snipped>
> >>
> >> So friggen what!?
> >
> > An RDBMS is not just a souped up version of MSOffice. It's an entire
> > development environment. As such, every database can be unique with
> > distinct requirements and bottlenecks. The more you can do to customize
> > the product to individual sites, the more likely you are to get the best
> > performance out of the RDBMS and minimize cost.
>
> Agree. But by the same token _many_ smaller companies can not afford a
> dedicated DBA. They are *not* using the db as a development environment,
> but are running shrinkwrap software. With the db being sold as an
> off-the-shelve product.

    Then it can only ever be highly limited in scope. Without some sort
of onsite expertise, it is absurd to expect any product to be fully utilized. If your notion of "sql server as imbedded rdbms" is realistic,
then my comment regarding postgres and mysql are certainly on point. If
you have no interest in ever maintaining the system, then Oracle is no less complex. You could even imbedd mysql with no ill effect.

>
> > Also, it's naive to expect the RDBMS to be effective at tuning itself.
>
> Agree again. But that does not change the fact that this is exactly what is
> required in the small to medium business market segment that SQL-Server
> sells so good in.

    That market typically doesn't hold Microsoft to it's claims.

>
> > Sometimes you need to be able
> > to "throw the manual away" in order to tune effectively.
>
> Bullshit. Exactly the opposite. If most people will just read the friggen
> manual on HOW the database works, then there will not be these huge
> performance problems.
>
> Performance problems are not solve by hacking the database with special
> config parameters and options. It is solved by doing it right from the
> start.
>
> And that means RTFM.

    Foolish consistency is the hobgoblin of little minds.

    If you are a professional, the book is only a starting point. You also
add your own experience and generalized knowledge of systems architecture.
You also don't really know what "hacking" is.

>
> > If your needs really so trivial, why not just use mysql or postgres?
>
> Get off it. Such little snides only show your ignorance.

    My comments are consistent with your own remarks on the matter actually.

>
> Business needs are not "trivial". But in many cases they can *not* be solved
> using as something as complex and expensive as Oracle (fact or not, this
> _is_ the market perception). Which is why SQL-Server is doing so good.

   You're being intentionally misleading here. There is a bit of a difference
between something being techinically trivial and it being trivial in terms
of business rules. The market that you claim needs Microsoft seem to be
merely using it as a security blanket as many customers do with Oracle.

>
> Technical superiority means jack shit when a company produces their year-end
> report.
>
> The right tool for the job. Sometimes its Oracle. Sometimes its not.

     Alternatively, it might not really be Microsoft either.

>
> What amazes me that some people to who Oracle is The Only Real Database,
> show the same ignorance and stupidity as what they claim Microsoft and
> SQL-Server supporters have.

     Ah yes, the "Lemming Mantra": Since you don't like Microsoft, you must therefore hate everything not FOO.  

[trash removed] Received on Mon Jun 30 2003 - 18:32:32 CDT