Comparing databases - Mailing list pgsql-advocacy

From Jussi Mikkola
Subject Comparing databases
Date
Msg-id 3FB164CE.4030109@bonware.com
Whole thread Raw
Responses Re: Comparing databases  (Robert Treat <xzilla@users.sourceforge.net>)
List pgsql-advocacy
Hi,

I have followed the discussion about MySQL database, and what we should
or should not do (comparisons, critique etc.).

I read (very quickly) the Oracle magazine, and here are some points I
noticed in there. After that there are shortly some points from IBM DB2
site (did not surf very thoroughly here either). At the end there are
some ideas about comparisons etc. Sorry, that this is a bit long.

First, Oracle tries to sell a database, but the database is just one
part of a larger system. They are telling a story, and there is a
hardware manufacturer, a consulting company, an application server
company, and perhaps the actual customer. Then they tell that how well
these all work together. And if possible, there are pictures of the
people. The pictures are important, since that makes you believe, that
those people are existing, and that thay really are saying what is there
on the paper. I also think that it is important to have the different
stakeholders. They are telling, that with these tools, we made this
success story, and that makes these tools good, and they work well
together. Okay, you say, that because of the standards compliancy, we
could make this well too. No, nothing so technical. It is not about JDBC
drivers. It is about having the systems work together. It really does
not usually matter very much, if you use Dell or Compaq, but having a
hardware maker saying, that these two work very well together just
sounds good.

The features. Self-managing and grid computing, are the two new main
features of the new 10g. Maybe there could be a third, but to sell 10
arguments is too much. People only remember a couple of features. And
they think that there are some points that are important. It can be,
that some day their database is huge, and they want to be sure, that
then this system will still serve them well, and they have not lost any
valuable work. Do you really need these features? Well, self-managing
sounds important. But if it did not exist in Oracle 7, then why did
people choose Oracle 7? It can be that it is good, but you can do
without it. But is has been chosen as a selling point, and that is why
they tell stories about it. Yes, one point that they sell is future. If
you now choose our product, it is good also in the future.

Foreign keys? Indexes? B-tree? No, too complicated. They can tell you
that they have a new improved indexing, and that the new system is much
faster than the old one, but thats it.

Then there is a story about the new release. Some interviews, pictures
of developers, and some new advances in the developement. What we have
done differently, and how that has made us even better.

I didn't see any database comparisons. I didn't even notice, that they
had mentioned DB2 or any other database. (Not even PostgreSQL ;-)



I also took a quick look at IBM's site about DB2.

IBM had one comparison with Oracle, and that was about TCO. The study
was very much focused on license prices.

DB2 also talked a lot about self managing stuff.



In conclusion, if we want to do product comparisons, we could quite
safely compare different PostgreSQL versions. That would be to tackle
the issues of the problems with earlier versions. Also it would show the
progress. If quality has been a problem, then stories about the process
and testing. Studies that show the reliability.

IBM and Oracle both have large marketing organizations, and I am sure
they have very carefully thought, what their message is. They have made
studies about the market, it's needs and put a lot of effort and money
in their material. Perhaps we could learn something from them? Not only
technology, but also sales and marketing. Okay, they have a price tag on
the database, but it does not affect everything.

Also, I think that _if_ we compare databases, we should always first
compare with the market leader. If people compare PostgreSQL and ms
access, or they compare PostgreSQL and Oracle, it has a difference.
PostgreSQL could win ms access 10 - 0, but still, that would not make
PostgreSQL very highly appreciated database. 5-5 against Oracle, and the
situation is a lot different. So, if someone would ask for a comparison
between PostgreSQL and MySQL, I think the comparison chart and text
should first compare PostgreSQL and Oracle. And after that PostgreSQL
and MySQL. But having clearly focus on the Oracle comparison. Because we
actually want people to compare PostgreSQL with Oracle, not with MySQL
or ms access.

Rgs,

Jussi



pgsql-advocacy by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: FW: [webmaster] Comparison to MySQL
Next
From: Robert Treat
Date:
Subject: Re: Comparing databases