PostgreSQL Advocacy

If you want PostgreSQL to gain mindshare then you need to go after
it. To do that well you need to understand your target audience. I
think that I am part of your target audience.

I am a programmer. I have bee a professional coder since 1990. I've
worked with and still talk to a lot of people who have ended up
writing diverse applications such as income tax preparation, revenue
forecasting for airlines, application servers, resource management,
project management, support call tracking, terminal emulation and
file transfer, and also lots of internal IT type software. We do a
lot of things and we use a lot of tools. I did a survey of various
colleagues to try and figure out how we as a group decide upon
tools. Guess what, perception means a lot.

If PostgreSQL can become prevalent in the manner that we search for
tools then PostgreSQL will gain mindshare in actual usage.

The clear #1 in my survey was bookstores. We look in bookstores.
Whether those stores are online or brick-and-mortar we read books.
For some reason most of us prefer dead trees to online docs. I guess
it is a visceral thing. If I am looking to decide on whether to use
MySQL or PostgreSQL one of the first things I want to know is how
much support is out there. I judge that, fairly or not, by the
number of books available on the subject. I go looking through
MicroCenter, Barnes and Noble, 1/2 Price Books, and Borders to see
what are on the shelves. During one recent foray I found three
distinct titles and six physical books on PostgreSQL in two
different stores. I found twice as many TITLES on MySQL as there
were physical books on PostgreSQL. Don't even get me started on the
number of titles for Oracle, MS SQL Server and Access.

The clear #2, with caveats, on my survey was online news and
reviews. The caveats are there because most of my colleagues feel
that online rags are exceptionally biased and often lack discipline
in their reviews. But, filtering through the bias and lack of
discipline can often reveal accurate trends.

   ZDNet recently did (a pretty poor in my opinion) benchmarking of
   top databases. Among those benchmarked were MS SQL Server,
   Oracle, DB2, and MySQL. PostgreSQL wasn't in there. An obvious,
   but inaccurate, conclusion that can be drawn from this is that
   PostgreSQL isn't important enough to even consider. Doing a
   search on ZDNet for PostgreSQL you get papers on the history but
   not much else. Care to guess how many hits MySQL comes up with?

   O'Reilly Network has a section that includes MySQL as a major
   player. No mention of PostgreSQL.

   RedHat has a database engine built on PostgreSQL but also ships
   MySQL and has a relationship with Oracle. If you do a search on
   RedHat's site for "postgres" and another for "mysql" you will see
   that MySQL has four times as many hits. Obviously more important,
   right?

   What about SlashDot?


Nobody in my survey said that they would look at an advocacy
website unless it was run by some major third party such as
O'Reilly.

Have you noticed that I haven't mentioned performance or features
yet?

How well do other tools support the tool under question? If I'm
wanting to use PHP to drive a website then what database will work
best with that? (For entertainment go to http://www.php.org to see a
nifty search result).

The next consideration from my colleagues was platforms. A vast
majority of my colleagues prefer Linux. But despite our preferences
we often are not given a choice. Most of our companies are hosted on
Microsoft Win2K. MySQL has a native Win2K version that doesn't have
to deal with the overhead of Cygwin.

I have noticed reading through the threads that there is a native
Win2K porting effort underway and if PostgreSQL can get some good
press (especially good benchmarks!) out of that then that will be a
big win.

I truly appreciate that PostgreSQL is committed to the architecture
and the Unix underpinnings. But, the consequences of that decision
have to be acknowledged. Cygwin is a useful little beastie but it is
also a pain to configure and manage. And, there are performance
limitations. Whether or not those limitations will ever be hit by
any but the most high-volume applications is mostly irrelevant. We
don't like those limitations even being there because we want our
apps to be pushed to their limits.

Features and performance are the last things on the list. We
typically have a set of requirements that MUST be met that will fill
80% of our needs. Other requirements can be worked around until they
hurt the application. I was working around the lack of sub-selects
in MySQL for the longest time because they had a native Win2K
version. But I really couldn't get around the lack of views without
cratering performance. So I'm running on Cygwin until I can convince
my IT manager to give me a Linux box in the cold room.

Like I said up near the top, perception means a lot. Right now I'm
not seeing PostgreSQL as much of a blip on the radar.

The next question is - do you care? I'm serious, do you care what
other people think of PostgreSQL or are you happy to toil on in
relative annonimity? If you don't care, don't do anything and keep
on truckin'.

rjsjr