Re: [GENERAL] trash talk - Mailing list pgsql-advocacy

On Thu, 2004-08-05 at 09:01, Scott Ribe wrote:
> > Interesting, IBM is saying that the code is worth 85M dollars!
>
> Well... Actually they're saying: "we paid $85M for it, now we're giving it
> away, aren't we generous?"

Hey guys, after reading the article, I began a dialog with Lisa Vaas,
the author, and we had quite a civil discourse over it.

It would seem that she, like many others, WAS under the mistaken
apprehension that PostgreSQL inc pretty much was THE company behind
PostgreSQL, and that PostgreSQL was built on the same basic business
model as MySQL is.  I republish a fair bit of our dialog here:

-----------------------------------

FIRST EMAIL FROM ME:

From: Scott Marlowe [mailto:smarlowe@qwest.net]
Sent: Wednesday, August 04, 2004 12:10 AM
To: Vaas, Lisa
Subject: recent article on IBM Java database

Hi Lisa, I have a question.  In a recent article on eweek, there's a
paragraph I will now quote:

BLOCKQUOTE

The developer community for Cloudscape now consists of about 80 IBM
developers, Rivot said. IBM of course anticipates that population will
explode when the open-source community gets its hand on the code, but
just because a product goes open source doesn't mean it will succeed, as
can be witnessed by the failure of the PostgreSQL database to thrive
under this model.

/BLOCKQUOTE

That last bit about PostgreSQL.  I have some issues with it.

1:  Did Rivot say it or did you write it, hence think it?

2:  It is completely incorrect.  PostgreSQL once had a few major
companies behind it.  One of them, GreatBridge, spent their budget
liberally on PR saying how they were gonna go all the way etc...  And
soon disappeared.  Several other companies, PostgreSQL inc., SRA
associates in Japan, RedHat, Command Prompt, Fujitsu, and another half
dozen or so, still support active work on that database.  The new
version, due out in beta this week, adds Point in Time Recovery, Table
spaces, savepoints, very efficient buffer management, improved
background writer, and a Windows 32 port.  Many of these features were
paid for by commercial concerns who wanted the features added and paid
for them.

This is less than a year after the last major release, 7.3.0 by the way.

You can read this list for yourself here:
http://developer.postgresql.org/docs/postgres/release.html#RELEASE-7-5

Oh, and when you go there, the IP address for developer.postgresql.org,
like all addresses ending in .org, will be served by a PostgreSQL
database.  Just like the .info domain.  Ask IBM how many tlds they
provide database services for with any of their databases.

This isn't a flame, I'm not mad, but I am amazed at how woefully
underinformed the average tech press is about PostgreSQL.  It's like
it's a stealth project that never gets noticed.  Which goes with the
system administrators mantra "When you do your job right, people won't
be sure you did anything at all."  (OK, I stole that from Futurama)

Thanks for your attention.  It was good story about an interesting type
of database, the kind I'd use in a Palm pilot or something.  I'll have
to go play with it now.

END FIRST EMAIL

-----------------------------------------

Lisa's response:

Hi Scott,

Thanks for the feedback. When analysts talk about failures to thrive
under an open-source model, PostgreSQL is one of the main products they
think of. I threw PostgreSQL out as a suggested example, based on what
I've heard plenty of industry sources say, and Rivot agreed.

My impression is that people point to PostgreSQL's failure not in terms
of technological shortcomings, but rather in commercial terms or, more
specificly, in terms of mindshare/marketshare. I know scads of people
running PostgreSQL who speak very highly of it feature/function-wise.
But for whatever reasons, whether it's mismanagement on the part of the
company or complexities having to do with licensing, it doesn't come
close to where MySQL is at in terms of market share/mindshare.

I should have been more clear, and I apologize for that. I'm sure I'm
going to hear from a lot of PostgreSQL users on this one! At any rate,
good luck with Cloudscape/Derby. Please don't hesitate to let me know
what you do with it and what you think of it.

All my best,
Lisa

-----------------------------------------

My response:

> But for whatever reasons, whether it's mismanagement on the part of
the
> company or complexities having to do with licensing, it doesn't come
> close to where MySQL is at in terms of market share/mindshare.

That one line kind of points out what I think is your basic
misunderstanding of the PostgreSQL development model.

PostgreSQL is much more akin to apache than it is MySQL in terms of
development.

The PostgreSQL Global Development Group is not a company, but a group of
core developers, each sponsored by a different company, and autonomous
in their actions in support of PostgreSQL.

In MySQL, one company acts as the clearing house for input from both the
company's and volunteer developer's efforts.  What goes in is decided by
that company, and how MySQL is marketed is decided by that company.
Therefore, a single, strong commercial entity is making all decisions
about the future of the product.

PostgreSQL, like apache, is beholden to no company.

Several companies each hire postgresql developers as needed, either to
accomplish a single project, or to do ongoing development, then the
output of those programmers efforts are funneled into the PGDG core
group who decide what does or doesn't go in.  No company in charge, just
like the apache foundation.  So, no single entity can say "implement
XYZ" even if it's a bad idea.  The PGDG core group must approve it, and
they make that decision based soley on the merit of XYZ.

I think the foundation oriented open source projects will eventually be
responsible for all major infrastructure type projects, i.e. those that
are ubiquitous, unseen, and just work.  Projects like the linux kernel,
the apache server, postgresql, openldap for instance.

The commercial "ride along" of companies like MySQL is only necessary
should that product still be fighting for dominance against closed
source equivalents or poorly performing open source equivalents.

Once the non-commercial, developer controlled open source program(s) in
a given niche achieve(s) dominance, commercial entities are really no
longer needed for just that one product, but instead, commercial
entities will combine such programs into larger distributions (i.e.
RedHat, et. al.) and provide easier interfaces to them, and better
documentation as their value add.  RedHat already pretty much does this,
and for the cost of one highly regarded PostgreSQL developer, they get
years of computer expertise to help guide them in their database
strategy.

So, while no companies are making a killing off of licensing PostgreSQL,
they are making money off of including it with dozens and dozens of
other commodity programs.

I think it quite likely that one day IBM's DB2 team, Oracle, MySQL, and
several other database vendors will wake up to a wholesale migration
towards the truly free databases like PostgreSQL and FirebirdSQL,
because they will be "good enough" and free.

> I should have been more clear, and I apologize for that. I'm sure I'm
> going to hear from a lot of PostgreSQL users on this one! At any rate,
> good luck with Cloudscape/Derby. Please don't hesitate to let me know
> what you do with it and what you think of it.

The nice thing about PostgreSQL users in general is how nice they tend
to be.  Vocal and loyal yes, but even flame wars in the mailing lists
are quite polite.

Thanks for your attention, and have a great weekend.  Scott Marlowe

---------------------------------------------

Lisa's reply:

Scott,

You're right, I didn't understand that. I appreciate the explanation. It
is markedly different from other companies in its structure. I always
contact Geoff Davidson, CEO of PostgreSQL Inc., with the understanding
that there basically was a somewhat standard company-like structure
behind the technology. I can see that I assumed wrong, and I'll proceed
differently in my reporting on the company henceforth. At the very
least, I'll get people to explain more fully their dismissal of
PostgreSQL so I understand how it's perceived, and I'll try to correct
my own perception.

Thanks again, and thanks of course for the level of refreshingly civil
discourse that is characteristic of the PostgreSQL community.

All my best,
Lisa

-----------------------------

End of discourse.  I got Lisa's permission to post her replies before
posting them here.  I provide them as an example of how to properly
approach a journalist (i.e. no flaming, no screeching, assuming they
mean the best and have made a simple mistake based on misunderstanding,
etc etc...)