Some comments on the marketing texts - Mailing list pgsql-advocacy

From Peter Eisentraut
Subject Some comments on the marketing texts
Date
Msg-id Pine.LNX.4.44.0310100016380.32317-100000@peter.localdomain
Whole thread Raw
Responses Re: Some comments on the marketing texts
first version of the PostgreSQL flyer [was: Re: Some comments on the marketing texts]
List pgsql-advocacy
We intend to print some flyers for LinuxWorld Expo in Frankfurt at the end
of this month, using the texts on the advocacy pages.  But I have found
these texts to be disappointing.  They exaggerate boring features, they
are fuzzy on complex things rather than taking the time to explain them,
the arrangement and coherence of the points isn't all that great.

regarding the text at http://advocacy.postgresql.org/:

> * World-class security

While certainly PostgreSQL can be regarded secure, touting the security
"world-class" is exaggerated.  What is the standard for comparison?

> * Worldwide Independent Software Vendor (ISV) network

What is that?  Besides gratuitous use of marketing-friendly acronyms?

> * Extensive support options

What?  By whom?  Where?  Who evaluates that?

> * ANSI Standards Compliance

There are thousands of ANSI and ISO standards.  Be concrete about what you
mean.  And the correct term is "conformance".

Then there is the page http://advocacy.postgresql.org/advantages/, which
partially seems to duplicate or extend the above.  The relationship ought
to be made clearer.

> With PostgreSQL, no-one can sue you for breaking licensing agreements,
> as there is no associated licensing cost for the software.

I can still sue anyone for breaking licensing agreements.  There isn't a
lot to break in the licensing agreements, but that doesn't invalidate my
point.

> In addition to this our training programs are generally regarded as
> being far more cost effective, manageable, and practical in the real
> world than that of the leading proprietary database vendors.

What are "our training programs"?  Does the PostgreSQL group offer
trainings now?

I think there is too much mixing between what the PostgreSQL community
produces and what other external entities may provide around that.  Those
external entities are certainly an integral part of the whole deal, but
you need to be honest and clear to users about what comes from where.
Considering, for example, that you cannot possibly know about all training
programs that are being offered for PostgreSQL, making statements about
how they are generally regarded is bogus.

> If your staff have a need to customise or extend PostgreSQL in any way
> then they are able to do so with a minimum of effort, and with no
> attached costs.

I think it's better to talk to "you" instead of "your staff".  If the
person reading this doesn't have a staff, he might be turned off.  "You"
addresses all sizes and types of audiences.

> 34 platforms with the latest stable release

I keep counting and I can only find 23.  Instead of trying to impress with
numbers, I suggest you give some concrete information, such as by listing
the operating systems on which it runs.

> Designed for high volume environments
> We use a multiple row data storage strategy called MVCC to make
> PostgreSQL extremely responsive in high volume environments.  The
> leading proprietary database vendor uses this technology as well, for
> the same reasons.

"Designed" is misleading.  "Suitable" would be correct.  As we all know,
"PostgreSQL is bloatware by design, it was built to house Ph.D. theses"
(Hellerstein).  Also, does "high volume" mean a lot of data in the
database, or a lot of traffic?  In the first case, the association with
MVCC is wrong.

> * ANSI SQL compliant

What version of the standard, to what extent?  (also here "conforming")

> * Native interfaces for ODBC, JDBC, C, C++, PHP, Perl, TCL, ECPG, Python, and Ruby

This is a bit too mixed up.  There are "native interfaces" for the
programming languages C, C++, PHP, Perl, Tcl (note spelling), Python, and
Ruby.

It provides "drivers" for ODBC an JDBC.

It supports the embedded SQL in C binding (which happens to be implemented
by a program called ecpg).

> * Sub-selects

The proper term is "subquery".

> * An open API

Where?  To do what?

> * Hot stand-by (commercial solutions)

That really isn't appropriate to mention when a few paragraphs earlier
you're rejoicing about how open and free the system is.

> * Better than row-level locking

This needs to be made more concrete.  Why not throw in a mention of MVCC,
and then explain that MVCC gives you better than row-level locking.

> * Loadable extensions offering SHA1, MD5, XML, and other functionality

Are SHA and MD5 really that exciting to be representative of the extension
functionality?  Maybe OpenFTS and Postgis are more exciting examples of
plugins?

> * Tools for generating portable SQL to share with other SQL-compliant systems

Really?  Where can I get that?

> * Extensible data type system providing for custom, user-defined
> datatypes and rapid development of new datatypes

The wording seems redundant.  If the system provides for datatypes, it's
implied that one can develop them.


Finally, let me point you to this page

http://www.postgresql.org/users-lounge/features.html

which was written a while ago as a starting point for marketing material.
Notice that it gives the reader a sentence or two of explanation of each
important item, rather than just saying "trigger" or "Unicode".

--
Peter Eisentraut   peter_e@gmx.net


pgsql-advocacy by date:

Previous
From: Justin Clift
Date:
Subject: People in New York?
Next
From: Josh Berkus
Date:
Subject: Re: Some comments on the marketing texts