PostgreSQL History Document

This is what I have regarding Postgres (not PostgreSQL only) history.
All of the quotes and people references were OK'd by the people
who made them. This supplements Bruce's history in that this
is the history as I experienced it, having been at UCB, worked at Ingres
and then at Miró, et al. Bruce's history from the perspective
of the postgres project and then PostgreSQL is better stated
from those who experienced that although I tried to cover the
transition to the pgdg.

This history was written circa late 2002. Obviously it can be added to.
I also have a picture of the original source tree branching to
the various entities if anyone can stand my corny graphics.

Robert Bernier--If people want to use it for part of the
historical documents it is freely given to the PostgreSQL project.
I wouldn't mind being cited as the original author, however,
since I got the quotes and self reference. My writing can always use
a good editing though.

--elein

INGRES
Ingres was the predecessor of Postgres. Ingres was developed at
UC Berkeley (1977-1985) under Dr. Michael Stonebraker's
tutelage. Ingres was productized by Relational Technology, Inc
which renamed itself Ingres in 1989. Ingres was known for its
good technology, lousy marketing and for its mass exodus of
engineers the day it was purchased by Computer Associates.

POSTGRES
Postgres followed Ingres (1986-1994) as the database project development at
UC Berkeley. It was built by many grad students led
by Michael Stonebraker. Postgres is an original
implementation of an Object-Relational Database system. It
used the query language _Quel_ as did early Ingres.

ILLUSTRA
In 1992, Michael Stonebraker, Gary Morganthaler, Michael Ubell
and Paula Hawthorn joined with a small team to create the
company Miró. Miró took a branch from UC Berkeley's Postgres
source tree and immediately adapted it to use SQL, implemented
(fixed) page level locking and a number of other key features.
Over the next few years, Miró became Montage and then became,
Illustra.

The Illustra product was based University Postgres but had a
much improved infrastructure. Illustra engineers and
consultants were able to clean up a lot of graduate students'
best intentions. Production tools and user interfaces were
added. The term "DataBlades" was coined by Michael Stonebraker
to mean a group of data types and their supporting functions.
For example, the group of geometric data types and their
functions was a DataBlade.

POSTGRES
In 1994 the University Postgres project was officially
completed. However, Postgres was to be used as a foundation
for the Tioga and the Mariposa projects. Andrew Yu, the Chief
Programmer for Mariposa and Jolly Chen, a PhD. candidate at
Berkeley, did some significant code clean up as well as adapted
Postgres to use SQL in order to use it with Mariposa and Tioga.
This version was released in 1995 as postgres95.

Andrew Yu, reflecting on the history of PostgreSQL, said,

"We didn't have postgres95 in mind when we [started] out
with the clean up project. The release [was] more of an
after-thought. We didn't set out to release it publicly.
The popularity was a nice surprise. And the fact that it's
going strong these many years after our initial work is
beyond our wildest dreams."

While working on their primary University projects, Yu and
Chen continued maintaining postgres95 and its mailing lists
until each left the University. After graduating in 1995
Andrew Yu went to work with Illustra and in 1996 Jolly Chen
left the University to work with TCSI in Berkeley and then
joined the start up WebLogic (acquired by BEA) which still
is the foremost Java application server on the market.

POSTGRESQL
By 1996, neither Andrew nor Jolly Chen had much time to devote
to Postgres95, but some members of the mailing list stepped up
to the plate. Jolly remembers,

"We appreciated [their]
enthusiasm over the project. Since the code was already
licensed in an open BSD-style, there wasn't much we needed to
do as far as the turnover was concerned. They just set up
some new mailing lists, ftp sites, and that was it. Neither
Andrew nor I felt like we "owned" the project. After all,
postgres was the result of many graduate students work over
the years, not ours alone."

This group became PostgreSQL Global Development Team. The
group was composed of Marc Fournier in Ontario, Canada, who
hosted the mailing lists and servers, Thomas Lockhart in
Pasadena, California, Vadim Mikheev in Krasnoyarsk,
Russia and Bruce Momjian in Philadelphia, Pennsylvania.
The name of postgres95 was changed to PostgreSQL.

Marc Fournier was able to provide server space and worked
on maintaining the source tree with others providing the key
patches and changes to maintain and improve the product.
The development and release processes were based on FreeBSD's
development model which Marc had been working with previously.
In 1999, the Hub.Org networking organization was started in
order to help provide funding for the infrastructure required
by the PostgreSQL maintenance.

INFORMIX
Also in 1996 in the commercial arena, Informix acquired
Illustra for its engineers and technology. The merge put
Illustra, the database, on the back burner while the team in
Oakland, California, along with engineers in Menlo Park,
California and Portland, Oregon implemented the Object
Relational technology into the existing multi-threaded
Informix 7 to produce Informix 9. I was a part of that team.

We worked against a killer one-year deadline for version 9.0.
There were nine teams implementing user defined, row and set
data, user defined types, user defined functions, the DataBlade
API (my area), large object, virtual tables and
virtual table indexes, DataBlades, and SQL extensions.
Informix 9 went through several name changes but was commonly
known as Informix IUS.

With the halt of active Illustra database sales, some of us
on the Illustra maintenance team approached Stonebraker with
the idea to make the Illustra database open source. Stonebraker
thought it was a good idea and so we initiated some informal
discussions with the PostgreSQL Global Development team to
consider turning over the code line to merge back their
development branch. However, corporate Informix would never
quite allow us to pursue the issue.

In 2001, Informix split. Half of the company seceded
and formed Ascential. The database portion of the
business was sold to IBM.

PgSQL,INC
In 2000, Postgres, Inc. was formed with Marc Fournier as
President/Director as a corporate front to PostgreSQL and to
offer the corporate world commercial support for the product.
The relationship between PostgreSQL and PgSQL, Inc has always
been symbiotic. Core developers were and are part of PgSQL, Inc.

GREATBRIDGE
In July 2000 the company GreatBridge was formed by
Frank Batten, Jr. (formerly with Red Hat) with the mission of
providing marketing and support for the PostgreSQL product.
It employed up to forty people, including Tom Lane,
Bruce Momjian and Jan Weick from the Global Development team.

Although Marc Fournier was a competitor of sorts, he
remarked on the benefits of GreatBridge on PostgreSQL.

"[T]hey allowed Tom Lane to focus on developing PostgreSQL
features as a full time job, instead of as a hobby
... and it allowed Bruce to travel and evangelize PostgreSQL
through providing talks to various organizations [and]
increased visibility."

Unfortunately, by August 2001, the money for GreatBridge was
exhausted and the company closed down.

POSTGRESQL
The early progress and work of the the PostgreSQL Global
Development Team is nicely documented by Bruce Momjian on
the PostgreSQL web pages.

The PostgreSQL project has been maturing and settling
into a regular development cycle. 7.3 was released
in November 2002 and 7.4 should be available late 2003.

There are twenty three mailing lists hosted on postgresql.org.
The one list one must subscribe to is pgsql-announce where
releases are announced. The other most active lists are:
* pgsql-admin
* pgsql-bugs
* pgsql-general
* pgsql-sql
* pgsql-hackers (for developers of postgresql)

The mailing lists are also available both as newsgroups
and as archives on the postgresql.org web site. It is
required that you join the mailing list in order to post
to the list, however, you can join and set an option so
that you do not receive mail. This option enables you
to read the lists from newsgroups and still post questions
and answers.

These mailing lists are the heartbeat of the project and
are well tended to by each contributor. It is clear
from reading these lists that the common goal of using,
developing and supporting a world class database system is
the foremost ideal held by the contributors.

An advocacy website was built in November 2002 at
advocacy.postgresql.org and enables people to contact
PostgreSQL advocates in eight different languages,
including English.

Also in November 2002, the weekly column PostgreSQL
General Bits was begun (REF: www.varlena.com/GeneralBits)
by this author and in early 2003 Robert Treat began summarizing
the weekly status of the project in messages to
pgsql-announce.

(REF: http://www.postgresql.org/docs/devhistory.html)
(REF: http://www.ndim.edrc.cmu.edu/postgres95/www/pglite1.html)
(REF: http://s2k-ftp.cs.berkeley.edu:8000/Sequoia2000)
(REF: http://s2k-ftp.cs.berkeley.edu:8000/postgres/papers/)
(REF:http://www.oreillynet.com/pub/a/network/2000/06/16/magazine/postgresql_history.html)

On Friday 03 December 2004 21:42, elein wrote:
> ILLUSTRA
> In 1992, Michael Stonebraker, Gary Morganthaler, Michael Ubell
> and Paula Hawthorn joined with a small team to create the
> company Miró. Miró took a branch from UC Berkeley's Postgres
> source tree and immediately adapted it to use SQL, implemented
> (fixed) page level locking and a number of other key features.
> Over the next few years, Miró became Montage and then became,
> Illustra.

Illustra was the other database I looked at when backending my
Navi^H^H^H^HGNN^H^H^HAOLserver installation (Illustra and the full text
DataBlade was extremely well supported by that beast). Just wasn't available
for Linux.

> Relational technology into the existing multi-threaded
> Informix 7 to produce Informix 9. I was a part of that team.

> With the halt of active Illustra database sales, some of us
> on the Illustra maintenance team approached Stonebraker with
> the idea to make the Illustra database open source. Stonebraker

> In 2001, Informix split. Half of the company seceded
> and formed Ascential. The database portion of the
> business was sold to IBM.

Any chance anyone inside IBM might be able to make that happen (the
open-sourcing of Illustra?).

The whole DataBlade mechanism was cool; I forget right off which DataBlade the
full-text indexer in AOLserver used, but it was working in 1997 very well.
--
Lamar Owen
Director of Information Technology
Pisgah Astronomical Research Institute
1 PARI Drive
Rosman, NC 28772
(828)862-5554
www.pari.edu

elein wrote:
> PgSQL,INC
> In 2000, Postgres, Inc. was formed with Marc Fournier as
> President/Director as a corporate front to PostgreSQL and to
> offer the corporate world commercial support for the product.
> The relationship between PostgreSQL and PgSQL, Inc has always
> been symbiotic. Core developers were and are part of PgSQL, Inc.

I don't think PgSQL, Inc can be called a "corporate front to
PostgreSQL". It is a PostgreSQL support company, and provides free
hosting services to the project. I also wouldn't use the word
"symbiotic". We could say Great Bridge was also "symbiotic" but it
conotates an equality between the two entities that isn't accurate.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

On Sat, 4 Dec 2004, Bruce Momjian wrote:

> elein wrote:
>> PgSQL,INC
>> In 2000, Postgres, Inc. was formed with Marc Fournier as
>> President/Director as a corporate front to PostgreSQL and to
>> offer the corporate world commercial support for the product.
>> The relationship between PostgreSQL and PgSQL, Inc has always
>> been symbiotic. Core developers were and are part of PgSQL, Inc.
>
> I don't think PgSQL, Inc can be called a "corporate front to
> PostgreSQL". It is a PostgreSQL support company, and provides free
> hosting services to the project. I also wouldn't use the word
> "symbiotic". We could say Great Bridge was also "symbiotic" but it
> conotates an equality between the two entities that isn't accurate.

When it was formed in 2000, that was its intent, since there was none ...
'corporate front' != 'controlling organization' ...

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy(at)hub(dot)org Yahoo!: yscrappy ICQ: 7615664

Marc G. Fournier wrote:

> On Sat, 4 Dec 2004, Bruce Momjian wrote:
>
>> elein wrote:
>>
>>> PgSQL,INC
>>> In 2000, Postgres, Inc. was formed with Marc Fournier as
>>> President/Director as a corporate front to PostgreSQL and to
>>> offer the corporate world commercial support for the product.
>>> The relationship between PostgreSQL and PgSQL, Inc has always
>>> been symbiotic. Core developers were and are part of PgSQL, Inc.
>>
>>
>> I don't think PgSQL, Inc can be called a "corporate front to
>> PostgreSQL". It is a PostgreSQL support company, and provides free
>> hosting services to the project. I also wouldn't use the word
>> "symbiotic". We could say Great Bridge was also "symbiotic" but it
>> conotates an equality between the two entities that isn't accurate.
>
>
> When it was formed in 2000, that was its intent, since there was none
> ... 'corporate front' != 'controlling organization' ...

Perhaps the issue is the wording:

The relationship between PostgreSQL and PgSQL, Inc has always
been symbiotic.

Which I would think is no longer accurate.

Sincerely,

Joshua D. Drake

>
> ----
> Marc G. Fournier Hub.Org Networking Services
> (http://www.hub.org)
> Email: scrappy(at)hub(dot)org Yahoo!: yscrappy ICQ:
> 7615664
>
> ---------------------------(end of broadcast)---------------------------
> TIP 8: explain analyze is your friend

--
Command Prompt, Inc., home of Mammoth PostgreSQL - S/ODBC and S/JDBC
Postgresql support, programming shared hosting and dedicated hosting.
+1-503-667-4564 - jd(at)commandprompt(dot)com - http://www.commandprompt.com
PostgreSQL Replicator -- production quality replication for PostgreSQL

On Fri, 3 Dec 2004, Joshua D. Drake wrote:

> Marc G. Fournier wrote:
>
>> On Sat, 4 Dec 2004, Bruce Momjian wrote:
>>
>>> elein wrote:
>>>
>>>> PgSQL,INC
>>>> In 2000, Postgres, Inc. was formed with Marc Fournier as
>>>> President/Director as a corporate front to PostgreSQL and to
>>>> offer the corporate world commercial support for the product.
>>>> The relationship between PostgreSQL and PgSQL, Inc has always
>>>> been symbiotic. Core developers were and are part of PgSQL, Inc.
>>>
>>>
>>> I don't think PgSQL, Inc can be called a "corporate front to
>>> PostgreSQL". It is a PostgreSQL support company, and provides free
>>> hosting services to the project. I also wouldn't use the word
>>> "symbiotic". We could say Great Bridge was also "symbiotic" but it
>>> conotates an equality between the two entities that isn't accurate.
>>
>>
>> When it was formed in 2000, that was its intent, since there was none ...
>> 'corporate front' != 'controlling organization' ...
>
> Perhaps the issue is the wording:
>
> The relationship between PostgreSQL and PgSQL, Inc has always been symbiotic.
>
> Which I would think is no longer accurate.

agreed ... the fun with 'history' is avoiding "revisionist history" ...
"formed as a corporate front" is accurate ... but with the growing # of
support companies like CP, SRA, etc, that role has gone from one to many
...

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy(at)hub(dot)org Yahoo!: yscrappy ICQ: 7615664

Hi there,

looks like very complete history ! But it's not what I expected.
I think that in community site we should have history of project
developing from community view, in other words - history of PostgreSQL and
its community. This doesn't mean that history by Elein, Bruce shouldn't
be at site, it's just another thing. We need history page which consolidates
us via common history, sort of "unofficial history". For example,
I joined PG community at 95 and my first contribution
to PG was introducing locale support. It took a several weeks to
understand source tree and to get PG95 1.0 working under Sunos,
patch was incredibly small, but it was very important for many
no-ascii people. In spite of locale support was completely rewrited by
Tatsuo, I think my contribution was important and could be mentioned in
community history page, while it's not enough noticeable for "official history"
page. Also, who remember Jason Wright (vnet.net), maintainer of pg95 mailing
list ? I think, Marc and Bruce should . There are many contributors, which
should be mentioned. Some of them retired, some - just use PG in their
projects and still in mailing lists. We could ask them to write about
their contribution to PG and we'll get a vivid history of PG project.

We need some numbers to see how PG community is growing, Marc, probably,
know number of subscribers per year. I could check my database and see
distribution of posts and authors. I have about 31000 unique authors in
mailware database.

Understanding community history helps us in understanding who we are.

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83

I'm going to look at this stuff next week. I will write up a draft and post it
for review, then after we've thrashed it around a couple of zillion times it
can be incorportated into the new site in the appropriate section.

comments?

Oleg Bartunov wrote:
> "unofficial history". For example, I joined PG community at 95 and my
> first contribution
> to PG was introducing locale support. It took a several weeks to
> understand source tree and to get PG95 1.0 working under Sunos,
> patch was incredibly small, but it was very important for many
> no-ascii people. In spite of locale support was completely rewrited
> by Tatsuo, I think my contribution was important and could be
> mentioned in community history page,

All contributions are attributed in the release history.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

On Fri, Dec 03, 2004 at 11:49:32PM -0500, Lamar Owen wrote:
> On Friday 03 December 2004 21:42, elein wrote:
> > ILLUSTRA
> > In 1992, Michael Stonebraker, Gary Morganthaler, Michael Ubell
> > and Paula Hawthorn joined with a small team to create the
> > company Miró. Miró took a branch from UC Berkeley's Postgres
> > source tree and immediately adapted it to use SQL, implemented
> > (fixed) page level locking and a number of other key features.
> > Over the next few years, Miró became Montage and then became,
> > Illustra.
>
> Illustra was the other database I looked at when backending my
> Navi^H^H^H^HGNN^H^H^HAOLserver installation (Illustra and the full text
> DataBlade was extremely well supported by that beast). Just wasn't available
> for Linux.
>

>
> > Relational technology into the existing multi-threaded
> > Informix 7 to produce Informix 9. I was a part of that team.
>
> > With the halt of active Illustra database sales, some of us
> > on the Illustra maintenance team approached Stonebraker with
> > the idea to make the Illustra database open source. Stonebraker
>
> > In 2001, Informix split. Half of the company seceded
> > and formed Ascential. The database portion of the
> > business was sold to IBM.
>
> Any chance anyone inside IBM might be able to make that happen (the
> open-sourcing of Illustra?).
>

Not a chance. I know exactly what happened to the source code and
it is No Longer Available unless someone made off with a copy.
(Unfortunately, I did not. Fortunately, that would have been illegal
so I guess it is best I did not.) I really tried to do my best fueled
by David Gould's enthusiasmand got Stonebraker behind us but the informix
corporate stonewalled us and anyone (except ex-illustra people) at IBM
would not know what you are talkiing about, let alone where to find the
dead and broken server that was the sole repository of code.
Big fish eat little fish and then get eaten by even bigger fish.

> The whole DataBlade mechanism was cool; I forget right off which DataBlade the
> full-text indexer in AOLserver used, but it was working in 1997 very well.

Note that postgres has the datablade "mechanism". It is only a matter of
marketing and packaging. Anything Illustra did could be or is doable by
PostgreSQL.

> --
> Lamar Owen
> Director of Information Technology
> Pisgah Astronomical Research Institute
> 1 PARI Drive
> Rosman, NC 28772
> (828)862-5554
> www.pari.edu
>
> ---------------------------(end of broadcast)---------------------------
> TIP 8: explain analyze is your friend
>

Elein,

On Sat, 4 Dec 2004, elein wrote:

>
>
>> The whole DataBlade mechanism was cool; I forget right off which DataBlade the
>> full-text indexer in AOLserver used, but it was working in 1997 very well.
>
>
> Note that postgres has the datablade "mechanism". It is only a matter of
> marketing and packaging. Anything Illustra did could be or is doable by
> PostgreSQL.

do you mean GiST ?

>
>> --
>> Lamar Owen
>> Director of Information Technology
>> Pisgah Astronomical Research Institute
>> 1 PARI Drive
>> Rosman, NC 28772
>> (828)862-5554
>> www.pari.edu
>>
>> ---------------------------(end of broadcast)---------------------------
>> TIP 8: explain analyze is your friend
>>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: the planner will ignore your desire to choose an index scan if your
> joining column's datatypes do not match
>

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83

On Sun, Dec 05, 2004 at 12:34:56PM +0300, Oleg Bartunov wrote:
> Elein,
>
> On Sat, 4 Dec 2004, elein wrote:
>
> >
> >
> >>The whole DataBlade mechanism was cool; I forget right off which DataBlade
> >>the
> >>full-text indexer in AOLserver used, but it was working in 1997 very well.
> >
> >
> >Note that postgres has the datablade "mechanism". It is only a matter of
> >marketing and packaging. Anything Illustra did could be or is doable by
> >PostgreSQL.
>
> do you mean GiST ?
>

Not especially. Most things in contrib qualify as a "datablade".
As I said it was just a packaging and marketing thing.

elein

>
> >
> >>--
> >>Lamar Owen
> >>Director of Information Technology
> >>Pisgah Astronomical Research Institute
> >>1 PARI Drive
> >>Rosman, NC 28772
> >>(828)862-5554
> >>www.pari.edu
> >>
> >>---------------------------(end of broadcast)---------------------------
> >>TIP 8: explain analyze is your friend
> >>
> >
> >---------------------------(end of broadcast)---------------------------
> >TIP 9: the planner will ignore your desire to choose an index scan if your
> > joining column's datatypes do not match
> >
>
> Regards,
> Oleg
> _____________________________________________________________
> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
> Sternberg Astronomical Institute, Moscow University (Russia)
> Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
> phone: +007(095)939-16-83, +007(095)939-23-83
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/docs/faqs/FAQ.html
>