It's always amusing to read a description of some DBMSs as "post-relational",
or to read that E.F. Codd's 1970 paper about the relational model was the
genesis of database technology.

"A Relational Model of Data for Large Shared Data Banks", E. F. Codd
(Communications of the ACM, Vol. 13, No. 6, June 1970)

When cleaning out some old files, I found a folder with vintage-1968 documents
related to the CODASYL Data Base Task Group. One of the documents is a "Survey
of Data Base Systems" dated 3 September 1968. It lists 51 database systems, and
identifies another 22 systems not included in the survey.

The systems in that survey are predecessors of the CODASYL database standard,
Codd's relational model, and SQL. Three of those systems have evolved in the 35
years since the survey.

IBM IMS is still in use. Dick Pick evolved GIM-I into the Pick database family.
The Mass General/BBN system evolved and is still available today as MUMPS, M and
Caché.


--
Ken North
www.WebServicesSummit.com
www.SQLSummit.com
www.GridSummit.com (coming soon)

"Ken North" <knorth2 (AT) deletethis (DOT) yahoo.com> wrote

Sweet! I've been tracking down pre-SQL database information and this is
most helpful. I haven't tracked ADABAS to its origins, but might that be
somewhere in the pre-relational? I, too, laugh at the "post-relational"
claims, often used with MUMPS or PICK products (even by companies like IBM
who ought to know better, eh?).
--dawn

The Pick system is currently marketed by:

http://www.rainingdata.com/

which is the current embodyment of the original
company.

Dawn M. Wolthuis wrote:

--
Kevin C. O'Kane
University of Northern Iowa
Cedar Falls, IA 50614-0507
(319) 273 7322 (Office + Voice Mail)
(508) 778 9485 (Massachusetts)
http://www.cs.uni.edu/~okane
okane (AT) cs (DOT) uni.edu

"Kevin O'Kane" <okane (AT) cs (DOT) uni.edu> wrote

You might want to take a look at
http://www.tincat-group.com/images/M...yTreeColor.pdf to see a diagram I
did when researching both history and current vendors of products based on
the Nelson-Pick "data model." IBM has the largest installed base.
Cheers! --dawn

Dawn M. Wolthuis wrote:

[Ken's post snipped but just a comment ...

Darn Ken, you're older than I am and I'm three days older
than dirt!

I had some info on ADABAS 20 or so years ago. It was
a product by a German company, but I don't remember
which one. I'll look up in the attic to see if
I can find it. I know it ended with -AG, but that's
the German equivalent of "Inc."

As I recall, ADABAS was a very interesting product.
If you were using a hierarchical model (like IMS), you could
use ADABAS. If you were using a CODASYL (network) model
you could use ADABAS. If you had a relational
model you could use ADABAS! (This was circa 1983-5,
I don't recall how old it was at the time.)

You could also define repeating fields within your
tables, as I recall, but memory is fuzzy.

The only reason we didn't use it back then was
that it only ran on mainframes and we were limited
to VAXen at the time.

--
"It is impossible to make anything foolproof
because fools are so ingenious"
- A. Bloch

Em Sun, 25 Apr 2004 23:34:23 +0000, Nick Landsberg escreveu:

Today it is still alive, and it is the origin of SAPdb, now
renamed MaxDB to be sold by MySQL AG as the big brother of its
namesake, substandard product...


--
Leandro Guimarães Faria Corsetti Dutra +55 (11) 5685 2219
Av Sgto Geraldo Santana, 1100 6/71 +55 (11) 5686 9607
04.674-000 São Paulo, SP BRASIL
http://br.geocities.com./lgcdutra/

"Dawn M. Wolthuis" <dwolt (AT) tincat-group (DOT) com> wrote


Try this page for a history of databases, including the 50s and 60s:
http://ourworld.compuserve.com/homep...th/db_hall.htm


adabas was one of the DBMSs that complied with the CODASYL (network model)
database standard. You can probably find the history at the Software AG web
site.

The 1968 survey was done in the formative years before the CODASYL DBTG
published its standard.

When there's time, I'll dig out the proceeedings of a comparative DBMS symposium
we did in 1974. adabas was one of the systems presented at that symposium.

There's a discrepancy between

1. Your diagram (http://www.tincat-group.com/images/M...yTreeColor.pdf )
2. this page (http://ourworld.compuserve.com/homep...th/db_hall.htm)
3. various Pick history documents over the years (e.g.,
www.picksys.com/company/corp.pdf)

and an article in my folder of 1968 documents.

References 1 and 2 show GIRLS to be a predecessor to TRW's GIM-I (1965). That's
in agreement with an article published in a Datamation article (1967 or 1968,
the data is illegible).

Some Pick documents refer to GIRLS as originally conceived as a TRW project. For
example, this Pick corporate history is incorrect (according to that old
article):

"In 1965, as part of a research and development project for the U.S. Army,
engineers at TRW Corp developed a software program
called Generalized Information Retrieval Lanaguage and System." Or GIRLS for
short."

The 1965 effort at TRW was part of the Cheyenne helicopter program, which was
eventually cancelled by the US government. The DBMS produced by that effort was
called GIM, ane eventually GIM-I when TRW started developing GIM-II for PDP-11
minicomputers. GIM-I was still running at TRW after Dick Pick was working on
Microdata Reality in the early '70s.

In "General Purpose Software" (Datamation), Donald H. Sundeen is pretty clear
about the origin of GIRLS:

"A joint effort by Advanced Information Systems and Douglas Aircraft Company in
1961 resulted in the development of GIRLS (Generalized Information Retrieval and
Listings System) for an IBM 7090 at the latter firm's Missile and Space Systems
Division."

So why the confusion? It was probably an overlap in personnel -- one or more
GIRLS developers who worked on GIM a few years later (possibly Dwight Buettell,
Don Nelson or Dick Pick).

"Ken North" <knorth2 (AT) deletethis (DOT) yahoo.com> wrote

Very interesting. I should have tried harder to find someone who was at TRW
in the 60's. I talked with several folks who had Dick Pick's stories from
TRW, having heard them after he left in '69. I didn't know there was
controversy on that issue. I do have the GIM-I & II information, but the
only original document I have was from '65. There are detailed flow charts
that Don Nelson prepared in the early 60's as the spec for GIRLS. I thought
he did that while at TRW, with Pick joining him later. Is Dwight Duettell
still living?
--dawn
P.S. Sorry this is offtopic, but I'm planning to do an update on the
diagram next year (40th anniversary) and have several minor corrections to
make, but sure would like to have the origins as accurate as can be.

Additional comments about GIM/Pick database history that you posted in another
newsgroup:

<< It was written at TRW in order to make it so that the military in Viet Nam
could query their data without technical folks in the field.

GIRLS (1961) and GIM (1965) were developed before minicomputers and
microcomputers were available. GIM-I ran on IBM mainframes. The IBM mainframes
in use in Vietnam and Thailand were not running GIM during the Vietnam War.

The original funding to develop the software was the US government's Cheyenne
helicopter program. Keeping track of a bill-of-materials (parts database) for a
major aircraft was understood not to be a problem for paper and filing cabinets,
even in 1965. Ironically, the Cheyenne never went into production.

<< Prior to the end of the cold war, it was used by the CIA

Some of the original GIM team worked on porting from the 32-bit IBM 360 (GIM-I)
to the 16-bit PDP-11 (GIM-II). The GIM-II port was done in the 1972-73 timeframe
at TRW's Washington Operation. Dick Pick had left TRW before that time. While
TRW was developing GIM-II for the PDP-11, Pick was doing a port for Microdata
minicomputers. That became Microdata Reality.

By 1974:

1. GIM-I was running at TRW Systems Group in Redondo Beach, California
2. Microdata Reality was just down the road in Irvine, California
3. GIM-II was running at the CIA (Virginia)
4. TRW had GIM licensees around the world, including CADAFE in Venezuela and
Matra in France.

The next port was taking GIM-I from 32-bit IBM mainframes to 36-bit Univac
mainframes. That was done in Houston, Texas for NASA's Manned Space Flight
Center. The Univac version of GIM was eventually licensed to a commercial
timesharing network (Infonet) as I-GIM.

In the mid-70s, TRW took delivery of an IBM 370/158 at TRW Systems Group in
Space Park (Redondo Beach). That was a virtual memory computer that supported
demand paging and provided hardware for direct address translation.
http://www-1.ibm.com/ibm/history/exh...me_PP3158.html

Before IBM offerred virtual memory on the 370, the GIM system had done its own
paging on IBM 360s. The DBMS code and data structures were too large to reside
in memory. It was necessary to analyze data structures and execution paths to
determine what modules belonged in what pages. The purpose was to swap pages
efficiently to/from disk and avoid thrashing.

Later in the '70s, the GIM system at Space Park was used for other important
activities, such as development of NASA's Space Transportation System (the Space
Shuttle).