Re: Oracle History??

afilonov_at_yahoo.com (Alex Filonov) wrote in message news:<336da121.0211260824.7528b3e2_at_posting.google.com>...
> joel-garry_at_home.com (Joel Garry) wrote in message news:<91884734.0211251748.43bbe98_at_posting.google.com>...
>
> <snip irrelevant stuff>
>
> > > > Mainframes and Minicomputers," and "the first SQL" whatever, as if
> > > > that is some kind of great thing. In the early to mid '80s there were
> > > > several superior languages to SQL, and Oracle was by no means the best
> > >
> > > Like what? I'm apparently becoming old and can't remember any DB language
> > > superior to SQL. BTW, IBM was actually first to create relational DB
> >
> > User-11 on PDP's, and it's various children and grandchildren.
> >
> > InTouch.
> >
> > Datatrieve, for that matter.
> >
>
> Datatrieve was based on network model, was it not? Just asking, I don't
> remember much from pre-SQL times.

Naw, sorta relational on hierarchical. The original Rdb was also layered on top of a hierarchical system. But that didn't last long. The Rdb that Larry picked up for a song was far ahead of Oracle. 8i/9i are finally implementing those bits of intellectual property.

>
> > Those are just ones I happen to know. I'm sure there are more that I
> > don't.
> >
> > Even the SQL creators recognized that it was only half-a-standard:
> >
> > http://ftp.digital.com/pub/DEC/SRC/technical-notes/SRC-1997-018-html/sqlr95.html
> >
> > > based on SQL (project R), but initially they decided against developing
> > > in further into marketable product. AFAIK Larry Ellison was working for
> > > IBM then... May be he's got some ideas.
> >
> > Yeah, I'm sure he went "Wow! Let's standardize on SQL as a data
> > interchange language!" Not.
> >
>
> I didn't say so. But he could've got the idea of marketable SQL-based
> database system.
>
> > >
> > > > system for any particular purpose. Like IBM, the greatness, if you
> > > > want to call it that, was in the marketing and financials. So what if
> > > > Star could join db's in New York and SF - in the mid-80's, I saw a
> > > > language that could join between different DBMS's!
> > >
> > > SQL can do it. There is no restrictions on phisical location of tables
> > > in the language definition. True, it wasn't realized until Oracle 6
> > > (I might be wrong, correct me if so), but it's not language restriction,
> > > it was an implementation restriction.
> >
> > Sorry, I don't know the syntax for SQL*Plus to open an Ingres
> > database. However, I do know the syntax for a language that could
> > open an Oracle database and an Ingres database and join the tables
> > together - at least if I root around in the basement and find the
> > manuals from 1986.
>
> SQL*Plus is not a language, it's a tool. SQL is a language and there is
> nothing in language definition which prevents it from working in the
> distributed system and/or with different database systems. Again, you
> are talking about implementation restrictions.

Well, if we don't talk about implementation restrictions, how can we talk about the history of the technology? If SQL were a complete language, we wouldn't need PL/SQL, TSQL, iSQL, etcSQL... What people tend to forget is, Codd proved R is potentially as fast as any other system - but the mapping of reality to the theory is another coddle of fish. (ouch, sorry)

SQL*Plus is an extension of the SQL language.

>
> The point is, SQL is just another notation for relational calculus, at
> least it was initially. I don't know any other math model for databases as
> good as relational calculus (relational algebra is a full equivalent,
> of course, but it's not as good to write complex queries).

True, as far as it goes.

>
> Oracle was first to put SQL database on the market. There were network
> based DBs at the time, I remember serios articles which claimed that
> SQL-based DBs would never reach network DBs performance. Somehow
> network DBs died out together with CODASYL. There were several attempts
> to create OO DBs, which, when you think of it, is just another
> implementation of network model.

Well, maybe even demonstrated performance pales against targetted solutions.

>
> Sorry, SQL found it's way to the top of DB world in tough competition.
> And Oracle is not the only implementation. DB2, MS SQL-Server, Informix,
> Sybase are all SQL-based databases.

That first sentence is not true. SQL found its way to the top of the DB world because it was subsidized by the US Government. Government contracts provided the cash flow both to originally build the product and to sustain a continually growing public corporation. The tough competition wasn't technical, it was marketing. And there was no competition in the standards creation, which later drove all those implementations. Most large infrastructural investment is done by governments (or trusts, which is another long story), so if you want major improvement, you need to have the revenue to drive it. So, we should raise progressive taxes, lay fiber everywhere and give everyone minimal free broadband access, eh?

Of course, there's always the improper booking of sales revenue to maintain stock prices, but most large software companies have done that anyways when times start to go bad.

jg

--
@home is bogus.
Larry should've offered to extend that California contract to
unsupported named licensing for all CA residents, in addition to
supported licensing for all state employees.