Re: An Analogy ?
From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Tue, 09 Nov 2004 08:42:57 -0500
Message-ID: <14u662-c2k.ln1_at_pluto.downsfam.net>
>> Tony Douglas wrote:
>>
>> > Here's a Monday morning throw-at-the-wall post for everyone !
>> >
>> > Reading some of the recent posts, I got around to thinking about
>> > FORTRAN and SQL as analgous to each other. In both cases, they were
>> > essentially experiments to show that something could be done (in
>> > FORTRAN's case, that a compiler could produce executables almost as
>> > efficient as hand-written assembler code, in SQL's case as part of an
>> > experiment to show that a "relational" system could work). In both
>> > cases, special cases and "clever" workarounds to problems to allow the
>> > experiment to work became embedded in the experiment (e.g. the close
>> > tie-in between FORTRAN's control structures and the IBM mainframe
>> > model it was originally built on).
>> >
>> > And, I would go on to suggest, both experiments have outlived their
>> > genuine usefulness and served to impede progress in their respective
>> > areas (that is, we're still stuck with FORTRAN 9x and SQL:2003).
>> >
>> > Any thoughts, anyone ?
>> >
>> > - Tony
>>
>> Well, if somebody were to produce a better relational query language
>> tommorrow, the commercial vendors could implement it as an alternate
>> interface to the data, without in any way touching their current SQL
>> implementations.
>> plans.
Date: Tue, 09 Nov 2004 08:42:57 -0500
Message-ID: <14u662-c2k.ln1_at_pluto.downsfam.net>
robert wrote:
> Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net> wrote in message > news:<kad462-uov.ln1_at_pluto.downsfam.net>...
>> Tony Douglas wrote:
>>
>> > Here's a Monday morning throw-at-the-wall post for everyone !
>> >
>> > Reading some of the recent posts, I got around to thinking about
>> > FORTRAN and SQL as analgous to each other. In both cases, they were
>> > essentially experiments to show that something could be done (in
>> > FORTRAN's case, that a compiler could produce executables almost as
>> > efficient as hand-written assembler code, in SQL's case as part of an
>> > experiment to show that a "relational" system could work). In both
>> > cases, special cases and "clever" workarounds to problems to allow the
>> > experiment to work became embedded in the experiment (e.g. the close
>> > tie-in between FORTRAN's control structures and the IBM mainframe
>> > model it was originally built on).
>> >
>> > And, I would go on to suggest, both experiments have outlived their
>> > genuine usefulness and served to impede progress in their respective
>> > areas (that is, we're still stuck with FORTRAN 9x and SQL:2003).
>> >
>> > Any thoughts, anyone ?
>> >
>> > - Tony
>>
>> Well, if somebody were to produce a better relational query language
>> tommorrow, the commercial vendors could implement it as an alternate
>> interface to the data, without in any way touching their current SQL
>> implementations.
> > isn't that what DataPhor(sp?) is said to do?? > > In fact, presumably it would generate the same execution
>> plans.
Have not looked closely enough to say, but I know it is more than that, an attempt I believe to implement a more purely relational database. Others who are more familiar can jump in perhaps and tell us more.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Tue Nov 09 2004 - 14:42:57 CET