"Corey Brown" <corey (AT) spectrumsoftware (DOT) net> wrote


Think about object orientation, it is older and it is still a complete
mess.

I don't see many advances in software after The Relational Model.

Can you mention any important advance in software in the last 30
years?

I don't think it is very difficult, but few people tried and they had
not a lot of resources. IMO it would be a lot easier than the average
video game.

There is a profound educational problem that is extremely hard to
solve.

DBMSs not databases, and yes they are so flawed that they should not
be called relational DBMSs.


Regards
Alfredo

"Alfredo Novoa" <alfredo (AT) ncs (DOT) es> wrote

Well sure. A lot of new significant programming languages have been
created
in the last 30 years. Not to mention all the things that have been
created from
those new programming languages including operatiing systems, graphical
user
iterfaces, databases, communications protocols, transport mechanisms,
etc..
I would think that the list is fairly extensive. Considering that 30
years ago
computers lived in closed shops and were only accessible to a handful of
people. Things have changed a lot since the 70s. But I can still see
your point
in that the languages themselves have not evolved much from their roots.
But
still, I think that modern languages like C++ and Java are a heck of a
lot better
than Fortran 77 and Cobol. Wouldn't you agree?

An educational problem in the sense that relational theory is not being
taught, or is it something larger than that?

But each vendor out there is trying to build a better product than the
next guy. If an implementation of the "True Relational Model" would
provide a significant business advantage over other implementations,
why are the vendors not rushing to do a correct implementation? There
must be another reason why the relational model, as the theorists see
it, has not been implemented correctly yet.

Thoughts?

"Alfredo Novoa" <alfredo (AT) ncs (DOT) es> wrote

No, but it confirms my opinion that your optimism is striking. What has been
done since November 2001?

"Bob Badour" <bbadour (AT) golden (DOT) net> wrote


I heared that they had legal problems, and the page dissapeared from
the net. Now their web says that they moved to California.


Regards
Alfredo

"Corey Brown" <corey (AT) spectrumsoftware (DOT) net> wrote


For instance?

Nothing really new.

But in the hardware not in the software.

Indeed, most of them are direct Algol derivations with a few things
taken from Simula 67.

Yes, but modern languages like C++ and Java are not a lot better than
Algol 58-60, Lisp (1958) and Simula 67. Even they are worse in some
aspects.

Something a lot larger. Many universities only teach to follow poor
cookbooks, without any critical thinking and sound understanding. This
is specially true in my country.

No, each vendor is trying to make more money selling any crap to the
next gullible.

A few years ago I worked for a DBMS vendor, and the managers didn't
have a clue about the fundamentals (and me too in that time).

A practical advantage, but probably not a business advantage because
very few people would accept it, and know how to use it.

Among other things because most of then don't know that their
implementations are fatally flawed.

Historical reasons. SQL was an early dirty and incomplete lab
experiment, but the circumstances forced its comercialization, and the
success was so great that the correct implementation and the theory
was forgotten. And when the limitations arised, they were attributed
to the model and not to the crappy implementations.

Another reason is that the understanding of the model in the early
70's was not as good as the current understanding. Even the current
understanding is not complete and there are still some open questions
about view updates, transactions, etc.

Too early standards can be a fatal dead weight for the industry.

Regards
Alfredo

"Alfredo Novoa" <alfredo (AT) ncs (DOT) es> wrote

Yes, hardware has been advancing at an astonishing rate for the last
30+ years. But I also think that the creation of some of the languages
that we use today helped accelerate the creation of personal computers
and other devices.

True, but better in others. That's why languages like C, C++, Ada,
Pascal (to some extent)
and SmallTalk, just to name a few, found mainstream acceptance while
other languages remain
as teaching tools in Universities. But of course, mainstream acceptance
does not prove
technical superiority in any way shape or form.

Ok, now I understand. Yes, i've been there too. Having been forced to
release software to the end user from a prototype stage.

Ok, so there are still open issues regarding the model itself and
possibly its
implementation. I certainly appreciate you taking a few moments out
to share your thoughts on this Alfredo. Thank you.

Regards
--Corey