Re: Is Ingres yet relational?

On Mon, 31 May 2004 09:05:25 -0300, Leandro Guimaraens Faria Corsetti Dutra <leandro_at_dutra.fastmail.fm> wrote:

>> > Ingres wasn't originally an SQL DBMS.
>>
>> Nor a truly relational one. SQL is only a part of the problem.
>
> I never studied QUEL. I am interested in it because it is
>said to be 'more relational' than SQL -- I don't know to which point.

I studied it many years ago, and it was closer to the relational calculus than SQL.

SQL was failed attempt to make predicate calculus less mathematical to the hackers (in the bad sense of the word).

> Indeed the current reference includes duplicates and NULLs, so
>it isn't *that* better.

And it has poor database independence. That's why I am highly skeptical about the conversion of Ingres in a TRDBMS.

>>> I submit to their own loss...
>>
>> It is pragmatic.
>
> I've heard the argument to the contrary by quite expert
>programmers.

It depends on many things. Short code well written in modern languages and well documented may be very useful, specially if you can extract complete routines and to reuse them.

But at practice it is difficult to reuse code that was not written with that intention, and then it is usually distributed in the form of documented libraries.

If you have a good compiled library that works well and that does what you need, you don't have reasons to modify the code. And if you want new features the most qualified people to implement them are the authors. The same for bug fixes. The vendors must fix the bugs for free in a short time.

If a big library does not work well the best thing you can do is to use another that works well.

>> And on many other things. For instance I have never seen a well
>> documented complex open source project.
>
> Free software programmers tend to say the best documentation
>is source code.

A poor excuse for lazyness. Most code is unuseful without good documentation, specially when the code was written in archaic and criptic low level languages like C.

This is one of the reasons because many open source projects progress so slowly or die when the creators abandone them.

I have to mantain a 400.000 lines app writen for others with poor documentation and I know what I am talking about :(

>> I use a lot of free software, and even some open source, but to study
>> and to modify the code of others is a last resource.
>
> Indeed. But there are quite a few instances when this last
>resource is needed.

Yes, it is a good practice to publish the source code when a product is abandoned.

Regards
  Alfredo Received on Mon May 31 2004 - 16:55:08 CEST