Re: The MySQL/PHP pair

"Dawn M. Wolthuis" <dwolt_at_tincat-group.comREMOVE> wrote in message news:cmej17$cug$1_at_news.netins.net...

> I was basing this statement on SQL-99 as much as Date & Darwin. Given
that
> nested structures are now supported by the current SQL standard, even
though
> not widely used, I figure that at least gives some credibility to nested
> relations. I do realize there is little consensus on whether or how to
> employ such features.

That's an important point. I was unaware of the SQL-99 standard. Maybe Joe can give us a little help in that regard.

The term "credibility to nested relations" is a little too fuzzy for me. I'd rather split it out into two issues:

First, do nested relations work?
Second, are nested relations useful?

For the first one, I'd point back to the mathematics, as I think you have. Nested relations certainly "work" in math, beyond all reasonable doubt.

For the second one, your comment about "little consensus on whether or how to employ such features" addresses this. My position is simple: I'll use them when I need them, and I haven't needed them yet.

> SQL still has its fans, but with nothing else on the horizon, I think
XQuery
> could be where we are headed. I'm not as sad about that as others are
since
> it does handle more of what I'm looking for than does SQL. It isn't
likely
> to make too many DBAs or DB professionals happy, however.
>

I'm a fan of SQL. It works. But I am not going to let my favorable opinion of it blind me to either the real objections that nearly all relational theorists have made of SQL, nor to the prospect that, sooner or later, something better than SQL will take its place.

I'm even open to the suggestions made by some of the theorists in this newsgroup that a few of these better languages are already in evidence. I'll learn one of those languages when I need to. I haven't needed to yet.

> And my experience, as mentioned before, is different in that I have seen a
> query language that works with not-exactly-relational structures (non-1NF)
> that is easier to use in many respects than is SQL.

Ease of use is only one aspect of cost versus benefit. BASIC is easier to use than Pascal. But there are some ways in which Pascal is, nevertheless, a better language. It depends on what you are trying to do.

> UniVerse & UniData are two implementations of the same data model as Pick.

[snip]

> I understand it -- is almost invisible to those who discuss the IT
industry.
> It has been pronounced dead many times.

Thanks for the detailed narrative about Universe and Unidata. The term "practically invisible" carries a lot of connotations.

For example, "object oriented programming" was practically invisible in the IT industry from the inception of Simula in the mid 1960s to the inception of Smalltalk in the mid 1980s. And OOP didn't really become mainstream among practitioners until the 1990s with C++ and Java.

Relational Databases remained a theoretical possibility from 1970 to the late 1970s when a Multics based Relational DBMS was produced. Even then, I would say that the mainstream database jockeys stayed with products like IMS and IDMS, for various reasons, until the late 1980s, when SQL emerged as the lingua franca for data exchange. Received on Fri Nov 05 2004 - 14:26:08 CET