Subject Re: SQL Standard (was Circular Foreign Key Dependencies)
Author Jim Starkey
At 11:57 AM 3/24/01 -0500, Ann W. Harrison wrote:
>At 10:32 AM 3/24/2001 -0500, Jim Starkey wrote:
>> But then
>>I'm too cheap to buy a copy of the standard. Exactly why
>>a standard is protected like a classified document is a
>>different question.
>Because even standards committees can't live on air.

This is a little off topic, but let's bat it around a bit

Let's take a short poll. How many people on this list,
which, after all, is devoted to the discussion of database
architecture, have access to a current copy of the official
SQL standard? Of those who have a copy, how many use it
as a reference work for writing standards compliant application

I don't have a copy. In fact I've never seen a copy. When
I have a question I mooch off Diane Brown.

Let's compare the SQL standard to the Internet standards (which
generally preceded it) and the Java standard (which followed it).
Both the Internet RFC's and the Java language, base class, and
class file specifications are available on line, for free.

The TCP/IP, name service, ftp, telnet, and httpd have thousands
of completely compatible implementations that make living in
this century a gas. When I want to write something that talks
to a SMTP server, I get a copy of the RFC, and if I follow the
same published rules at the guys who wrote sendmail, qmail, etc.,
my program will work just fine. This is good stuff.

Java is arguably the most compatible computer language written.
Write a Java program against the spec (the best reference work
on the base classes comes from Sun), compile it with any of a
dozen Java compilers, and the resulting class files will run
in any a dozens JVMs (yea, even mine, often). This is really
good stuff.

Now look at the SQL standard. Unlike the Internet and Java
standards, you can't just get a copy. And if you had a copy,
it wouldn't help you one bit one iota because:

1. Not a single DBMS in the world complies with the

2. Mutually incompatible systems can each comply with
the standard.

3. No DBMS implementor would ever dream presenting an
idea to the SQL committee for a blessing before
implementing it.

Now SQL isn't the only worthless standard; POSIX is nearly
as bad (just watch the trace from "make config" for an open
source product to see a catalog of 10,000 mutually incompatible
ways to comply with the standard). And surprise! POSIX is
also a secret standard!

The difference between the Internet and Java, on one hand,
and SQL and POSIX, on the other, is that the Internet and
Java are intended to promote interoperability, and SQL and
POSIX are intended to let vendors claim standard's compliance
which protecting their market share with non-interoperable
implementations and a high barrier to entry.

"Because even standards committees can't live on air."
Ridiculous! What are their expenses? The members have
their salaries and travel paid by their organizations.
Postage? Maybe they could figure out e-mail and the web.
Meeting sites? Maybe some the member organizations have
conference rooms.

This is absurd. A living, accessible standard is an important
precondition for technology evolution -- if we can't agree
on the foundation, how can up build the upper stories?

Jim Starkey