Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!usc!elroy.jpl.nasa.gov!turnkey!orchard.la.locus.com!prodnet.la.locus.com!jfr
From: jfr@locus.com (Jon Rosen)
Newsgroups: comp.databases
Subject: Re: SQL Standards
Message-ID: <19915@oolong.la.locus.com>
Date: 26 Nov 90 20:15:00 GMT
References: <5124@baird.cs.strath.ac.uk>
Organization: Locus Computing Corp, Los Angeles
Lines: 50

In article <5124@baird.cs.strath.ac.uk> jnw@cs.strath.ac.uk (John Wilson) writes:
>I have heard mention recently of SQL 2 and SQL 3.
>
>I suspect that SQL 1 is the 1986 ANSI standard and that 2 and 3 are
>subsequent revisions.
>
>Is anyone out there aware of these revisions? Is there published
>documentation describing the modified standards?
>
>John.

This gets really complicated but try to follow it if you can...
 
ANSI SQL is the adopted standard for the US, as well as by ISO 
(international standards organization) and FIPS (the federales)...
 
ANSI SQL is a wimp (i.e., weak immature miserable and puny, not
windows, icons, mouse pointing) standard which copped to much of
IBM did in DB2 and left out major areas such as DROP <anything>,
indexes (indices if you prefer), referential integrity, etc.
 
The Integrity Enhancement Feature (i.e., referential intgerity)
was defined as an extension to ANSI SQL but I do not believe that
it has been fully adopted (I may be wrong on this point)...
 
SQL2 has been proposed as a draft standard.  It adds DOMAINs, full
referential constraints including general ASSERTIONs, DROP <anything>,
FETCH orientation (i.e., fetch last or previous, as well as next),
embedded SQL definitions, outer joins (actually any join can be
expressed with SQL 2), DATETIME, and some other stuff.  
 
SQL3 is integrated into the SQL 2 proposal with a bunch of *** This
is only for SQL 3 *** type remarks... SQL 3 includes triggers and
a lot of extensions in the manipulation language (I think joined
updates are included but the document is pretty hard to decipher).
SQL3 is also only a proposal.
 
Along with all of this, there are two ISO standards being considered
for international use and the SQL2/3 document merges all of this
into one big mish mash.  

 
I think SQL2 and SQL3 are quite a long way off from reality.  Even
if a consensus develops (and it is certainly like designing a camel
by committee), approving the draft standard and then having vendors
respond will take a long time.  Remember that ANSI SQL was in fact
a reduction of practice to definition.  SQL2/3 will require building
new practice out of a definition.  This is much harder to accomplish.
 
Jon Rosen