Re: Why is "group by" obligatory in SQL?

From: paul c <toledobythesea_at_oohay.ac>
Date: Fri, 24 Jul 2009 16:01:21 GMT
Message-ID: <lVkam.37959$PH1.15247_at_edtnps82>


Cimode wrote:
,,,
> BS12 was developped after IBM started having second thoughts about the
> massacre of System R massacre and *after* Oracle was on the way to
> build an empire. The context explained below
>
> http://www.mcjones.org/System_R/bs12.html
>
> gives very interesting clues about the context of that second
> attempt...
> ...

IBM's "Service Bureau" operations on all continents were basically a time-sharing service for customers who couldn't afford their own mainframes or couldn't justify a second mainframe for a single application. I suspect there were other reasons, eg., political, the various semi-autonomous IBM country organizations were famous for their individual 'not-invented-here' thinking. As well, there were likely big operability questions regarding the use of SQL in the various Service Bureaux because most of the customer work was submitted in a batch mode that didn't encourage timely error correction. I remember some of the time-sharing operations in N.A. from those days, the instrumentation experts who were responsible for cpu and other resource billing algorithms had god-like status in their organizations. Once in a while a salesman would quote a fixed price based on loose specs that turned out to be ridiculously low - you should have seen the s... fly in the executive suite when that happened. Just as now, practically no significant decision was made on technical grounds. Received on Fri Jul 24 2009 - 18:01:21 CEST

Original text of this message