Re: Hashing for DISTINCT or GROUP BY in SQL

From: paul c <anonymous_at_not-for-mail.invalid>
Date: Fri, 15 Oct 2010 20:52:40 +0000 (UTC)
Message-ID: <i9aeun$ksl$1_at_tioat.net>


On 15/10/2010 12:00 PM, Anne & Lynn Wheeler wrote:
> in the late 70s, there was contention between the 60s "physical" IMS
> dbms group and the system/r group. The IMS group contending that the
> implicit index of system/r doubled the physical space needed on disk and
> there could be 4-5 times increase in disk i/o (processing index). the
> system/r group countered that the "implicit" index of system/r
> significantly reduced the human and adminstrative overhead involved in
> managing a large IMS database (direct record pointers exposed as part of
> the data a programmer had to handle ... and be updated if the DBMS was
> re-organized).

I've heard that top IBM management didn't understand much of these kinds of arguments so they employed glib advisers to tell them what the above meant, except that they were usually instead told what they wanted to hear. By the 1990's all large customers were doing the same, only with 'advisers' from the big consulting houses. The ones I met were at heart very slimy characters. By then, most programmers, if not their managements, knew that publications like ComputerWorld and Datamation were respected only by idiots, since that's what most of the editors were.

Also heard that IBM's mainframe salesmen who made big commissions selling the hardware to run IMS (not sure exactly when the software itself was 'unbundled') ran an active campaign within IBM to discredit Codd. It was very personal and nasty and may have caused him to have a stroke.

Later I met an Amdahl salesman who said, "give me more of this relational stuff, I can sell all the 'boxes' it needs". When a boss of mine plunked for an Amdahl cpu, the local IBM field manager invited him over for coffee and 'career counselling'. That was in the mid-1980's. By the early 1990's even IBM could see which way the wind was blowing and introduced the RS6000, basically a Unix machine, an attempt to hedge their bets, having already blown their early PC lead. In other words, just like other people do now, they knew something big was happening, they just didn't know what.

Just before Clinton hired him, an academic named Robert Reich toured the continent giving speeches about the new 'service economy' and touting IBM as a company that embraced the new world order. A few hundred Canadian liberals hissed at me when I told him it was only economic desperation, today not many popular pundits want to talk about service economies. Amdahl people imagined that they fell down the mainframe toilet when their genius president bet on the wrong kind of cmos (Gene Amdahl had left his namesake company long before), but in the bigger view the mainframe 'bricks and mortar', as it were, was irrelevant.

At the same time (early 1990's), Steve Balmer had briefly retired from Microsoft with his early profits and Amdahl tried but failed to hire him, still dreaming that technology can be managed by the right person (today, whatever crystal ball skills he has aren't stopping erosion of the Microsoft franchise, who knows when that will turn into an avalanche). Not much later, Codd was willing to go to Amdahl to take over one of their software products and there was no argument about money (it was big money), it fell through because they wouldn't give him the title he wanted.

Commodity prices are probably the much bigger economic influences, but computer development in the last twenty years has effected similar upheaval and confusion as the first fifty years of the printing press except for much more economic waste. 'Programming skill' is now a commodity, just as the typical university degree is, even if the word 'skill' doesn't mean what it once did. (I see US President Obama is now touting 'all kids need a university education', a theme that Canada embraced in the 1960's but which is now discredited, the average university graduate having less command of English than the high-school graduates of fifty years ago. One such math and cs graduate, frustrated because he couldn't explain his new feature, said to me "well, we aren't English majors!")

(There were lots of other revealing anecdotes, those are just ones I saw personally.)

Meanwhile google might have the biggest 'db' (as far as many people are concerned) ever and which happily ignores RT, massive disjunctions being its idea of a 'join'. Bound to happen, if it hasn't already started, that businesses will use google methods to manage their accounts. A skeptic like me can't wait to see that chaos. Received on Fri Oct 15 2010 - 22:52:40 CEST

Original text of this message