Re: Hashing for DISTINCT or GROUP BY in SQL

On Mon, 18 Oct 2010 14:33:47 +0000 (UTC), paul c <anonymous_at_not-for-mail.invalid> wrote:

>On 18/10/2010 2:31 AM, Cimode wrote:
>> As years pass by and database science goes further and further into
>> collective IT industry amnesia, I somehow believe*old timers* (note
>> that I do consider myself an old timer) have simply failed to explore
>> the opportunities openned by Codd, and remained, implementation wise,
>> too bound to the IBM historical context of how relational
>> implementation is to be conceived. Such lack of perpective limited RM
>> expressive power and constrained it(unhappy pun!) to empirism rather
>> than applying it to how we conceive information encoding as a whole.
>
>I remember a lot of amnesia! I posed a question I was curious about
>even though it is probably pointless today, to the folklore usenet
>group, about the origin of the term 'field' as applied to file systems.

     alt.folklore.computers.

> It might have been an unfair question since that group is mostly
>interested in the cultural history of hardware and operating systems as
>opposed to all software, but I thought it was more apt for that group
>than this one. There are several mathematical definitions for the term

     Well, no. We deal with all aspects of it. Maybe, nobody knew. I had never considered the issue myself.

>'field' but the term also dates back to the 1940's use of Hollerith
>punch cards and possibly much earlier for all I know (I remember seeing
>punched cards that had pre-printed 'nested fields', eg., telephone
>number might have sub-fields of area code and number.
>
>Anyway, it may be because I wrote the question badly, but so the
>reaction of that group seems to be that there was no mathematical
>inspiration behind the term. I had guessed that possibly the
>originators of Fortran might have thought it had special meaning being
>that the early Fortran files often consisted solely of numbers of
>different kinds and there is a definition of a 'closed field' of numbers
>that some mathematicians use. But logicians at one time also used the
>term to describe the union of what they called 'converse domains'.
>Maybe neither influenced the computer use of the term but so far nobody
>seems certain of that.

     It might just be by extension to another type of delimited area.

Sincerely,

Gene Wirchenko Received on Mon Oct 18 2010 - 20:59:27 CEST