Re: Whatever happened to BS-12?

From: paul c <toledobythesea_at_oohay.ac>
Date: Mon, 20 Nov 2006 02:17:09 GMT
Message-ID: <Fe88h.344288$R63.15040_at_pd7urf1no>


paul c wrote:
> jlfoster wrote:
>

>> "Roy Hann" <specially_at_processed.almost.meat> wrote in message 
>> <news:8JidnWPKlZm9O8HYRVnyvw_at_pipex.net>...
>>
>>
>>> "Bob Badour" <bbadour_at_pei.sympatico.ca> wrote in message 
>>> news:vX07h.20787$cz.317781_at_ursa-nb00s0.nbnet.nb.ca...
>>>
>>>> jlfoster wrote:
>>>>
>>>>> And how in hell did System R (and the SQL nightmare) get so 
>>>>> popular? Feh.
>>>>>
>>>>> http://www.mcjones.org/System_R/bs12.html
>>>>
>>>>
>>>> It was good enough.
>>>
>>>
>>> And it had an IBM badge on it.  For modern readers, consider how 
>>> today the Microsoft badge protects the technology buyer from rebuke 
>>> when it turns out badly.
>>
>>
>>
>> Wasn't BS/12 also developed by IBM?
>>
>>

>
> Yes, and Hugh Darwen (of third manifesto fame) was one of its developers
> for some years - I don't know if he was chagrined when IBM then assigned
> him to the SQL standards committee. There is a brief summary of BS12
> written by Darwen somewhere on the web (and another by David Maier in
> one of his books, sorry I don't remember where and which). Maybe that's
> where Darwen mentions how the Yank IBM lab took a look and decided to go
> their own way (I suspect because they didn't understand it, Darwen is
> probably too polite to say that).
>
> I believe some of the early bright relational lights such as Stephen
> Todd also worked on BS12. If the microprocessor and copious main memory
> had been dominant then, I wonder if developments might have ended up
> quite different.
>
> p

Another thing that puzzles me is why BS12 didn't last. My guess is that   a big reason has to do with politics. There was one part of the BS12 development model that I think was exactly right - I believe it was used for customer applications as soon as possible, maybe sooner than those of System R, but because it was a service bureau offering, my hunch is that customers did very little programming in it and likely no modifications at all - the developers were forced to stay close to actual customer problems. The reason I think this is a good way to go about things isn't that I think the developers are necessarily smarter, they are just fewer but more focussed and varied incoherent competing approaches are minimized that way. Contrast with the typical open-source project. I remember a teenage postgresql developer giving a presentation which concerned mods that I thought were extremely dubious.   His father was in the audience video-taping him.

p Received on Mon Nov 20 2006 - 03:17:09 CET

Original text of this message