craig 81366 (11/1/2012)
All your transcript link provides is a recollection of the circumstances in which it was discovered.There are a number of unanswered possibilities:
- When the problem was discovered, did the team consider INSERT / DELETE scenarios at all?
- Even if they hadn't originally considered the possibility of INSERT / DELETE exhibiting similar unexpected behaviour for the same reason, would they consider those to be a different problem?
- Did the team consider the Halloween problem as exclusively applicable to the UPDATE statement?
- Or perhaps even more strictly, did they consider it to be the specific problem of updating Salaries over the Salaries index?
- When they use the word updates, do they mean changing existing records (UPDATE statement), or any update that changes the state of the database?
When I read what the team published, something over 30 years ago, if I recall it correctly, it was about an update being repeated several times on a row, instead of once. Yes, they looked for the underlying cause, but the problem they addressed was the repeated update. Unfortunately I can't recall the reference - there were rather a lot of papers being published in those days, and it's not easy to remember which journal and with what title everything appeared in. But the System R tyeam published very freely in those days (although they did later see the beginnings of the later "keep it all secret until we have a patent" frenzy) and they published that information quite quickly. You may have seen some earlier comments of mine on this site about the relationship of the IBM library with NRL in the 1960s (I was at the NRL end, while NRL still existed) and that free exchange of information carried on well into the 70s.
I will however, give you something to consider.
Languages, words and names are in continuous evolution to aid communication. Believe me, you're not doing yourself any favours if you reject a logical definition on an historical 'point of order' - regardless of the accuracy of your claim.
It's not a "point of order". It's about what people mean when they use that term. What makes you believe that you youngsters own the meaning, rather than us boring grey-haired old ****s?
Nevertheless, I'm willing to agree to disagree.
OK, so am I. We disagree. That means only one of us can be right, and I know which one it is - while, I believe, you don't.
Up to now, nothing nasty. I'll have to address the rest of your post in a separate message, because I don't want to mix the civilised part of this exchange with the uncilvilised part (and that, incidentally, illustrates one of the differences between us).
Tom