Re: Object-oriented thinking in SQL context?

On Jun 12, 10:42 pm, Marshall <marshall.spi..._at_gmail.com> wrote:
> On Jun 12, 2:58 am, David BL <davi..._at_iinet.net.au> wrote:
>
> > On Jun 12, 3:46 pm, Marshall <marshall.spi..._at_gmail.com> wrote:
>
> > I'd be interested to know what argument won you over - and led you to
> > believe that personal attacks could be a good thing.
>
> > > The thing you're arguing against is incivility.
>
> Once upon a time in cdt, a choice between style and substance
> was made manifest in the form of two of the most frequent
> posters. One of them was entirely pleasant and quite polite,
> even obsequious at time. I am quite sure that if I knew that
> person socially, I'd really like them, and invite them over to
> barbequeues on a regular basis. That person was also entirely
> anti-theory, and argued *against* any kind of formal understanding.
>
> The other person was Bob. I assume he needs no introduction.
>
> The two argued quite a bit, and Bob was quite harsh. I joined
> the fray in opposition to Bob. I argued quite forcefully that, in
> essence, if he couldn't say anything nice he shouldn't say anything
> at all. My first surprise was that Bob had vast popular support
> and I didn't.

Bob's support comes from the substance of his posts, not his vitriol.

> But the big shock was the realization that I was
> here for database theory (like the sign says on the door)
> and not pleasantries, and that it actually kind of *did* come
> down to a choice between one or the other, given how much
> anti-intellectualism there is to go around.

For most people that choice is made in what they choose to read. One of the good things about the Internet is freedom of speech. Do you really think it should be controlled, so viewpoints that you or Bob consider objectionable or harmful can be suppressed?

> Bob's tirades really do decrease the noise.

Brian responded to this point very well. In any case, I fail to see what the problem is. If a poster has nothing interesting to contribute then ignore them. This censorship crusade you're defending is evil.

If you don't trust people to make up their own mind, you could always respond to an obviously bad poster with something short and factual. E.g.

   "That post confused the definition of 5th normal    form making it unintelligible".

It takes less effort and has more venom than a longer post full of insults. In any case, you don't need to be told how to suck eggs and you mostly take the high road. It's one reason I think highly of you.

I came here because I'm researching the concurrent editing of data using Operational Transform (OT). I'm interested in recursive data types that for example are associated with scene graphs. Unfortunately Bob has been so effective at "cleansing" this NG it is difficult to get anything but D&D viewpoints here. Someone interested in a domain like CAD and interested in recursive data types would be viewed as an OO loving idiot by Bob and not last long here. He attacks people that may actually have something interesting to say.

When I first arrived at cdt I was "censored" for suggesting there could be OT based applications where a simple rwlock is more appropriate than a conventional lock manager based on strict 2PL. Bob was derisive, yet his arguments were merely simplistic generalisations - all of which arose from ignorance of OT. For example:

    "Are you suggesting that it is possible to acquire     and release a universal exclusive lock over a     distributed system in less than 200 microseconds?"

In another post he implied incorrectly that distributed database systems must support distributed transactions. Again this stems from ignorance.

His only interesting contribution was a reference to a paper by Jim Gray that claimed update anywhere-anytime-anyway transactional replication was inherently unable to scale because reconciliations grow with cubic order. However in reality Bob was just a VI, and that's the part I find so ironic. He was unaware of the literature on OT and the sense in which Gray's result was inapplicable (even the abstract mentions that commutative update transactions avoid the instability), and yet pretended to be an authority on the subject.

You may be interested in some of the mathematics behind OT. For example lattice theory is relevant to causality preservation in a distributed system. Received on Mon Jun 15 2009 - 09:38:57 CEST