Re: [xml-dev] functional databases noodling ...

[Frank Manola <fmanola@acm.org>]

Dana--

I've no problem (and to a great extent agree) with what you say about the value of looking at the technology developed in the O2 work.  I wouldn't have commented at all except I was frankly a bit niggled by your remark that "IRIS is just a sort of a rip-off from the O2 object oriented database".  "Rip-off" doesn't suggest the usual interaction of scientists and their ideas, but something much more serious.  I was working on the Probe project at CCA at roughly the same time as the IRIS work was going on, and what Probe was trying to do (at least from the data model end) was extend Dave Shipman's Daplex functional model (also done at CCA).  We understood that IRIS was going along the same route, just as Michael suggested.  Neither of us, as far as I know, was borrowing from O2, let alone "ripping off" from it (I think this is pretty clear from the various papers and their dates).  Also as Michael suggested, neither of us originally thought of ourselves as building an object database as the idea came to be understood, although there was a lot in common (Stan Zdonik, who later consulted on Probe, convinced us differently!).  

--Frank

On Feb 15, 2012, at 1:07 PM, daniela florescu wrote:

>> 
>> What Michael says.
> 
> I don't think "what Michael says", Frank. But I don't think it matters either.
> 
> Most ideas in CS can be be lineaged down in one way or the other in work that
> has been done in 50-60ties (those guys had all the good ideas..:-)
> 
> However, the closest approximation, the and most correct and complete implementation of a
> running functional database was still O2: it was not Oracle, yet was not only a small prototype 
> with a bunch of slides attached to it.
> 
> I strongly advise people to look into query processing in the time of the O2. Lots of lessons to learn.
> 
> Best
> Dana