Re: C.J. Date class "Relational Remodeled"

On 6 Feb 2002 06:05:48 -0800, michael_at_preece.net (Mike Preece) wrote:
>If anyone is interested, there were quite a view threads concerning
>RVAs on comp.databases.pick back in '93 & '94.

I found one.

>> http://www.uclaextension.org/shortcourses/spring2002/relational_remodel_sp02.htm
>>
>> Did I read that correctly?
>>
>> >>>> RVAs are legal!
>>
>> Wow! No more first normal form then? :-)
>>
>> -- Jan Hidders

OK, what is a relation valued attribute then?

Hmph. Get Date book off shelf, copyright 2000 edition, p 127ff, ... had this stupid book on the shelf for two years, barely opened it, ... Date's writing is kind of opaque, is that a negative recommendation for his course?

OK, back to RVA's. Book seems to suggest actually stuffing row(s) into attribute, rather than, say, an RVA being a pointer to a row in a conventional table.

"No more 1NF then?" you ask with a smiley. Hmmm.

Well, for ten years we've been watching as our good friend C.J. slooooowly tried to adopt himself to any database idea not purely and simply relational. The early years were not pretty. Even his recent Manifesto is ambivalent, IMHO. Perhaps what we have with the RVA is, even in the 2000 year book, not quite ready for prime time?

Hmm. Chapter 25, object/relational, sez "first great blunder" is saying the relvar=class, as opposed to domain=class. Sorry, C.J., I can't agree to that. Especially if you then go on to sell me RVAs, in place. And, what does this chapter say is the "second great blunder"? Mixing pointers and relations. Oh, my. Do I really want to take a class from this guy?

That's kind of why I started this thread. And, you may have prodded me to an answer, in the negative.

Hmm, I may never actually force myself to read this book unless I fork out $1k for the class, reason enough, I suppose, ... even if I disagree with a lot of it?!?!? If only the price included donuts and coffee.

Joshua Stern
JRStern_at_gte.net Received on Wed Feb 06 2002 - 18:58:55 CET