Re: sql views for denomalizing

From: David Cressey <david.cressey_at_earthlink.net>
Date: Sat, 30 Jul 2005 22:55:40 GMT
Message-ID: <MzTGe.7529$6f.6337_at_newsread3.news.atl.earthlink.net>


"Marshall Spight" <marshall.spight_at_gmail.com> wrote in message news:1122759998.556491.146100_at_z14g2000cwz.googlegroups.com...

> For sure. As an aside, it's quite clear to me we need a really
> good combined data/programming language. I'll get right on it.
> Writing code in a statically typed language to assemble untyped
> strings which are then passed to a statically type SQL is
> teh suck. I want unified typechecking between my application
> language and my data language!

You might have liked "MDL". You can look it up in Wikipedia or FOLDOC. I was one of team of five people that originated that language, back in 1970. My job was memory management, the automatic garbage collector being the bulk of the effort.

You can bet I was interested in "transitive closure" at that time. Although I didn't know that term for it until I read it in here, from you. My formal education is sadly lacking.

Anyway, we started with Lisp and added explicit data types, arrays, and a bunch of other neat stuff to the language. It actually lasted for quite some time before it died out.

The reason I say you might have liked MDL is that the data types went with the VALUES, and not with the variables. So variables were not bound to types, and type checking was deferred to run time.

But of course it wasn't at all like Dataphor or SQL.

> with a new language. As Brooks made clear, there is no silver bullet.

yep. Received on Sun Jul 31 2005 - 00:55:40 CEST

Original text of this message