Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: OO versus RDB

Re: OO versus RDB

From: Tony D <tonyisyourpal_at_netscape.net>
Date: 9 Jul 2006 09:22:06 -0700
Message-ID: <1152462126.409523.43920@35g2000cwc.googlegroups.com>

Marshall wrote:
> Daniel Parker wrote:
> >
> > I think what I would like to have is a hybrid language, that would
> > allow me to implement a function with imperative techniques, permitting
> > mutable data structures for building immutable objects, aka
> > StringBuffer/String in Java, and a pure functional higher view. Does
> > that sound sensible? Or stupid?
>
> Sounds exactly right. Functional by default, imperative when necessary.
>

Sounds like a horror; the best of neither world.

> It's also entirely possible (as I think you're implying) to have a language
> that allowed one to write a function that is "pure" (in that its outputs
> depend solely on its parameters) but whose implementation was
> imperative. The functional guys *hate* it when I point this out. :-)
>

I don't see why; it's a trivial observation. You can write pure functions in Pascal or C, after all. All it takes is a little discipline.

> There might even be type system support for distinguishing pure functions.
>

To what end ? And even then, I doubt it; this would be a job for some kind of abstract interpretation, which, since you've allowed side effects and referentially opaque subroutines, becomes a non-trivial pursuit. Received on Sun Jul 09 2006 - 11:22:06 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US