Re: Principle of Orthogonal Design

From: David BL <davidbl_at_iinet.net.au>
Date: Fri, 8 Feb 2008 18:41:19 -0800 (PST)
Message-ID: <d3188d42-b65a-403b-861c-b7a7066a066a_at_e25g2000prg.googlegroups.com>


On Feb 9, 9:26 am, Marshall <marshall.spi..._at_gmail.com> wrote:

> However
> quantifiers in logic come with their own lexical scope. (They
> must, since they introduce local names.)

This is a good point, and there are constructs such as for integration, product, summation and limit that also introduce local names that are invisible to the containing context, and are said to be "bound", not "free".

I think an explicit syntax for introducing a lexical scope for a variable that is explicitly bound to some expression works nicely.

Eg (let a=x in f(a)) is equivalent to f(x)

and the query

    P(a=y, b=x) & Q(a=x, b=z))

where P,Q are relations with attributes named a,b would instead be written as

    (let a=y,b=x in P) & (let a=x, b=z in Q) Received on Sat Feb 09 2008 - 03:41:19 CET

Original text of this message