Re: Bit strings and semantics
Victor Yodaiken (yodaiken@sphinx.nmt.edu)
Thu, 27 Oct 1994 08:27:21 -0600
On Oct 26, 4:52pm, Randall Holmes wrote:
Subject: Bit strings and semantics
>
>With a suitable ADT interface, bit strings can be made to look
>like objects of the most amazing and diverse sorts!!! (lists
>(even infinite ones!), trees, formulas of your favorite language,
>arbitrary-precision integers, maybe even ZFC sets :-)
>
>So the argument against the semantic, non-formalist attitude is
>belied by C. S. practice. We are not exhorted to remember that objects
>of the usual CS types are implemented as bit strings in various odd
>ways -- in fact, we are usually supposed to forget this!!!
With dismal consequences for those who mistake floats for real
numbers, for example. In any case, I don't want to make an argument
for "formalism", I just want to recall that any theorem proving
program will, ultimately, just be shuffling syntax.