Re: Model Theory

On Thu, 2002-01-03 at 13:39, Jim Hendler wrote:
[...]
> All
>     without taking a side in the particulars, I would like to point 
> out that there are many kinds of semantics and languages and the 
> notion of knowing "he meaning of what we write down in that language" 
> is not as clear as you say -- consider C++ -- seems to be a language 
> that is quite heavily used, yet I'm having trouble finding either a 
> model theory or a formal semantics for it.


Yikes! C++? Are you sure you're not taking sides, rather subtly?

Working on porting a million-line C++ program from
solaris to AIX in the early '90s taught me much
about the value of formal specifications.

The C++ disaster is perhaps the loudest argument in favor
of precise formal semantics that I have ever seen!

In contrast, take a look at Modula3 some time.

  "Shamefully, Java didn't steal nearly enough from it."
	-- http://c2.com/cgi-bin/wiki?ModulaThree

I think perhaps the C# guys stole a bit more; I hope
to study it more sometime soon...

>
  I think it is important 
> for all of us to keep in mind that the language we are trying to 
> create needs to satisfy (as best as possible) many conflicting needs
>   1) Need for being able to represent information (KR issues)
>   2) Need to be used on the web
>   3) Need to be usable by computer for processing in various ways 
> (implementation issues)
>   This working group is trying to make sure we come up with a language 
> that all of us can use for any of these and sometimes the needs will 
> be in conflict.  Please remain very aware that we are engaging in a 
> CONSENSUS process, and sometimes that requires some give and take on 
> all sides.
>   -JH
> p.s. Note the W3C Guide [1] is titled "The Art of Consensus"
> 
> [1] http://www.w3.org/Guide/

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 3 January 2002 16:01:46 UTC