Re: Verified compilers?

[Georg Bauhaus <rm.dash-bauhaus@futureapps.de>]

On 08.03.12 21:40, Dmitry A. Kazakov wrote:
> On Thu, 08 Mar 2012 20:02:25 +0200, Niklas Holsti wrote:

>> How else could you explain why "2*3+4*5" means 26, and not 70, nor
>> something else?
>
> Explain? It is a subject of proof. There are more than one way to do it.
> BTW the "meaning" you are advocating for looks rather something like
> 23*45*+.

No!

It must mean

  "at least zero twos followed by a three or more of these,
   with maybe fours coming next, and finally a five".

+ is clearly a macro. And, for proof, in which school did they ever
teach that a star denotes multiplication?  A Kleene star is a Kleene
star.


No!

"2*3+4*5" must have a syntax error. What it really means is this:

Python>>> "2"*3+"4"*5
'22244444'


What? No!

Dumb animal, this Python. The mathematical interpretation is obviously
right:

Perl-DB<1> print "2"*3+"4"*5
26


Excuse me, no! A Programming Language was there first, and it says,

*     2*3+4*5
  Inf
Because, obviously,
*     2*1027
  Inf


Uh, no, not any more.

Are you guys still using that old APL dialect? Look at J for the
real meaning:

    2*3+4*5
46
Because
    3+4*5
23


Nah!

Ever since 1969, SNOBOL-4 clearly shows that
	2*3+4*5
         	2*3+4*5
         	^
stdin:1: *** Illegal character in element

Because a string needs quotes around it, for one thing!



Right, that's been known by Lispers for a long time:

* 2*3+4*5

Error in KERNEL::UNBOUND-SYMBOL-ERROR-HANDLER:  the variable 2*3+4*5 is unbound.
    [Condition of type UNBOUND-VARIABLE]




Really, you can't remove the quotes from a string and expect
it to have meaning? Look, all it means when output is just itself!


SNOBOL4 (Version 3.11, May 19, 1975)
     Bell Telephone Laboratories, Incorporated

	OUTPUT = "2*3+4*5"
END
No errors detected in source program

2*3+4*5
Normal termination at level 0


--
> LISP
* 2 + 3 * 4 + 5

2
*
2
*
3
*
3
*
4
*
4
*
5
* (QUIT)