TUNES digest, Vol 1 #356 - 8 msgs

Pascal Bourguignon pjb@informatimago.com
Fri May 30 11:17:02 2003


Alaric B Snell wrote:
> Now, I see a lot of fighting between the strict typing people and the 
> non-strict typing people; it's part of a large battle between "rigorous 
> careful engineering" and "throw it together". This battle has come up in 
> debates about schema languages on xml-dev, and in things like the 
> Extreme Programming movement.

When talking about typing, please keep in mind this:

[pascal@thalassa pascal]$ clisp
;; Loading file /home/pascal/.clisprc.lisp ...
;; Loaded file /home/pascal/.clisprc.lisp
[1]> (+ "toto" 2)

*** - argument to + should be a number: "toto"
1. Break [2]> (quit)

[pascal@thalassa tests]$ cat c-faiblement-type.c
#include <stdio.h>

    int main()
    {
        printf("\"toto\"+2=%d\n","toto"+2);
        return(0);
    }

/*** c-faiblement-type.c              -- 2003-05-29 18:46:36 -- pascal   ***/
[pascal@thalassa tests]$ cc -o c-faiblement-type c-faiblement-type.c ; ./c-faiblement-type 
"toto"+2=134513910



With respect  to compilation/interpretation, I would say  that this is
only   a  efficiency   optimization   that  should   not  concern   us
programmers/system builders.   That is, if you need  to have automatic
type checking (and  yes we need it!),  it has not to be  linked with a
compilation step.   Linking any such kind of  analysis/checking of the
program/system  to  a  compilation  step  is too  limiting,  and  miss
possibly a  lot of errors.   Of the kind  of the bug which  caused the
Mars Climate Orbiter crash.



Jeff Cutsinger wrote:
> A lot of the literature on software engineering incorrectly assumes 
> total knowledge is available to begin with, so the designer produces a 
> single monolithic system that precisely meets those needs...

That would be an argument for dynamic type checking. If the programmer
can't have a total knowledge of  the system he's building, how can you
expect his tools to have it?  Even in a pure open source environment.

 
James Michael DuPont wrote:
> The entire thing that makes perl great is that you can write baby perl,
> simple online statements like in sed or grep that do things that are
> useful. That is the macro system to the extreme.....

That's not a language feature, but a tool feature:

[pascal@thalassa pascal]$ clisp -q -norc -x "(prog1 (values) (with-open-file (in \"/dev/stdin\" :direction :input) (loop with lines = nil for line = (read-line in nil :eof) until (eq line :eof) do (push line lines) finally (dolist (line lines) (format t \"~A~%\" line)))))"<~/tmp/misc/wang |head -8
32 -that's all-
31 
30 
29 
28 
27 publie en 1962 par MIT Press, un des maitres-livres de l'Informatique.
26 1960), et il a ete repris dans le celebre 'LISP 1.5 Programmer's Manual'
25 de  LISP  (LISP  1,  sur IBM 704 justement, le manuel est date  de  Mars


What's more, it's trivial to add this feature to any language:

[pascal@thalassa pascal]$  ~/tmp/tests/c-script 'int i=0;while(i<5){printf("%03d\n",i);i++;}'
000
001
002
003
004
[pascal@thalassa pascal]$ cat ~/tmp/tests/c-script
#!/bin/bash
cat >/tmp/$$.c<<'EOF'
#include <stdio.h>
int main(int argc,char** argv)
{
EOF
echo $@ >> /tmp/$$.c
cat >>/tmp/$$.c<<'EOF'
    return(0);
}
EOF
cc -o /tmp/$$.exe /tmp/$$.c && /tmp/$$.exe



-- 
__Pascal_Bourguignon__                   http://www.informatimago.com/
----------------------------------------------------------------------
Do not adjust your mind, there is a fault in reality.