Skribent: Didier Kryn Dato: Til: dng@lists.dyne.org Emne: Re: [DNG] Tcl as a scripting language: 4/2/2025 7pm Eastern
Standard time
Le 02/04/2025 à 23:58, Hendrik Boom a écrit : > On Wed, Apr 02, 2025 at 11:32:07AM +0200, Didier Kryn wrote: .. >> AFAIR (this was long ago), the very Tcl intrepreter is a function
>> your C program can call with, through the calling arguments, an access
>> to C functions belonging to your application. There is, of course a
>> specific Tcl API these C functions must match to, but this is all
>> documented. The Tcl command you can invoque from the command-line is
>> just a main program which calls this function.
>>
>> This very simple but extremely powerfull idea is the very strength
>> of Tcl in my opinion, the weak point being the primitive and boring
>> Basic-like syntax of the Tcl language. I just wonder why not all other
>> interpreters offer the same feature -- IBM's Rexx was doing so.
> This is also hoe Guile interacts with C programs.
> And Guile is an implementation of Scheme -- a richer programming language.
Sure, Guile, IIUC, is a extension of Scheme, which is based on
Lisp. But, OTOH, Tcl comes with a graphical toolkit (Tk), which is
uncommon amongst languages, so that it may be used to provide a GUI to
your application.
Yet, both languages have this pretty ancient feature I dislike
which is "everyting is a function": eg arithmetic operators are
functions and you must express "3+4" as "(+ 3 4)" in Scheme and I guess
it's the same in Guile. This is a return to the past almost back to the
primitive pocket calculators for which you would type "3 4 +" . BTW
Postscript works like this but it was not primarily intended for humans.
I understand it makes writing the interpreters more trivial, but, well!
we're in the 21st century and the computer languages should adapt to
humans, not the opposite. Interpreters and compilers have plenty of cpu
power and ram available to do that.