GOTO intoxication
Skip Tavakkolian
fst at mcgp1.UUCP
Tue Apr 26 06:17:46 AEST 1988
In article <310 at unisv.UUCP>, vanpelt at unisv.UUCP (Mike Van Pelt) writes:
> In article <733 at ambush.UUCP> kimcm at ambush.UUCP (Kim Chr. Madsen) writes:
> >Even Mr. Dijkstra had once compared the goto's with alcohol, it's nice
> >when used properly, you get headaches when you use it too much and you
> >don't give it to minors.
> >
> Now THAT'S the best description I've heard yet, and from Dijkstra, no
> less. Who would have ever thought...
>
> Now, if we can just get rid of the Carrie Nation types who go around
> bashing those few necessary goto's with axes... :-)
> --
> Mike Van Pelt Unisys, Silicon Valley vanpelt at unisv.UUCP
> Bring back UNIVAC! ...uunet!ubvax!unisv!vanpelt
I think I may have a solution or two to the goto dilemma:
1) Language sensitive editors running on a bit-mapped display so that the
labels could be printed in larger font sizes (like 96 points :-{) ).
For example:
# # ###
# # ###
# #
goto # ###
# # ###
# # #
# # #
.
.
.
# # #
# # ###
# # #
# do_something();
# # #
# # ###
# # #
This would obviously make debugging easy.
2) Writing a C preprocessor patterned after Ratfor in reverse (lets call it
UnratC). which would take goto statements in C and turn them into structured
control loops (It is time for another great book from Kernighan and Plauger,
the "Software Tools in GNU Lisp, Including UnratC").
If I may make a serious point here, I believe that the use of gotos in machine
generated C code is very useful, Since nobody is ``supposed'' to see them.
This is the case with some special purpose C ``preprocessors'' which take
an embedded statement in something, and convert it to C code. Indeed the gotos
made writing something like Ratfor possible (maybe easier?).
Sincerely
Skip Tavakkolian
UUCP ...!uw-beaver!tikal!cti3b1!fst
More information about the Comp.lang.c
mailing list