sommar@enea.UUCP (Erland Sommarskog) (10/31/86)
I've followed the by now quite long discussion on this subject and I fear
I haven't seen all yet. We're still some days back on news in Sweden (Europe?)
Anyway, I'd like to make some comments to what's been said.
Endif, endwhile etc:
Some one - the original poster? - advocated the style of Modula-2 where
you must match an IF with an ENDIF and so forth. So far I think this is a good
approach, Ada uses the same style. However these languages also requires ";".
This solution does not cover assignments (and subroutine calls). To solve
this poster proposed something like
a := b END
This doesn't seem very smart to me. I'd prefer a discreet semicolon than an
clumsy END.
Fortran:
In the discussion of using end of line as a statement delimiter some one
mentioned good old Fortran as an example on this. However this is not
really true. Fortran uses end of line or column 72 whichever that comes
first.
Indentation-sensitive languages:
I think this is good idea. In a way I can't understand why it should be
allowed to write:
PROGRAM Nisse(input,output);CONST A=10;TYPE Vector =
ARRAY(.1..A.)OF char;VAR i:integer;B:Vector;BEGIN FOR
i:=1 TO A DO B(.i.):=0;END.
And having the compiler of an missing END in:
WITH InputData DO
BEGIN
This := 1;
That := DoSomething;
MakeSomeFunOfThis(InputData, Result);
is frustrating. A child could see where the END should be. (OK, computers ARE
stupid, I know.)
An indentation-sensitive language would surely enforce an indentation
standard, which of course be a pain fro any one would like indent in other
way. On the other hand different programmers would have easier to read each
other's code.
Other objections that have been arised I think are minor problems that
can be solved. As been mentioned, there are such languages today. It just
surprises me that no one has mentioned Occam.
And for x<y<z: (And a=b=c): There are to few languages supporting this!