Python and COBOL

[François Pinard]

[Ian Parker]

> It did have interesting concepts:  

> (iii) The PERFORM for calling procedures once or repetitively (which I
> remember as being deeply cool).

Notably, you had the "PERFORM paragraph-1 THRU paragraph-2.".  You never
knew for sure where a subroutine started or ended, just looking at it.
You had to find those PERFORM first, to understand subroutines boundaries.

Even worse was "ALTER paragraph-1 TO PROCEED TO paragraph-2.", with the
limitation that the first given paragraph should contain a single GOTO
statement.  It had the effect of altering that GOTO statement at run-time,
so it jumps elsewhere than what it textually looked in the source.  So, you
might not even believe that a GOTO was being executed as written.  I once
worked in a series of programs making a notable abuse of these things,
and in my opinion, that system was both unreadable and unmaintainable.

COBOL was designed to be readable, but to succeed in this, any team also
needed books to rule tons of conventions and verboten writings.

Python is designed to be readable.  The only rule is to avoid lambda's :-).

P.S. - And before anyone feels hurt, remember that I'm always a bit kidding.

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard