At 04:11 PM 3/4/98 -0800, Donna Garverick wrote:
>Glenn Cole wrote:
>
>> Tail recursion says RECURSIVE PERFORM with "Q"uestionable severity.
>> Linking, however, results in a runnable program that works fine.
>
> (getting out my old fogey hat...) i remember way back when on my series
>70, i wrote a little cobol program that used recursion.

One of the advantages of a stack architecture was its relatively easy
way of utilizing recursion which simplified some types of tasks.  Obviously
one was limited by stack space for the markers and any argument pointers.

Back in '76 one of my first contributions to HPIUG(?) (now Interex) was
a FORTRAN subroutine for performing a type of 'partition' sort which employed
recursion.  I recall having wrote the same routine in both SPL and COBOL
also at the time, but only the FORTRAN version survives.  I seem to recall
it sorting a 1,000 values in well under 1 second on a Series I system but
I'd have to go to the garage and get into my archives to dig this out.

The routine itself was published in Volume 1, Issue 1 of the new publication
for HPIUG (The HPIUG Journal?).

Might be interesting to see what it would do on MPE/iX in native mode.... :)

(Donna....I'll pass that old fogey hat back now... :)




/jf
                              _\\///_
                             (' o-o ')
___________________________ooOo_( )_OOoo____________________________________

                        Thursday, March 5th

           Today in 1770 - The Boston Massacre took place.

___________________________________Oooo_____________________________________
                            oooO  (    )
                           (    )  )  /
                            \  (   (_/
                             \_)