Re: J4 - presentation/discussion on "Future of the COBOL Standard"

Actually, I have used MAX, MIN, ORD-MAX, ORD-MIN with an ALL subscript and
a 
fixed number of occurrence table.  The "trick" is to fill with high-values
or 
low-values before populating the table.  for example pre-fill with
low-values 
and then use
  Function Max (elem (all)) ...

and it will happily find the maximum entry.

SUM could work if you initialize to zero.

I don't know how "performance" OR "maintainability" would work when
compared to 
using an ODO for the same logic (and not having those "extra" table
entries).

As far as Pete's question, the date, numval and the min/max, ord/char
functions 
are the ones that I think have been MOST commonly used (at least in IBM 
mainframe and Micro Focus applications that I know).  All of these were
from the 
'89 Intrinsic functions amendment and not added by the '02 Standard.  (OK,

random also gets used - but it always surprises people how it actually
works - 
and I have  also seen REM/MOD in more than one application.)  I guess I
have 
seen everything except the statistics and trig functions <G>.

The two functions that I "like" from the '02 Standard - that I don't think
are 
used much (and I am not certain who has implemented them) are:
  highest-algebraic
    and
  lowest-algebraic

As someone who (back in my paid programming days) occasionally fell into
the 
"trap" moving zeroes or 9's to a numeric field and NOT getting its 
highest/lowest value, I think these were really "nice to have" new
features.

-- 
Bill Klein
 wmklein <at> ix.netcom.com
"Pete Dashwood" <dashwood@[EMAIL PROTECTED]
> wrote in message 
news:63p7ugF27vleuU1@[EMAIL PROTECTED]
>
>
> "Rick Smith" <ricksmith@[EMAIL PROTECTED]
> wrote in message 
> news:13teonpobo7kve4@[EMAIL PROTECTED]
>>
>> "Pete Dashwood" <dashwood@[EMAIL PROTECTED]
> wrote in message
>> news:63p2sfF28n6tgU1@[EMAIL PROTECTED]
>>
>> [snip]
>>
>>> (Yawn)... so I should use ODO to save a few milliseconds when I need
to
>> sort
>>> an array?
>>>
>>> But I DON'T need to sort an array... ever. And if I did I'd take the
>> hit...
>>>
>>> Next point?
>>
>> Intrinsic functions using the ALL subscript will work with
>> ODO tables, but likely will not give the correct result when
>> using fixed length tables with a variable number of items.
>>
>> MAX, MEAN, MEDIAN, MIDRANGE, MIN,
>> ORD-MAX, ORD-MIN, PRESENT-VALUE,
>> RANGE, STANDARD-DEVIATION, SUM, and
>> VARIANCE are those affected.
>>
>>
>
> Fair enough.
>
> Again, facilities I wouldn't use often enough to lose sleep over... :-)
>
> I don't think I've ever used ALL as a subscript, and MAX, MIN, and SUM
seem 
> more like SQL to me than COBOL... (I have used those in SQL...). So it
looks 
> to me as if my COBOL use may be "limited", notwithstanding the fact I
used it 
> successfully for decades. :-)
>
> I'm wondering what pro****tion of the COBOL user base are using these
features 
> on a daily basis?
>
> Perhaps this is what Bill meant by a "bloated" standard...? Some cool,
but 
> esoteric features. Should the Standards committe be spending time on 
> do***enting them and then expecting vendors to implement them, if there
is a 
> tiny user demand for them? \
>
> Or is it me? :-)
>
> Pete.
> -- 
> "I used to write COBOL...now I can do anything."
>