RE: CLEAR Opcode?

I have the impression that many of the users of this technique
mistakenly believe it's more efficient. I seem to remember messages on
this list about coding to avoid a series of data structure subfields and
alphanumeric moves. It's only one instruction, so it must be better!

The second reason it's popular, I think, is that programmers by nature
enjoy demonstrating to other programmers how smart they are. It's a kind
of badge used to identify oneself as a member of the clever programmers
club.

Dave Kahn, TCO, Kazakstan
=========

kahn@tengizchevroil.com   (to November 25)
dkahn@cix.compulink.co.uk (from November 26)


>-----Original Message-----
>From:  Jon Paris [SMTP:paris@ca.ibm.com]
>Sent:  Saturday, 22 November, 1997 01:51
>To:    MIDRANGE-L@midrange.com
>Subject:       Re: CLEAR Opcode?
>
>>>>>>Whereas you should really be shot for using MULT to convert dates  <VBG>
>>>>>> }
>;-)
>
>>>>Boo, boo, boo, Jon.
>
>I don't know who I'm responding to as your name doesn't seem to appear in the
>message, but since there's no smiley or anything in the text,  I'll assume
>that
>you disagree with my comment.
>
>I make _no_ apology for that it.  I just can't understand why anyone would
>code
>this way.  When I was an MIS manager I would have looked long and hard at any
>of my programmers who did this kind of thing.  Other than for our own
>amusement
>from time to time, shouldn't one of the major aims of any programmer be to
>make
>his code both efficient and as simple as possible to understand for those who
>come after them?  In particular I dislike this technique because:
>
>1) It's non-intuitive - if you've never seen it before it can take a long
>while
>to convince yourself it really does work.  The Italian company who are
>working
>with us on Y2K tools had never seen it before (since only North America uses
>MDY it has little applicability in other countires).  It took us quite a
>while
>to convince them there really were people in the world who coded that way.
>
>2) It's horribly inefficient.   Multiplication is one of the slowest
>operations
>you can ask any computer to do (outside of I/O).
>
>3) Of all the systems in the world to use this technique on, the AS/400 is
>about the worst possible choice.  The error condition  (the overflow which is
>implicit in _every_ use of the technique) is bubbled up from the hardware
>through multiple levels of the OS until it eventually reaches the RPG code -
>where its promptly ignored!!
>
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to "MIDRANGE-L@midrange.com".
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---