[TriLUG] Money in mysql

Chip aviatrexx at gmail.com
Thu Oct 14 19:27:52 EDT 2004


On Thu, 14 Oct 2004 08:05:50 -0400, Phillip Rhodes
<mindcrime at cpphacker.co.uk> wrote:
> 
> len boyle wrote:
> 
> >
> > Mike  is known in the old world as the author of Rexx and NetRexx  See
> > his web page listed on the jcp page for many refs beyond java. Matt you
> > will like this site.
> >
> > http://www2.hursley.ibm.com/decimal/
> 
> Interesting... what are the odds of two references to Rexx coming up
> on the Tri-LUG list, on the same day, in pretty much completely
> un-related contexts?
> 
> It's a weird world we live in... :-)

Weirder than you suspect, I'll bet.  And smaller.   :-)

As it turns out, the President of the Rexx Language Association is a
proud member of TriLUG, lives in SW Wake County, and would be at the
meeting tonight were it not for the crush of work that has been
created by IBM's transfer of Object REXX to the Open Source community.

And no, Mike Cowlishaw is not retiring, although he has the honor of
being not only a long-time IBMer, but one of the longest-tenured IBM
Fellows (no small feat when you have to come up with something
brilliant every 2-3 years to justify  your position).  For instance,
Mike was the guy who convinced Lou Gerstner to invest $1G in Java in
the first place.

Mike wrote the first IBM port of Sun's Java to an IBM platform (OS/2
of course) and then (since he didn't have anything to do over the
Christmas holiday) he designed and implemented NetRexx, an
object-oriented Rexx for the JVM.  The first version of NetRexx was
written in Rexx, the second in NetRexx.  (So much for the "Rexx is
just an old mainframe scripting language" drivel.)

To keep this OT, the Java Decimal class was a "sorta-decimal"
implementation, so Mike wrote a true decimal arithmetic class
(BigDecimal) for Java which is now part of the base distribution. 
BigDecimal is simply an implementation of the arithmetic defined by
the ANSI X3J18-1996 Standard for the Rexx programming language which
offers infinitely variable precision and avoids all the "decimal
anomalies" previously mentioned.

IBM thought  this was such a benefit that for the last couple of years
Mike has been working on implementing Rexx decimal arithmetic on the
chip. (Watch for it to come to an IBM processor near you!)

Which allows me to respond to the naif who opined that "I don't think
BCD is found much these days".  Jeremy does not appear to be aware
that there are _billions_ of lines of S/360/370/390/zSeries code still
in use out there.  An great deal of it is written in  IBM Assembler, 
which is one of the languages (with Rexx, BTW) that I make a living
teaching.  I cover BCD (packed arithmetic, actually) on the second
day.

And to follow-up on another post, yes there is a movement afoot to get
IBM to open-source OS/2, but it's very complicated and delicate, and
far from a sure thing (having a lot to do with certain pieces of
Microsoft IP, as you might suspect).

BTW, the Linux version of Object REXX was written by another  IBMer
who now lives in Apex.

If any of you might be interested in contributing to the Open-source
Object REXX (OORexx) effort, please feel free to contact me at
president at rexxla.org and I'll put you int touch with the development
team.

May the PARSE be with you,

-Chip Davis-
President, RexxLA
www.rexxla.org



More information about the TriLUG mailing list