Re: It's That Time Again! The 2002 RPG Enhancement Poll

["Simon Coulter" <shc@xxxxxxxxxxxxxxxxx>]

Hello Rob,

You wrote:
>Having IBM do the trivial sometimes comes up with some better idea's.

Not always :)

>For example, everyone had a set of date routines for dealing with
>numeric numbers.  However I like the functionality of true date fields
>much better.

This is not a good example.  Dates, times and timestamps were implemented
in the database therefore they become accessible to any data processor
(Query, SQL, HLL, etc.) therefore it makes sense to provide built-in
support for those data types.  (But look at how many systems **still**
use numeric 'dates' thus requiring the language to support conversions
when we should be using the native types.)

What I object to is the seemingly endless requests for Toronto to provide
trivial RPG IV specific functions whose sole purpose seems to be to
provide a BIF to replace an op-code so free-form RPG looks better by
avoiding the switches out of /free and in again.  Hence my comment that
the proposed %TESTD, %TESTT, %TESTZ functions are a waste of Toronto's
time and money because:

D testDate      PR              1N
D  aDate                       10
D  aFormat                     10

P testDate      B                       EXPORT
D testDate      PI              1N
D  theDate                     10
D  theFormat                   10

D validDate       S              1    INZ(*ON)

C                   SELECT
C                   WHEN      ( theFormat = '*YMD' )
C     *YMD          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*MDY' )
C     *MDY          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*DMY' )
C     *DMY          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*CYMD' )
C     *CYMD         TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*CMDY' )
C     *CMDY         TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*CDMY' )
C     *CDMY         TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*JUL' )
C     *JUL          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*ISO' )
C     *ISO          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*USA' )
C     *USA          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*EUR' )
C     *EUR          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = '*JIS' )
C     *JIS          TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = $LONGJUL )
C     *LONGJUL      TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   WHEN      ( theFormat = $JOBRUN )
C     *JOBRUN       TEST(DE)                theDate
C                   EVAL      validDate = NOT(%ERROR)
C                   OTHER
C                   EVAL      validDate = *OFF
C                   ENDSL

C                   RETURN    validDate

P testDate      E

took 15 minutes to code and verify with the reference manual and allows
me to code:

C               IF      testDate( birthDate )
or
C               IF      NOT(testDate( birthDate ))
or
C               EVAL    validDate = testDate( birthDate )

Of course, if variables were allowed in Factor-1 for the TEST(x)
functions it would be simpler still.  Wack that function in a service
program, shove that in a binding directory, name the binding directory on
a common H-spec that is /COPYed into the program template, and we have an
instant new language extension.

This example needs testing and possible additional support for *YYMD,
*MDYY, *DMYY, etc.  Consider it open source; if someone creates the
corresponding time and timestamp (really easy) functions we'll be done
and Toronto can concentrate on complex things.  The same thing could have
been done with the date duration BIFs and there are other examples such
as ALLOC.

Having said that, perhaps Toronto do have something grander in mind but
that doesn't change the essence of my proposal which is to improve the
LANGUAGE rather than the code.  If the language improves the code will
too.  Toronto have given us the ability to extend the language via
procedures so let us stop asking them to do the stuff we can do for
ourselves.

Regards,
Simon Coulter.
--------------------------------------------------------------------
   FlyByNight Software         AS/400 Technical Specialists
   http://www.flybynight.com.au/

   Phone: +61 3 9419 0175   Mobile: +61 0411 091 400        /"\
   Fax:   +61 3 9419 0175   mailto: shc@flybynight.com.au   \ /
                                                             X
                 ASCII Ribbon campaign against HTML E-Mail  / \
--------------------------------------------------------------------