[comp.lang.rexx] Some Beginner's Questions/Impressions on REXX.

hhg1@GTE.COM (Hallett German) (07/05/89)

I am impressed with the REXX language but have some questions which
are probably beginnerish but since this group (Is there a BITNET version
may sound beginnerish (Is there a BITNET version of this group? Any
more traffice there).

1. It seems that some of the beginner's Traps for using REXX are:
   a. confusing function call with subprocedure (using CALL)
   b. Thinking DELSTR and similar built-in functions have a default
      length of 1 character rather than end of string.
   c. Trying to use SIGNAL as a GOTO for all situations
   d. Thinking internal subroutines variables are local. 

Did I miss any?

2. What type of REXX tips not in Colishaw book which would be useful to
   know? (I don't have a copy of the other REXX book by Ohara I think.)

3. Is there plans for a later version of REXX than 3.50 (that is the
   latest version I have) What new features are planned?

4. THe I/O model wasn't adopted for CMS and I think TSO. Are there any
   efforts afoot to add the CMS extensions to the REXX language (such
   as MAKEBUF,DELBUF,EXECIO,SENTRIES). This appears to be the de facto
   standard.

I am sorry if these are real basic questions. But I haven't seen the 
answer anywhere and am the only person using REXX at my site.

Hal German
GTE Labs 

C
C
C
C
C
 
    last version I have. What i

terry@uts.amdahl.com (Lewis T. Flynn) (07/06/89)

In article <7154@bunny.GTE.COM> hhg1@GTE.COM (Hallett German) writes:
>1. It seems that some of the beginner's Traps for using REXX are:
>   a. confusing function call with subprocedure (using CALL)
>   b. Thinking DELSTR and similar built-in functions have a default
>      length of 1 character rather than end of string.
>   c. Trying to use SIGNAL as a GOTO for all situations
>   d. Thinking internal subroutines variables are local. 

They are if you use the procedure statement, not if you don't.

>Did I miss any?

Probably, but the only thing that comes to mind is that after a while,
you become a fanatic about quoting everything that isn't a key word
or a variable. REXX is one of the kindlier, more forgiving languages
I've come across. Very much unlike PL/I, one of its forbears.

>2. What type of REXX tips not in Colishaw book which would be useful to
>   know? (I don't have a copy of the other REXX book by Ohara I think.)

Be careful with your quotes, as mentioned above.

>3. Is there plans for a later version of REXX than 3.50 (that is the
>   latest version I have) What new features are planned?

Which (whose) REXX are you running? CMS versions are part of the system
(SP6, I think is the latest), TSO is part of TSO-E Version 3 (2?), and
the current version of Personal REXX is 2.0 (2.0d or some such).

>4. THe I/O model wasn't adopted for CMS and I think TSO. Are there any
>   efforts afoot to add the CMS extensions to the REXX language (such
>   as MAKEBUF,DELBUF,EXECIO,SENTRIES). This appears to be the de facto
>   standard.

The ones you mention are all CMS commands which you can use like any
other CMS command. I would speculate that the implementers of the
CMS version didn't feel the need to install the I/O parts when the
functionality was already there (for use by the other two command
procedural languages already in CMS). A minor note, the queued()
REXX function serves the same purpose as SENTRIES but in a more
convenient fashion.

As far as additions go, all the common CMS commands used in REXX
execs are part of Personal REXX for compatibility reasons (so you
can port your CMS execs to MS-DOS). I never use them, though as the
standard I/O functions are easier to use.

>I am sorry if these are real basic questions. But I haven't seen the 
>answer anywhere and am the only person using REXX at my site.

Not to worry. That's what these groups are for. And besides, this group
can use a little volume.

>Hal German
>GTE Labs 

Terry

Standard disclaimers apply.