RE: Of chalk and cheese... -- MIDRANGE-L


On Wed, 25 Nov 1998, David Murphy wrote:

> 
> We're a COBOL shop (not using ILE) and I'm curious why you say that? Did
> you have problems? Should we not look into ILE COBOL? 
> 
> David Murphy

David,

no, ILE COBOL is fine. With ILE COBOL you will get a few neat things.
Reason for my comment is that level of improvement made to COBOL isn't
nearly as great as it is in case of RPG.

        Some of the new stuff in ILE COBOL (just on top of my head):

- Linkage type "procedure" (naturally, we are talking ILE)  

- External data definition (don't confuse with externally described data)
realized through "weak exports" (a few modules sharing same data space,
open file in one module, read in another, update in third ...) 

- Subprograms (more that one program in a single member, with defined
hierarchy, help localize data, structure programming etc...) 

- Better compiler, faster programs, higher limits ...  
- ACCEPT 4 digit year (from certain release of OS)  
- Built in functions, some new ways to CANCEL program ...  

But, I just have a feeling (it's just me, no flames, please) that ILE RPG
developers took opportunity to really shake RPG. It looks to me that they
addressed real weakneses of RPG and made such changes that it looks they
were almost following "wish list" of RPG programmers all over the world. 
They also choosed the right moment, all this Y2K mess, to introduce date
type and wery good set of tools to manipulate it. Programming in RPGIV
simply feels good. 

On the another side, ILE COBOL developers didn't go all that way. Date
manipulation is weak (better than in "regular" COBOL, but not good
enough). Hey, again, it's just me, no flames, please. It seems that
activation groups concept doesn't play that well with cobol run unit
concept. There is "strange" behaviour of GOBACK statement (it can close
files under certain circumstances even if it is not first COBOL program). 
They tried to correct it with "without closing AG" (or something similar) 
sentence, but... Also, START statement (to I-O opened file) locks record
and won't work with already locked record (behaves like READ and gives FS
"9D"; that can be suprrising for old COBOL guys, and for old COBOL
programs which you might wish to translate into ILE). Above mentioned
subprograms are nice but not right match for prototypes and subprocedures.
And last, but not least, I can't accept that incidentaly "walking" out of
the boundaries of array or table (i.e. index is set to value higher than
that in OCCURS clause)  can corrupt data in other modules within same
program. I would expect modern compiler to protect me from that. 


Again, there is nothing wrong with ILE COBOL, if you have experienced
people in house that's way to go. You would benefit from ILE environment
"big time". Only reason for mu comment was disappointment that I like ILE
RPG more, and COBOL was my first AS/400 language.

Take care

Vanya Jovic



> -----Original Message-----
> From: owner-midrange-l@midrange.com  On Behalf Of Vanya Jovic
> Sent: Tuesday, November 24, 1998 10:56 PM
> To:   Midrange-L
> Subject:      RE: Of chalk and cheese...
> 
> 
> 
> Hi, 
> I'll go one step further. After 1.5 yrs working with RPGIV, I would never
> write new program in RPG/400 (a.k.a. RPG III). Also, don't think I would
> even remotely consider accepting position with company which doesn't use RPGIV
> and other ILE features and benefits. 
> 
> Unfortunately, can't say the same for COBOL/ILE COBOL relationship.
> 
> Regards
> 
> Vanya Jovic
> 
> On Tue, 24 Nov 1998, John Carr wrote:
> 
> > 

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@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
+---