Re: Using RDP 8.0 i Projects

wdsci-l-bounces@xxxxxxxxxxxx wrote on 08/01/2011 12:28:39 PM:

----- Message from "Mark Murphy/STAR BASE Consulting Inc."
<mmurphy@xxxxxxxxxxxxxxx> on Mon, 1 Aug 2011 10:37:18 -0400 -----

To:

wdsci-l@xxxxxxxxxxxx

Subject:

Re: [WDSCI-L] Using RDP 8.0 i Projects

Yes, that is true, but with other environments like .Net, Java, or
PHP, there is a directory structure that helps you out. Not true on
i. Just a flat structure, with potentially multiple source files if
one is inclined to break things up a little bit. However, mostly 1
source file is used per language rather than per "application area".
So in other environments I can navigate to
/app/customer/controllers/ and I see a handful of source files.
Easy to find stuff, I don't even really need to spend a lot of time
remembering the name of the object. With i, I have to sort through
the 1500 source members in QRPGLESRC, or maybe it is 250 source
members in CUSTSRC. Then I have to find the right source file, or
wait while the filters find all the objects. Yes, I know what I am
looking for, but I prefer just having the 5 or 6 objects I am
working with there for me.

Mark Murphy
STAR BASE Consulting, Inc.
mmurphy@xxxxxxxxxxxxxxx

From what I've seen, .Net and others don't force you to use a directory

structure. You can put it all in one directory or break it up into
multiple directories if one is inclined to do so. Just because your
company (and many others) blindly throw everything in the bid drawer of
the file cabinet doesn't mean the file cabinet can't handle hanging
folders with manilla/file folders inside those. I've always advocated
multiple libraries (with meaningful library names) and meaningful source
names within those. I will admit you are limited to libraries and source
files within that, but *no one* forcing you to put all your RPG in
QRPGLESRC in one library.

I also understand why some shops have chosen the single source file in a
single library model. It depends on the needs. If all your applications
are interrelated, it simplifies finding something. If you don't have good
naming conventions and require meaningful/useful source member
descriptions, finding things can be tough. But if you have either of the
above, you can find anything pretty quickly and easily. (And just to
establish that I'm not dealing with a smaller set of source members, I
just checked--our main COBOL (gasp!) source file has 1587 members.) It
ain't religion let alone rocket science--it's just programming.