RE: ILE philosophy

["M. Lazarus" <mlazarus@xxxxxxxx>]

Scott,

> What are the steps you refer to.  I just don't see that it is that diffucult
> to manage.  What am I missing that you are referring to.
> I don't miss a MAKE tool that much.  I can only think that it would be real
> useful if you bind modules by copy.  Since I don't advocate that approach I
> don't see much need for a MAKE tool.
1)  All compiler options s/b able to be put into a source.  While the H
spec allows most, it does not allow all.

2)  Both binding by copy and reference require a 2 step process, if you've
changed the module.  Recreating the module should either create / update
the service program or mainline program.

3)  In cases where there's a circular reference, i.e. module A refers to
module B and module B refers to module A, one of the modules must be
created w/ *NORSLV (sp?), then the next one would be created, then the
first one must be recreated.

4)  Now that the module info is buried one additional level deep, and more
importantly, the fact that the system knows (to some degree) the
relationship between objects (due to binding info being known at compile
time), there s/b some hierarchical display to show that relationship.

 The list goes on.  A lot of the info is available, but it's difficult to
get the big picture, since it's buried several levels deep.  This is basic
system management info (what IBM s/b providing), not application software
(what vendors s/b supplying).  JMHO.

 -mark