RE: Strange behavior w/%editc()

[boldt@xxxxxxxxxx]

Well the latest digest rolled in while I was writing this,
and so now some of my points have been rendered redundant.
But what the heck, I'll post my comments anyways.

Joe wrote:
>I thought I'd get a response from this.  To be honest, if I had to choose,
>I'd rather my business languages were designed by people with business
>application experience than someone with no business experience.  I'm from
>SSA.  We had a language developed by compiler experts.  It was called
>AS/SET.  While it was great for generating code, it was useless for
>developing business applications.

I agree we need both domain experts and compiler experts,
and I believe we have both.  For example, the free-form
stuff was reviewed by a lot of people, both inside and
outside the lab, who had a lot of influence in the design.
Our first design included support of all opcodes.  But one
of our internal reviewers, who had much more contact with
our customers and more knowledge of the needs of application
programming, insisted that we should not support all the
older opcodes.

Furthermore, we do have that regular "RPG Enhancement Poll",
in which we ask RPG programmers to rate proposed
enhancements.

But ultimately, it is our responsibility to manage the
language.  Maybe some particular people aren't happy with
everything we do, but tough!  We can't please everybody.  I
think most people understand that.

>I suspect you know very little of what I know or don't know about Java,
>Hans.  Perhaps you've read some of my articles, or attended some of my
>classes on the subject.  Since I have little knowledge of your Java
>capabilities, I can't comment on them, but I have a commercial Java-based
>business application on the market.  That's more than most Java
programmers
>can say.

I've written a few lines of Java in my time too, but then,
I don't want to get into any kind of pissing match.  My
point is that it's ludicrous to think that there's the
remotest chance that RPG will turn into Java.  Since you've
had some experience with both Java and RPG, you of all
people should know that.

>But in any event, it's not my Java skills being discussed here.  It's my
>opinion that removing the MOVE opcode is a misguided concept.  I've yet to
>hear one good reason for it's loss.  On the other hand, I can say that
>replacing the MOVE opcode with some C-syntax BIFs is definitely a
backwards
>direction, and I have a feeling many other programmers would agree.  RPG
>isn't about "cool and geeky", it's about getting a job done in a limited
>time frame with a limited budget.  RPG is the best language in the world
for
>that, and a large part of that capability is the lowly MOVE instruction.
>But hey, I've been wrong before.  And I could live with it if one of the
>decision makers could actually tell us lowly programmers why they did it.

Well, I first wasn't going to comment since I've already
written more than I care to on the subject of free-form
calcs.  It boggles my mind to think that this is THE
most controversial issue in RPG in the past 20 years,
when practically all other programming languages have
been totally free form for the past 40 years!

I think Barbara put it best yesterday when we were
discussing your comments.  What exactly does MOVE A B do?
What happens with arguments of different types?  Or of
different lengths?  Face it - MOVE is a rather overloaded
monstrosity of an opcode.

By excluding MOVE from /FREE, we haven't removed any
functionality.  Practically anything you could do with
EVALs and BIFs.  That was the main reason MOVE was left
out of /FREE - we just felt that we shouldn't have too
many different ways of doing the same thing.  We left
out MOVE for the same reason we left out IFEQ and DOWNE.

Regarding your statement about RPG being the best
language in the world for getting things done in
limited time and limited budget, I would suspect that a
lot of programmers around the world would argue
differently.  Perl programmers argue that Perl is the
best, Java programmers argue that Java is the best, C++
programmers argue that C++ is the best.  (On the other
hand, Python programmers KNOW Python is the best ;-) )
I've made this point before, and it looks like I have to
say it again: Whenever someone says that X is the best,
they are really saying that X is what they know best.

Is RPG the best for business applications?  Well, I've
seen successful business apps implemented in many other
languages as well.  I would argue that implementation
language is one of the lesser factors in the success of
any project.

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com