Re: framework question - OO is useless -- JAVA400-L

Joe Pluta wrote:


Actually, if you want any credibility with me, you will have to
establish your credentials.  I'm comfortable explaining my background
and why that gives me some justification for my opinions.  So far,
you've yet to do so.  You're just sort of telling me "OO is best".  Even
the OO guys don't do that.  And in any event, you're the one who started
the personal crap with the "quick hack" comment.

I dont stated that you do "quick hacks". I stated that "quick hacking" aka fast coding can be better done with RPG. Whats personal in that statement? Nothing. So calm down.

I wrote about 10 Articles on MidrangeServer.com focusing on Java on iSeries. This backed up with about 11 years experience in software development, 6 years of that in the iSeries space. You can read everything at www.logemann.org if you like. But i dont think its useful to compare BIOs here, its about the topic.

The third party package deal is beaten to death.  Yes, you can find some
third party libraries.  And those are almost all for things which have
nothing to do with application business programming.  For those things,

If you completely focus on business development, aka boring data i/o and calculations (btw i know ERP systems and i know how complex price calculations can be) RPG is indeed not too much behind or even if you like. But i still cant see why you cant do it in java. If one doesnt know how OO works, he still can code procedural in java. If you think the performance is the point, ask IBM why java on the iSeries is so damn slow compared to the machine price.

But we should define "your" scope of application business programming.

Also, I've seen more than one project get badly burned by
incompatibilities between third party libraries.  And then what do you
do?

Simply know what 3rd libraries to use. Just be skilled. Its really no magic to use 99% percent of them and 98% of them can simple be mixed without a problem.

No, it seems you don't know what pricing means.  You can't follow me,
because you've never written an enterprise-level pricing module.  I find
this all the time: Java theorists use simplistic application "examples"
to prove their point, but in truth the real world model doesn't fit.

Why does SAP still exists? They (and thousands of other developers worldwide) seem to master business development with java.

For example, in pricing you often have to know how much a client has
bought in a certain period.  This is historical information.  You may
even have to know which store he bought from, in order to provide
quantity discounts.  In an OO environment, if you haven't already made
that information available to the class that calculates price, you will
have to modify the interface.  For example, if you only provided the
item and customer number to the pricing class, then you will need to add
the store number.  And as has been shown in every study of OO
programming, changes to the interface are the single most expensive
thing in OO development.  Whereas in a procedural program, it's simply a
call to another server.

Two ways. First you can just provide another method with a different signature, this wont break any interface. But a better approach is to use a PriceCalculation object on the signature. This object can be modifed in any way without breaking something. I hope this wasnt your ultimate prove that you cant do a price calculation with java.

But more important, i dont follow your procedural programming statement, if your price function only has two paramters, how to add the third parameter without breaking the (non-existant) interface? What do you mean with a call to another server?

Excuse me?  How exactly do you redeploy an EAR file without restarting
the application server?  Please show me the EXACT procedures to redeploy
an EAR file.  I'd love to see this, actually.  (By the way, I know you
can get around some of this by hot-applying changes, but then you need
to know the directory structure of your server, and that gets pretty
scary.)

Its not scary, its only scary when you dont have a clue what an EAR file is and how an AppServer deals with it. This is AppServer Administration Fundamental stuff. If you just use the Browser-Console of the AppServer to maintain it, then i would suggest never doing anything critical with J2EE.

Creating a project with a RPG/JSP mix is the last option i would use
when creating a web-based project.

And it's the first one I would use.

You see, many oppinions which is good. I dont blame you for your approach, but let others do theirs.

Maybe because RPG and COBOL are doing just fine in the business rules
space.  Most of the concentration has gone on in the Internet
programming arena, and frankly we know exactly what happened there,
don't we?  Basically a highly inflated market of software tat was never
delivered, leading to a meltdown of the entire tech economy.

Software that never delivered? I would think most serious internet sites, shops and portals are running with Java (or Perl or PHP) these days. There was a meltdown, but not because of the languages, but because of crazy venture capitalists. Cobol is doing fine, because most companies dont have the power to reprogram that, so they have all kind of Corba Middleware to let this old-stuff communicate with a modern software backbone. I know a lot of german banks who hate having still cobol programs running.

In the meantime, of course, RPG and COBOL keep on going and going...

Yeah, but mostly not because the people want it, but its hard to do a transition.

Marc, you really need to revisit RPG.  The changes to RPG have been
incredible.  As have the ILE concepts and everything else.  You're just

Perhaps a statement i can agree with you. I left RPG where ILE just came into the picture. So i missed all that freeform thing and perhaps some more advanced ILE techniques, but at the end its RPG isnt it? Just with some sugar on top, to be able to modularize better and to get rid of these column based programming which was allready obsolete 8 years ago.

Marc, you can spend some time heeding your own words.  You called my
programming a "quick hack", simply because it doesn't fit into your OO
methodology.


This isnt true. See above.

In any event, you have yet to provide any reason that Java is good for
business applications.  You rant and rave about third party packages and
namespaces, but frankly I don't find any real business experience in
your statements.


Try this article, perhaps you will find some more.

And I'm not flaming you, I just don't like people coming onto the
mailing lists and telling everyone that OO is the way to go when they
really don't have the experience to back it up.  There might be someone
who doesn't know any better who might listen to you.  So I am here to
make sure the other side of the argument is heard, and to allow the
readers to make up their own mind.

I am propagating OO on iSeries since a long time, so i am not someone who is completely new to this. We spoke on the same iSeries Conference last year if i can recall it correctly and i am constantly doing this through my articles. You can have a different oppinion, but dont think that you have the key to all iSeries problems. There are some others out there also knowing some details about programming and design.

Remember, I love Java.  I love OO.  I think OO may indeed be one of the
few true advances in programming since the Von Neumann machine.  But I
also think RPG is a better language for developing complex business
rules, and I have real world examples that, at least to me, prove why.
This is based not on a dislike of the language, but on over 25 years of
development experience.

I know your development experience, but to me statements like "the world has no time for OO" sounds like an headliner for an RPG usergroup. Nothing more.

regards
marc