System iNetwork

To bad this was not available a couple of years ago before my company made a decision to go a completely different direction and now will not even talk about the new opportunities.

I too am being forced out of a System i job after a merger/aquisition. The New Company is moving us over to a Microsoft-based product....

I worry IBM is going to forget us now that IBM i is virtualized on the POWER platform. Even if investment in the core OS and ILE is cut to a minimum, IBM i will continue to sell and run on POWER hardware into the future.

What are the features planned for the next release of IBM i? How many programmers are working on it? When is the expected release date?

The Microsoft model is a good one to follow in this regard. Ian Jarman should be blogging about what is in store for IBM i. Release preview versions of the new OS to the developer and cutting-edge user community.

-Steve

There's a cutting edge to the System i?

The feature the System i needs most is for IBM to deprecate and end-of-life many of the features they support.

IBM makes the same mistake as old school programmers, thinking that "sunk cost" equals "investment". The code on most AS400s ("AS400" is easier to type and how do you pluralize or make possesive System i?) is a ball and chain that prohibits change and growth.

What new feature does i5/OS and the attendent infrastucture need? It needs to eliminate old features. If companies are going to run code from the 1980's, they don't need an upgrade path to the next release of the OS.

BTW, the Dallas Techfest was today. There were presentations on Ruby, Java, .Net, Flex and Air along with free admisssion and free food. iPods and Xboxes were handed out as prizes at the end of the day.

Greg - I wish to refute your claim that "If companies are going to run code from the 1980's, they don't need an upgrade path to the next release of the OS." My company runs software from the 80's, 90's, 00's, and my customers still require new functionality. Shops like mine need the new functionality and the old!

I may be wrong, however I believe the decision to move to another platform (and OS) has more to do with software selection vs. finding software that runs on their hardware of choice.

Greg

In the real world, every company has his old features running along with the latest features available.

And don't worry about the old features: they all are part of a system with the lowest TCO in all IT. If you don't understand this, you don't know System i.

You can be sure that all your brand new features will be legacy in a few years, and soon you'll be praying IBM doesn't drop your old features.

A. No one pointed out any cutting edge features.

B. An older Java app, for instance, one that uses Struts 1.2 doesn't need support from anyone (vendor or open source community). In the open source community, new versions of frameworks aren't backwards compatible with previous versions. That's how the open source community makes progress. Daddy takes the old dog on a ride out to the country, to a farm where he can play with other old dogs.

RPG should lose its backwards compatibility with RPG 3 style of coding. IBM should remove support for subroutines, C-specs and F-specs from the RPG language. None of the dozens of RPG programmers I know attend user group meetings or read about new features in the language. What else could make them learn the new (ten-year-old) features in RPG? Do you want to work on twenty thousand line programs subdivided into just a few subroutines for the rest of your career?

I think my argument has merit. Most of the RPG written this day (5/2/08) was in a program main or a subroutine. RPG 3 was great for machines that had 100 kilobytes of memory and half a GB of storage. Today, in order to change one piece of logic, companies have to change duplicate code in dozens or hundreds of RPG programs. Java, on the other hand, has its domain model, services and the concept of DRY (look it up) that all guide developers toward code reuse so a change only happens in one place. Java will be cheaper to maintain until RPG programmers "get it". ILE RPG can compete with Java on the point of reuse.

I know you guys think I'm wrong. I had more than a dozen years of RPG experience before I started working with a group of Java architects for five years. I try to apply what I learn from them about program design not just to Java but to RPG, too.

Reuse code. Copy and paste is not code reuse, it is duplication. In RPG, reuse is simply calling a procedure.

Write code that can be reused, i.e., called from other programs. Write ILE subprocedures.

Write code that can be tested. Write tests, using Don Denoncourt's test framework. Run your tests.

In my download the Spring framework I find 7.74 MB of test code for 13 MB of source code. Apparently those Java guys are pretty serious about quality. You can only have automated tests for your code if you can reuse that code in a test framework. So write reusable code. Write subprocedures. If you code subroutines, you are hurting your employer and yourself.

I know you guys think I'm wrong. I have heard it all before. I'm a Java developer now. I'd be happy to code RPG (I logged on an AS400 green screen over the weekend and did some coding) but I wouldn't do it in a way that is empirically wrong.

I work for county government. With budget restrictions, we have to make old code (from the 80s/90s) work. Our philosophy is if it isn't broken, don't fix it. A lot of the old code still works and does what the user needs. However, we do need to keep support current so the new OS needs to support the old code as well.