iSpeak

Hear from our iSeries experts. Put in your two cents.

September 16, 2005

The Perception

There's a major schism evident in this blog between those that believe that RPG is a language for traditional programming and those that see it as a capable language for the future. The former perceives RPG as a language for 5250 applications, text-only reports, and batch jobs. It's not hard to understand where they get this perception, since that's 99% of what you see when you look at RPG programs. The other side thinks it's capable of much more, and I fall into that category.

If you read the comments posted to this blog, both in response to Paul Conte's Free Format D-specs are LONG Overdue! and my own It's Time to Improve Our Image, you'll see that the vast majority of arguments boil down to this difference of perception.

From my perspective, RPG and Java are direct alternatives to one another. The things that you can do in Java or .NET or Visual languages can also be done in RPG.

Many people, such as myself, have written web applications in RPG, and have found it to be an elegant language for doing so. I use RPG to generate not only plain text printed reports, but fancy, colorful reports with pictures and lines and different fonts on them. I use RPG to generate Word and Excel documents that can be built off-hours and e-mailed automatically to our sales force, our customers, our vendors, etc. I use RPG programs to consume web services, to provide security for network applications, and much more.

From the perspective of many others, that's not the case. RPG to them is for writing green screen accounting software and drab reports that all use a common fixed-width font and are often printed on green bar or blue bar paper.

It's important to understand that this perception is absolutely central to the problems facing the iSeries today. If you believe that RPG is only for old-style programming, then RPG is doomed. If RPG is doomed, I have no doubt that the iSeries is also doomed. Like it or not, the two are inextricably linked. You can run Java or .NET anywhere. Without RPG, the iSeries is just a really expensive database server.

It's RPG that makes it all make sense! It's an elegant language for writing not only business rules, but also for web programming. I've written web applications in PHP, Perl, C, and even Java. CGIDEV2 -- an RPG method -- is the most elegant way that I've found! Most of these other methods either involve embedding the HTML/XML code in the program or embedding the program in the HTML/XML code. In either case, it's awkward. Separating the markup language from the code, the way CGIDEV2 does it, is much more elegant in my opinion.

The problem isn't that RPG isn't capable of these things. The problem is that the programmers who make up the majority of RPG people are not aware of the capabilities that it has. Or, perhaps they haven't upgraded their skills to the point where they're able to understand modern techniques.

That's not usually a problem with a Java or .NET programmer.

What this all boils down to in the end is how someone perceives of the iSeries and of RPG. A management team that's out looking for a software solution looks at the alternatives, and sees demos of the available software packages. Or, if they're looking at custom software, they look at what consultants have done for other companies. Or, for in-house development, they look at what other companies have done to provide a starting point for their in-house staff.

About 70% of the solutions that they see are Windows solutions, because it dominates the marketplace. If they keep looking, they'll eventually find the iSeries solutions.

And when they see iSeries solutions, they occasionally see a really unbelivably convoluted WebSphere solution that costs big bucks and requires big iron. The rest of the time, they see an ancient 5250 system that looks pathetic when compared to the friendliness and flexibility of a Windows system.

When 70% of what they see is Windows, and those solutions are friendlier and more flexible, will they bother to look at iSeries solutions the next time? Especially when they're much more familiar with Windows?

It's true that it'll take a long, long time for all of the existing iSeries shops to migrate away. So the system won't be dead soon, but it will fade away very slowly!

If we want to change this, folks, tt's the programmers who need to step up to this challenge. We need to upgrade our skills so that we can compete!

Now, this is the point where you choose RPG or Java. Do you re-use the skills you already have, and expand upon them, to write efficient, modern RPG code? It'll run faster than the Java code, use much less memory, and RPG's syntax is the best in the world for writing business rules. Or do you bite the bullet and do your work in Java? There are many more Java programmers in the world, since it's not an iSeries-only technology, and your code will run on any system, which is a big advantage to some people.

Believe me, remaining with the 5250 technology, or drab reports, or nightly batch type coding isn't an option! Not if you want to have a future.

Fortunately, no matter how you approach it, the iSeries is capable of doing it. And it'll be stable. And it'll be easy to troubleshoot problems on. But it may not be around forever if we don't upgrade our skills and improve our image. Because whether it's technically better or not, it's all about perception!

Posted by at September 16, 2005 3:04 PM

Comments

Hello Scott, I must say there I do agree with virtually every point in your post. As far as I can tell, you do have the correct perspective on Iseries future and as you say, RPG "the ecosystem as IBM would call it", is key.

And yes, ofcourse RPG can handle GUI screens and reports perfectly. And it can do it even without a web server application. Back in 2000, I played a bit with socket programming, and I managed to communicate with a web browser on a PC directly from within my program. The HTML pages were stored in a variable length indexed file, and the index was in fact comparable with a screen name. The function to read or write was simply PutScreen(screenname) or ReadScreen(screenname), again comparable with the traditional rpg exfmt. The first reaction when I showed a demo was WOW!, does this come from an AS/400?

But, but, but, I am however afraid for this systems future. Certainly in Europe, the system is losing CRITICAL MASS fast, and the Mass loss is so heavy that at least in my view, sorry to say, the system is beyond recovery. The Iseries won't die, but the tradtional eco system itself will. Iseries will become one-out-of-dozen-servers running linux, cheap web server, cheap data base server apps, maybe the only value that remains is SAP (yes, SAP runs on OS/400) and Domino.

Posted by: U.Geerts at September 17, 2005 6:19 AM

Hi,
The differentiator I believe is the extent of help when you get for writing a code. Java has excellent "free" documentation, I believe so is the case of .net. Can we say the same about RPG? Even the ones we have is cryptic. I am yet to meet a person who has not made the error of using 4B 0 fields when using API's.

Everybody wants to improve the skills, but few are willing to take the risks associated with it. The heart says go for it, the brain says go for the tried and tested method. In most situations toleration for untried method is less and errors even lesser.

Posted by: Sunil at September 17, 2005 6:55 AM

Your post is a perfect example of the restistance to change in an Iseries shop, although your reasons are understandable. The responibility for anyone to take in case something goes wrong, if you change or modernize core applications, is often too high.
But then again, we've arrived back to square one.

Posted by: U.Geerts at September 17, 2005 1:23 PM

Scott, great post. Modernization of applications is going to happen, it's just a matter of time. It's time to teach some of us old iSeries dogs some new tricks.

In my opinion, we have to grasp new technologies and implement them on the iSeries and make the iSeries an important part of that modernization.

Only developers can do that, but as UGeerts says, there is alot of resistance to change in most iSeries shops. If that doesn't change, the platform will not be a viable solution to anyone!

Posted by: Real Developer at September 18, 2005 7:17 PM

I'm one of the ones that based on my responses would seem to fall into the RPG is for 5250 camp, even though I wrote the web server for the apparently defunct JOBS/400 web site in ILE RPG. On the other hand it was interface independent code and I tested it with 5250 screens, but that's not common out there as we all know, nor is it necessarily the right answer for most of our AS/400 programs. It has its place. So do 5250 programs.

Our traditional strength has been the complex 5250 subfile programs operating against an integrated OS/400 operating system and DB2 database. There is no future for us as a me too Websphere database server. Websphere is common denominator, and OS/400 doesn't bring that much to bear for companies to require OS/400 expertise to run a lowest common denominator Java application server.

As a result, IBM is now selling the AS/400 as something to lower costs by sharing hardware across multiple operating systems and implying they will lower server admin personnel costs as part of that from what I can tell, but they are so vague one can never tell just what they are up to. The implication to me from their marketing for the AS/400 is that as you shift to other OS's such as Linux, AIX, and Windows you can repartition OS/400 smaller so that your box doesn't become obsolete with OS/400. They would never openly say that, but that's the implication.

I think that RPG and even Java with the OS/400 Toolkit extension is a powerful business programming centric environment that's in between the thinness of web page programming and the thickness of Windows client server programming. 5250 terminal programming is the right mixture of interface complexity and centralized server code. It's different, and it has its place in the world of solutions.

I argue that we fill that niche by continuing to provide powerful, cost effective solutions driven by our RPG programs but that the EXFMT buffer be opened up by IBM so that the output is 5250 binary to terminal emulators but XML with 5250 keywords and DDS extensions to be displayed by any interface, from web pages to Java GUI to Unix forms to Windows rich clients. It's an XML stream. For that matter, it could be machine to machine at that point. It completely opens OS/400 up and lets us continue to provide the powerful interactive programs we have provided all these years. It allows our programs to be interface independent even though we're fundamentally programming to the 5250 functionality, but with DDS extension keywords we could cause richer interaction to optionally take place.

There's a lot more to this. I believe interaction among apps on the client interface is what is most desired by users and web pages and Java GUI are islands insulated from that interaction. That's why Microsoft will win if IBM does not give us the ability to stream our complex program screens to any interface via XML from the EXFMT buffer.

I know that is politically incorrect to say that in these days of the requisite separation of interface, logic, and data access, but we write more complex programs than fill a web page with an SQL SELECT statement. The key is all that RPG draws upon from OS/400 delivered to any interface anybody has, not just web pages.

Our programming needn't change, we just need our interfaces opened up in OS/400 to deliver our powerful solutions anywhere, 5250 or web pages or better yet an open source Eclipse based richer client that is an OS/400 interface that will compete with other OS'es.

rd

Posted by: Ralph Daugherty at September 18, 2005 7:38 PM

I agree with Scott for almost everything. However, real RPG-ILE programming on the server in combination with JAVA on the client (both rich JAVA clients and browser clients) allows modern, object-oriented programming for the creation of components (Services) that comply to the SOA specifications. I have this reported this many times and with our JAVA Workflow Design framework (JAWFLOW) we have created an application development platform omitting the complexity of WEBSPHERE (EJB, containers, etc) but including all the beauties in i5/OS. We do better then Microsoft with NAVISION, etc. The best of this story is that we achieve subseconds response times and that is what iseries users want.

Posted by: Ludo Dierckx at September 19, 2005 7:17 AM

I view the dichotomy of expressed opinion less along the lines of RPG's purpose and more along the lines [warning -- red flag ahead] of a split bewteen those who understand the benefits of contemporary language features and designs and want to bring RPG up to date, and those who see only through the lens of their habituation to RPG and are uncomfortable with significant change.

The current state of RPG remains far too cumbersome and lacking in function to compete as a leading (or even acceptable) language for contemporary programming practices. Java and C# are certainly compelling comparative reference points to demonstrate this fact.

Any vendor trying to introduce ILE RPG (the language, not the ILE environment) into _any_ other community (Windows, Unix, mainframe) would be laughed out of business.

The "fixed format provides beneficial discipline and free format leads to chaos" opinion is patently ridiculous, as evidenced by the millions of lines of failing and impenetrable fixed format RPG code that's ubiquitous in AS/400 and iSeries shops and the many millions of lines of clear and robust free format code (mostly in other languages) used for a variety of complex systems.

What's true is that you can write better or worse code in any language. But, for a competent and thoughtful programmer, it's much easier to write comprehensible code in a well-designed language and free-format is an essential element.

There are special cases, of course, but for general-purpose, imperative programming, a free-format syntax is something that's almost universally accepted as a basic element of all contemporary languages.

It's painful to see the iSeries programmer community still "debating" this issue when we desperately need to be looking at _real_ issues with iSeries development tools so we can compete.

IMO, Java is a well-designed business language with a few deficiencies (notably lack of a built-in decimal data type). To assess fully Java, however, you must consider it in conjunction with contemporary IDEs (such as WDSc), which provide "integration" capabilities (e.g., with the database and user interface). Roughly, Java IDE tools fill some of the roles (and more) that the iSeries HLL compilers played to great competitive advantage in their early years.

Someone posted an accurate insight that another level of comparison is J2EE vs. ILE. This isn't exactly one-to-one because J2EE covers a lot that ILE doesn't. But one thing is clear, J2EE is complex and (IMO) is viable for small and medium-sized IT groups only with some tooling that simplifies development.

I don't believe at this point there's a single set of strategies or tools that are the "right" choice. It depends on a variety of factors, including what you need to get done, your IT group size and skill set, your current application portfolio mix, etc.

[Shameless promotional pitch: If you live near New York, Toronto, Chicago, or Boston, I hope you'll attend the "iSeries Core Decisions -- Development Strategies" one-day roadshow that Carson Soule, Jef Sutherland, and I will be presenting this Fall.]

There's much more to be said about skills development, but as others have pointed out, any iSeries programmer who wants to be enagaged in something other than maintenance needs to stay up-to-date with ILE RPG features, incluidng the syntax and (most importantly) use of procedures and service programs to support a modular structure. Beyond that, for the iSeries, I think a basic command of Java (which is really a fairly easy language to learn for "core" programming) is important. Then it becomes more difficult because I do think it's more productive to be able to use suitable tools (e.g., in WDSc or an application generator) than to become a "guru" in the deepest reaches of direct J2EE programming (e.g., coding Enterprise Java Beans directly).

As a side note, I just finished reading "Hackers & Painters" by Paul Graham, who makes a strong case that Java itself is far to constraining and primitive and that the truly suitable language(s) for today's rapidly changing application requirements is Lisp or one of the Lisp-like languages. Maybe IBM should trump the Windows and Linux movements by introducing CAR and CDR built-in functions in the next release of ILE RPG !

-- Paul

Posted by: Paul Conte at September 19, 2005 11:10 AM

I am reminded of when IBM said PL/I was to be the end-all be-all of computer languages. I learned it 30 years ago and haven't used it since. I did leverage the skills to learn COBOL, RPG (all flavors), ALGOL, C/C++, etc...

The market will drive what is to be used and the last I heard the 10's of thousands of as/400, iSeries and like shops won't be jumping to a "new" language until they are forced. I firmly believe in expanding my skills, but frankly, RPG gives me the quickest simplest way to resolve many of the problems posed to me by my clients. By not having to find the end delimiter to insure I have correctly phrased my statement saves countless hours of debugging and problem resolution. Of course, if I want a real challenge I can always find the one shop using PL/I as the production language. Til then I'm content to write code that doesn't require me to count characters.

Posted by: Martin Foglesong at September 19, 2005 11:41 AM

I agree with most of what Scott says. Having used CGIDEV2 since it came out, I've been very impressed with the ease of use and efficiency that is offered by CGIDEV2. No surprise to anybody though who keeps up with my views, I don't believe that text based systems are dead. They have their place. I will not do GUI for the sake of GUI. Text based systems are extremely fast over "challenged" connection types such as dial-up. And despite what some experts say, a well written text based screen is easy to use and presents screens in a sequential manner, a necessity for heads-down data entry. Many retail and most larger hospitality companies still use text based systems at the front desk. It isn't because they lack the money, software or the talent to get another interface installed. It is because the text based systems work!

Posted by: jef sutherland at September 20, 2005 4:56 AM

I understand where Scott and others are coming from. But the problem with most ISeries shops is they are running a canned package like PRMS or BPCS or something else. And to provide a GUI interface these companies are doing screen scraping. So it is very confusing to tell someone to go to the green screen to run this report but now go to the web browser to run this report. So this is why a lot of us are stuck with the green screen.

You know we can make all the excuses in the world! :)

Posted by: Ricky Thompson at September 20, 2005 7:55 AM

Toronto has done an excellent job of keeping RPG with, or ahead of, the times. The deficiencies generally are tied to /FREE, including incomplete support for all fixed-format opcodes and the failure to free-form "F" and "D" specs. Yes, even the incorrectly-maligned GOTO and TAG deserve /FREE support if for no other reason than to "send the message" to all those programmers hanging back that /FREE is good, RPG is still growing, and upgrading your RPG skill set is not hard.

We've had a tremendous growth in hardware, i5/OS, and database/SQL capabilities in the last few years. *But* IBM remains focused on what they want to sell: Java, WebSphere, and (today) SOA.

The real issue isn't RPG, it's how we can use it in the iSeries environment.

IBM's Developer's Road Map is IBM's way of recusing itself from dealing with the reality of the green-screen dilemma. Even if you accept the WebSphere value proposition, there's a chasm between the green-screen environment and the full-tilt Java/RPG/WebSphere environment IBM wants for its customers. IBM proposes HATS and WebFacing: HATS is limited use; WebFacing is not much different than being invited to tie the noose at your own hanging. CGIDEV won't die, in spite of IBM's bumbling efforts to treat it like an old Eskimo (push it out on an ice floe and let it die).

Since the early 70's, IBM has, generally speaking, provided an integrated solution for midrange customers. Consider IBM's investment in the S/36-to-AS/400 conversion and the AS/400 S/36 environment, and you have to wonder how IBM could miss the strategic importance of having a Blue Roadmap for the moving to browser-based applications instead of throwing the responsibility out to 3rd parties. I believe moving off green-screen applications is the most signficant change the iSeries community has ever faced, and IBM's response has been to stand on the other side of the river, waving, and saying the water's not too cold.

As an application developer and ISV, I face competition from othere platforms every day. Microsoft has targeted the SMB customer, and the individual who poo-poo's Microsoft's efforts should look at Microsoft's marketing skills and remember there are a lot of very smart, very competitive folks working in Redmond. Microsoft has tripped in the past but they know how to get back on their feet; I believe Microsoft will turn out to be a formidable competitor.

Remember that IBM's fighting Microsoft with WebSphere, and that makes WebSphere a strategic and tactical product.

RPG is just a battle in the war. It's an important battle, and I believe we've won it. But there are other battles, and I'm not convinced IBM has given us the resources to fight them.

Posted by: Reeve Fritchman at September 20, 2005 11:20 AM

I think the biggest advantage of Java over RPG is that there are millions of people writing utility code for you.

Easier to leverage that with Java or JNI that reinvent it as RPG, imo.

Posted by: john elliott at September 20, 2005 12:34 PM

Below an email I received from Asna, an ISV based in GB. I'm not expressing any opinion here, but it makes interesting reading to see what Iseries ISV's are up to.
________________________________________________________

For a simple iSeries and .NET integration strategy...

Dear Sir/Madam,

Further to the launch of ASNA's Visual RPG for .NET (the RPG compiler that sits within Microsoft's Visual Studio .NET) to the North American market last August, ASNA are pleased to inform you that this development environment has now been launched, on schedule, to worldwide markets including Europe.

AVR for .NET was developed specifically for iSeries/400 (RPG) programmers to maximise the potential of web development offered by the .NET framework, inlcuding XML and Web Services. We would therefore like to invite you to attend a one-day hands-on workshop being held in Guildford, UK, on Tuesday 16th December, 2003.

This hands-on workshop will introduce you to programming with Visual RPG for .NET. You will code, run and debug Visual RPG Windows and Web applications within Visual Studio for .NET. You will also discover just how easy it is to integrate your iSeries/400 into the .NET framework. Workshop highlights:

· Overview of the .NET Framework and Visual RPG
· Become familiar with Visual Studio .NET
· Write a Windows application
· Write a Web application
· Introduction to Web Services

Please be aware that places are limited to 10, so we recommend you book early to avoid disappointment.

Email seminarsuk@asna.com to request a booking form.


SIT BACK, RELAX AND LEARN...

Posted by: U.Geerts at September 20, 2005 2:21 PM

Right on to the problem Scott. As I have said over and over, the people on the Iseries will kill the machine.

I just went through this. Showed a programmer about User Spaces through my API. Turns around and hard wires all the User Space API's into his program. Refuses to use the API. "I have to see all the code in the program". We have this wonderful language with all these capabilities and everyone is still writing monolith code. Like I said to this guy.

"How long do you figure companies are going to be willing to have you writing the same code over and over again in a different version everytime when they can hire a Java or .Net programmer who is more than happy to use the API's provided to him and wouldn't think of doing it any other way?"

Our Java programmers are just amazed at the stuff they see being done by AS/400 people. They cannot comphrend the stuff that we do on this machine.

Change or die. That is about our choice and I don't see that happening anytime soon.

Posted by: Alan Campin at September 20, 2005 2:43 PM

There are many topics worth to discuss. I have the feeling many discussions boil down to fi. this question: which party should lead the Iseries out of the black hole it is being sucked in?

1) Scott says: Iseries programmers. Hmm, what power do they have to force management decisions in favor of Iseries? But, isn't it because of the enthousiasm of young software entrepreneurs that made Linux and Java and Eclipse what they are today? Unfortunately, I think everybody agrees that the average Iseries shop shows anything but the eagerness and enthousiasm of youngsters anxious to explore and to experiment the rpg or java universe.

2) ISV's: Hmm, they are a little frustrated because their green screens apps aren't selling for years and because of the confusing roadmap of IBM. A close friend who is the managing director of a small ISV, said to me "I wouldn't go back to that ... (AS/400) even if IBM held $10,000 in front of my nose. Then he starts complaining; I have some green screen accounting package still lying around and customers are asking for maintenance ... help.. my only remaining RPG guys are 2-5 years away from retirement, can't you take it out of my hands?

3) IBM. Yes, in my opinion they created the machine, they layed a strong solid foundation on languages and development standards in the 80-90's, they are the only ones who can pull it off, imho.


Opinions anyone?

Posted by: U.Geerts at September 20, 2005 3:24 PM

A clarification on ASNA and their AVR product line (ASNA Visual RPG).

ASNA Inc. is a U.S. company based in San Antonio. (They also have a U.K. office.)

AVR is a derivative of RPG that runs on Windows. As a consultant, I was one of the original group that developed AVR's CL-derived, free-format syntax for AVR. In the early days, there was a Windows-based editor that _presented_ the AVR code in fixed format, but underneath it was always free-format. This design predated RPG IV's free-format syntax by many years, and RPG IV _still_ lacks free-format declarations, which AVR has had from the beginning.

Overall, ASNA has advanced AVR much faster than IBM has advanced RPG IV. For example, AVR has a "form" data type, which provides a handle to all the GUI forms and controls for the program using a class-like "dot" notation. AVR also supports class definitions -- a fundamental programming language construct that _still_ hasn't found its way into RPG IV.

AVR for .NET can now be developed within Microsoft's Visual Studio environment, providing a first-class IDE.

Relative to VB, C++, or C#, I believe AVR represents only a small share of Windows development.

Although AVR is one example of a form of "RPG on Windows" (there are others, I believe), I don't think it provides evidence that RPG IV would be considered even minimally "acceptable" in it's current form by any significant segment of Windows developers.

BTW, I'm not recommending for or against AVR, just trying to provide some details on where it fits into the picture.

-- Paul

Posted by: Paul Conte at September 20, 2005 3:44 PM

There's a constant background of comments about RPG programmers being aging and out of it that I need to respond to. Bottom line: not in my experience. At 53 I have been one of the oldest AS/400 programmers at every shop I've been in for the last 16 years (I was 37 when I started on the AS/400 in 1989), and I've been in several large shops with dozens of programmers.

In addition, they have been anything but tired. They pushed hard into RPG IV and Java, pushing the limits of interworking with the existing bulk of RPG/400 packages and custom code. There was everything from Delphi/400 to ASNA Visual RPG to Java to web pages.

In every department on project after project there were good programmers integrating AS/400 RPG with every kind of communications and client platforms you can think of. The only thing holding them back was the need to maintain some kind of cohesive architecture. At no time did I see any tired programmers waiting for retirement.

Maybe in small shops is what people's experience is as they make those comments, but I just spent a year and a half job hunting in AS/400 shops of all kinds of sizes, and nearly everyone I saw was younger than me and just as energetic.

I don't know if people are making thiose comments from a limited personal experience or repeating a stereotype, but it's appropriate that this thread is entitled Perception.

That perception is wrong.

rd

Posted by: Ralph Daugherty at September 20, 2005 10:19 PM

Being close to retirement doesn't mean a person is tired or burnout. I've never met any Iseries programmer who was burnout, on the contrary, most are very experienced, focused and have for example enough preseverance to debug monolith 10000 lines rpg programs. You don't ask that of a 25 yr old java programmer, he's going to quit after half a day. But the demographic reality is that most of us are in the 45-50 year age bracket. Considering that in Western Europe, only 60% of the workforce is active above 55 and only 25% above 60, this spells trouble if there is no inflow to replace us in the next decade. I consider it a serious threat for Iseries future.

Posted by: U.Geerts at September 21, 2005 4:28 AM

I've struggled with the questions posed along with everyone else but always end up going around in circles. The problem I think is that the strength of RPG under OS/400 is a unique proprietary solution which everyone dances around but which is the ultimate question.

Either custom coded RPG record level logic programming provides powerful cost effective business systems as we have done for decades or it doesn't anymore, but if it doesn't there is no place for RPG programmers and the AS/400 anymore.

That's ok if true. Systems written in SQL under J2EE or .NET just like the vendors keep saying, and we all move on. Or we can beat the crapola out of those vendors lowest common denominator software solutions with all the power of OS/400 and RPG record level logic. I say we use every unique capaibilty of OS/400 to its fullest advantage and make people happy doing it.

I'm not sure how to change that perception so that organizations once again eagerly embrace the AS/400 as the powerful, custom, cost effective computer to run their business on, but here are some ideas for what they're worth.

Build in the best 5250 to HTML converter IBM has into the EXFMT pipeline as a native interface. Accept commands to generate HTML, XML, or XUL through the built in Websphere web server and of course 5250 binary to terminal connections.

It isn't done right until users can consider the open source Firefox with perhaps an XUL plugin the native AS/400 interface, or even trump that with an Eclipse Java based interface that handles a 5250 keyword XML stream.

Take advantage of the remaining window while Oracle struggles to put out a J2EE Fusion ERP and Microsoft struggles to put out a .NET Green ERP to make open source with something like the Apache license some class RPG ERP with the level of functionality, quality, and aesthetic appeal of perhaps a VAI System 2000.

Maybe get a piece of the action from IBM for every AS/400 running an open source ERP. A small opportunity exists for a global standard ERP base for businesses that aren't ready for a SAP or JDE to run, and run on a Sys/3X, just like businesses used to. The maintenance fees and service for a global standard ERP could be there for the right ERP at the right time.

Along those lines, make OS/400 nothing but cool as the government and healthcare system of choice to run open source solutions for our many government and healthcare massive system needs, needs that have seen $100 million to $200 million J2EE on Oracle systems literally so useless they are thrown away and never implemented.

It takes record level logic in RPG, Cobol, or SAP to handle that much data with logic that complex as we know from experience. We know why Java J2EE fails and we know what we do that works, but the vendors pulling in those hundreds of millions and billions upon billions of dollars for massive failures only have one mantra, open standards that suit their lowest common denominator cross platform dreams.

We can beat those vendors with open code that actually works, and it will become a standard. Governments consider Windows and .NET Basic in that realm. Why is OS/400 and RPG any less open than Windows and .NET Basic? It isn't. What is standard is what is wanted.

I can't think of a better time to help our government than now anyway. They are reeling from massive development failures and out of money. Done right with OS/400 features such as message files any government could use and contribute to it. This is a battle of OS/400 and RPG against J2EE and Java and .NET and CLR, even though we run those as well. It's a battle we should have already won, but we never marched onto the field. It's time to show that OS/400 and RPG is the software solution that goverenments should run on successfully because business .runs on it successfully. There's a reason for that. Open source solutions would highlight and demonstrate that success.

In that battle, ILE should be used to full advantage to include C++ and Cobol with RPG and Java to provide effective open source solutions. In that blur of software merging, IBM can help by implementing some suggestions I've made in the past, allowing /free to accept open and close braces in lieu of a semicolon and END statement for block commands and allowing other Java notation such as the == and != comparators. There is no reason for the code to vary from Java syntax.

Better yet, they should just include the equivalent Java and Java Toolit code syntax as /free syntax. Just because it looks like Java doesn't mean it has to generate bytecodes to run in a JVM. I don't know if it could be slick enough to have new() statements treated as D specs, but what if the code would compile in RPG /free or Java? Would that be a transition or what?

rd

Posted by: Ralph Daugherty at September 21, 2005 11:44 PM

I went on the ASNA seminar in 2001 and wrote a long comparitive report between AVR and VARPG. Took a big yawn ... and recommended Websphere!

Posted by: Hassan Farooqi at September 22, 2005 5:33 AM

Sorry, but RPG is outdated, except for a limited range of applications. These applications would include anything that does not require lengthy, formatted text or does not require more than 80 column text.

Granted, the display can be enhanced via add-on products, but why? This means you are using a patchwork set of development tools just to create someting that a kid could do using a program as simple as Microsoft Access with an SQL Server back end.

Yes, the standard reply changes the subject to the reliability of OS/400 or the stability of DB2/400 ... but as I just said .. this is changing the subject. RPG and the green screen environment is obsolete except for basic accounting applications.

This raises the question, "Why have two systems in house; one for basic accounting applications and another applications beyond the scope of unaugmented RPG?"

Adding JAVA will provide a gui and text, but you also add the need for labor intensive applications. Most of which can be purchased and be far superior to what most in house coders are capable of.

JAVA is also obsolete in many respects. While it will be the power behind systems programming for years to come, it is a pure dud as a business application language. It has a large learning curve, requires expensive programmers in house, creates long and often incomprehensible programs, lacks maintainability by inexperienced coders, and, in summary, is very costly.

Posted by: bob2006 at October 4, 2005 10:37 AM

I believe IBM provided poor leadership in the AS/400 and later the iSeries marketplace . It is difficult to go into an accounting department, where there are tailored RPG green screen application that have been developed over years and justify rewriting everything! I really thought that IBM would challenge MS Visual Basic with a Visual RPG of its own that would have given us an easier migration path; the ability to retain some of our code and gain the advantages of GUI, web browsers and Internet. It seems that IBM put Visual Age RPG on the market place, but then encourage everyone to go to Java and scrap what we did in RPG and Cobol. Many on the iSeries believe that CGI programming-so called e-RPG- is the way to go. There is ASNA's Visual RPG which has merits. Personally, I like the kind of things I see with RPGsp, which is like a mid range case tool that writes GUI web based applications using CGI. IBM dropped the ball! Now we have to pick it up and run with it, but we must hold onto the ball and run towards the right end zone!

Posted by: John Polucci at October 4, 2005 4:32 PM

G'day Bob2006,

To say "RPG and the green screen environment is obsolete except for basic accounting applications." is incorrect and indicates to me that you only see computers being used in a clean-room office environment.

I have worked for a great many S38/AS400/iSeries/i5 shops where this is not the case. I have seen terminals deployed to dusty, noisy (from a aural and electronic perspective), hot and extreme environments. Normal PC's do not function in these environments and the most robust terminals are the old twinax terminals which have no moving parts and very little circuits.

Data entry in extreme environments - especially cool rooms where people wear thick layers of clothes and gloves - presents another problem in that using a mouse is impossible.

I still advise clients to use their 5250 screens as they are more productive and reliable for the environments they are deployed to.

Posted by: Darryl at October 5, 2005 8:25 PM

Scott, Wow. So many good points. Re the last one about perception - so true, yet so few of us seem to understand the importance of this anymore.
Mgmt's perception of their iSeries programmers is something to think about. We're aging and greying, and from what I can see many of us aren't challenging ourselves. In the ultra-dependable environment of the iSeries it's easy to adopt an attitude of 'conservative-efficiency'. We can provide very minimalistic solutions in a relatively short period of time, and that often becomes the default mode by which we operate. This is fine for maintenance and help-desk issues. But for new apps or major revamps this attitude makes us and what we produce look tired and simple.

We all lead busy lives. So where do we find the time to learn and play with new languages/techniques? As Tony Robbins would say, it all comes down to pain and pleasure. What are we willing to do to avoid the pain of not getting a raise or not having a position in a few years? What are we willing to do to have the pleasure of being an integral part of the forward momentum of the enterprises we work in?
The way each of us answer these questions, shown through the actions we take, will determine the perceptions we shape in those we work with and for.

Posted by: Buzz at October 6, 2005 6:46 AM

Good points, Scott. I recently learned that HTML can be effective for stylized reports, and that RPG is excellent for formatting report blocks, controlling absolute vertical and horizontal positioning, and controlling page breaks for portrait and landscape printing - not just traditional data access.

Clicking my Name should show a template for a student transcript which may illustrate.

Using RPG for Web applications, stream files, and email would be worth putting on a list of things to learn.

Posted by: Nathan Andelin at October 7, 2005 5:09 PM

I will hand it to you, Nathan, the web page example has a lot of data on it. I've seen some pretty paltry web page examples before, but not yours.

With a 5250 screen we would have a year select drop down list with the data in a subfile, probably with a function key to see and add comments and another one for medical records. There's a natural way to work with data, and the web page isn't it.

There's no navigation on this example screen but if there were the user would have to tug the screen down to get to the navigation buttons.

I have an old browser so maybe you have navigation in a sidebar and I just don't see it, but in any event I find the whole concept of a business user dragging screens up and down to work with data ludicrous unless it is a powerful screen such as a spreadsheet.

When I hear real people say they want to work this way instead of IBM saying it then I'll change my mind, but as far as I'm concerned Microsoft is going to eat IBM's (and my career's) lunch because the web page won't compete with a Windows interface.

An Eclipse based AS/400 interface that interoperates with the desktop will compete. This sorry IBM web page stuff won't, not for business users.

rd

Posted by: Ralph Daugherty at October 9, 2005 4:19 PM

Highly interesting post Scot -

My feeling is that we will see a scenario something like this by the end of the day:

-) a DSPF consisting of constants and variables.

It would suffice to extract the variable in a WYSIWIG and prepare the HTML/JAVA pages where:

-) one "format" = one page.

The developer would retrieve the DSPF variables managed by RPG in a WYSIWIG and would insert the graphical objects in place.

These HTML/JAVA pages would be stored in a DSPF *file object with the 5250 image which simply means that an RPG program can manage the variables. It would then do to display the 5250 screen formats or the web formats through SOA.

On the one hand, it would be the HTML/JAVA presentation and on the other, there would be constants with DSPF attributes.

IBM would have to manage through SOA that it works also for the web.

The web presentation should not be the developer's problem; his/her aim would be to reply to user needs.

Next, either the green screen or an HTML/JAVA screen would be just a question of display format.

SOA should bring an elegant solution that would help the iSeries to be launched for the future.

Please remember what Dr Frank (Soltis) says: "on the iSeries, it is users who are asking for solutions and not the technical people".

And finally, we will see what solutions are available by the end of year...

Posted by: jean at October 12, 2005 7:34 AM

Jean: Isn't that what HATS is? Or WebFacing? Or about a dozen different 3rd party packages?

The problem is, GUI has a lot of capabilities that the green screen can't have, and it doesn't follow the same programming paradigms as a green screen.

So, while it's possible to do what you're suggesting (and there are many solutions that do it already) you don't get the complete GUI look and feel. Instead, you get all of the limitations of the green screen, but reproduced as a colorful web page.

Look around the web, on each web page you visit, think to yourself "how would this page have been generated from a 5250 display file? Would it even be possible?"

Posted by: Scott Klement at October 12, 2005 2:13 PM

To think in terms of 'display'files or 'printer'files is completely outdated. No other development system thinks of a display screen or a printed report in terms of "a file". Subfiles were for example a methaphor invented by IBM when they thought of it like they were rolls of paper scrolling behind a terminal window. Then they extended this notion to printerfiles and communication device files.

Objecttype *files with attribute dspf, prtf, icf?? All rubbish concepts from the past, destinated for the wastebin.

Posted by: ugeerts at October 12, 2005 3:05 PM

Scott, I have a couple of examples:
Imagine that you have a user friendly graphical user interface like FrontPage for example...
Then imagine you have to insert a big form and an RPG program that does it properly. The only thing remains is to pick the data from the existing program and you bind it whit SOA.
So SOA is THE solution to insert your RPG Program.
Another exemple: I want to insert a list into a page. My RPG program knows how to do that. I only pick the data of the subfile.
So I will have the same graphical possibilities as FrontPage and RPG program together.
If I put the HTML/JAVA pages into an object *file with or without the 5250 image I solve the problem of iSeries with SOA.
In fact, it is the RPG program that manages data, not the 5250 green screen.

Posted by: Jean Mikhaleff at October 13, 2005 1:51 AM

I have been working in the green screen environment since its inception in the mid 60's. Our existing problem has been caused by IBM. They have had their heads burried in the sand for a good number of years. Our company is now in the process of positioning ourselves to move forward with or without the iSeries. If IBM does not step in and take an active roll in providing and supporting a modern versatile UI; not to mention capable tools for forming and handling spooled type output, it is only a matter of time for the demise of the iSeries. We are looking to hedge our possition by using the ASNA products that will allow us to move up to the Visual Studio.NET world and then be able to use the iSeries or other servers if necessary. In addition, we are using the Looksoftware tools to help us move in this new direction. Our transition will be an evolutionary process, rather than a revolutionary process.

If IBM had any forsight, they would buy up a few companys like ASNA, Looksoftware, ReCrystallize Software, etc. and bring some fresh vision to how the iSeries can thrive in the new SOA world.

Posted by: Larry Gustafson at October 13, 2005 8:14 AM

Hi all,
Ideas, so much ideas ...

I work with PDM and VB since 15 years.

If IBM provide us a development environment (WDSC, Eclipse, ...) that permit us to make GUI interface in front of RPG program as quickly as (at least) Microsoft permits with Visual studio, I can convince my boss that developping with the 400 is cheaper than with windows.

And my boss will buy 400 instead of windows.
And I will learn this new tool .

Today, my key to choose an IDE is the documentation.
sample : when you try a new api in MS IDE, you cut and paste the code, it runs, then you learn ...
on the other hand, when I try to uses an IBM API, the sample is not in RPG, and if you try to cut and paste it in the good source text, it don't run.

And someone say programmers are afraid to uses API ... for me it is normal. I don't agree with spending one day to decript the API when it needs 5 minutes to integer it after having understood how it works.

Posted by: jp lamontre at October 13, 2005 8:34 AM

To Ugeerts:
Your question/comment:
Objecttype *files with attribute dspf, prtf, icf?? All rubbish concepts from the past, destinated for the wastebin.

Mine:
As I look at Explorer, my latest J2EE incantation(IRS project), lists files with extension .jsp as labeled as "java server page", the .js files as labled as "javascript file", the .java files are labled as "java file". Now, other then the .js/.jsp/.xml..tld.etc files everything else(the vast majority) are lumped into category "java file". In fact, I can and have written complete swing applications with mostly .java files, which means without naming conventions one file is indistinguishable from the next, but I digress....

Please tell me how this is any better than labeling files prtf, dspf, icf?

Posted by: HelpsUsUnderstand at October 13, 2005 9:02 AM

Dear HelpUsUnderstand,

I don't think Ugeerts was trying to start a war. Rather, I think he was relating to the limited range of use for the green screen and all of its appendages. Everyone who designs complete and complex applications will mix a number of file types to achieve a given result. My speculation is that he wants to limit the number of file types and required development tools and, thus, bring application design closer to the advanced user than the professional coder.

The need for the professional coder has greatly diminished. All you have to do is see how many are unemployed or have had to change professions to know this. All good managers who need applications will first try to meet this need without an IT department. If this is not possible, then the good manager will use tools that minimize the need for a bottleneck such as a coder.

Good advanced development tools reduce variability and improve quality by removing coder fads, personalities, politics, and personal skill levels from the equation. What used to be nearly impossible for a coder is now commonplace for an advanced end user.

Since RPG requires coders, wallows in fads, has limited visual appeal in the resulting applications, and can not handle text or graphics without add-on tools from miscellaneous vendors, is will pass / has passed on for all except maintenance work.

Personally, if I were managing an IT department for a small or medium sized business, JAVA would be a minor player. It is too labor intensive and too complex to be easily maintained by lower level employees.

Posted by: bob2006 at October 13, 2005 9:58 AM

The exaggerations regarding the iSeries "green-screen" seem to control these discussions. Let's put this whole thing in perspective.

First off, my hat is off to all the fine engineers that developed green-screen to begin with. Without which many professionals would not have these good paying jobs or advanced careers.

Second, green-screen and many technologies were the fore-runner of todays GUI environments. Without the "dumb-terminals", advancing to intelligent terminals Windows would not exist. Advanced graphics existed long before Windows ever hit the market. So from a technology perspective Windows really isn't anything special.

However, there have been two threads to this discussion, one GUI the other features.

Regarding GUI I do not believe this will change much of anything because it really is all wrapped up into the features. However, if we explore the GUI issue, then a plausible solution would be to re-invent OS/2 as the iSeries GUI. Why waste all that technology, just ramp it up a bit. That said, IBM will never do this. No matter how many years or dollars, whatever is done will always be compared to Windows in the end, so why re-invent the wheel everyone already owns.

As far as features, we have to roll back a few years to the good old green screen PC(or amber). It was a huge success, no windows, no freaking mouse, no 256 billion colors to give us headaches. Just an instant hit. Why? Because it allowed us to do things my big brothers couldn't or wouldn't let me do. It's the same today, anyone with a PC can run a world enterprise from their desktop so to speak. It's all about empowerment and control. And yes DOS apps still thrive and exist today.

As a gov't contractor, day in a day out I see Windows farms, and day in and day out they crash, mail gets lost, etc,etc. But you know what, no one says, we need a mainframe or we need a Sun Box or we need an iSeries. They put up with Windows because they control their own environment. GUI is not the reason Bill is rich and famous, it's because Windows made many,many ISV's rich along the way and it still persists today.

It's all about choices, adding a GUI to the iSeries is nice but GUI will not change all that much except for those already staying put and any new converts. I admit it certainly will help with the pros and cons checklist, but now it's GUI on both platforms, so then the next question is available applications. Okay, so how many spreadsheets, spreadsheet tools, and spreadsheet scripting languages on the iSeries can anyone name - none! Same could be said for many other popular tools.

My shopping list includes allowing complete ISV machine level access so I can have as many choices as my Windows counterparts. The days of IBM doing it all themselves have to be over. In other words, if another vendor wants to build and sell an HTTP server, application server, etc well let them and help them do it!

And your comment "has limited visual appeal in the resulting applications, and can not handle text or graphics without add-on tools from miscellaneous vendors" underscores what I'm stating. The very thing you accuse of needing on the Iseries is actually sorely lacking on it but exists in abundance on Windows and other platforms, of which seems to be the rant of many pro-Windows zealots on this topic. In other words, without 3rd party solutions Windows is dead. If IBM embraces 3rd party then it will continue to thrive. Bottom line message to IBM would be allow ISV's to build it and they will come.

I'm not sure about your Java comments, but yes it is complex and it requires a solid base of OO experience in order to use it effectively. That said, after 7 years of Java and 25+ years of programming many env's and languages, Java is far over-rated and is more like the proverbial shotgun to kill a fly approach. Personally I think Scott's marketing plan went awry and Java took off shocking to all those involved with it.

But anyway, I'm a big iSeries fan but do not see GUI or RPG as an Achilles heel. There would be no Windows w/o 3rd party, no Word w/o WordPerfect, no Excel with Visicalc and SuperCalc, no GUI w/o dumb terminal. Either way the technology planet stacks up, it can still thrive on the iSeries, it's all up to IBM.

Posted by: HelpUsUnderstand at October 13, 2005 12:30 PM

We use Profound Logic RPGsp and can develop true browser based (all you need is a browser on the desktop) applications just as fast as the old fashion green screen DDS apps. My programmers did not have to re-learn a new language (java), they just had to learn html and other browser tricks (which is a lot easier to learn than a new language, whether that is java, rpg, cobol, etc.) Nothing reads native AS/400 DB2 files faster than native AS/400 RPG because it was designed that way from the beginning (no ODBC, etc). I have always been disappointed in why IBM did not take RPG to the next level of making HTML just another I/O device (no api’s or CGI, thank you) like DDS. This is basically what Profound Logic did, simply (or almost simply) replace DDS with html. I looked at all of the Web-RPG tools and converters and it just was not the same as good old fashioned RPG interacting with html like RPGsp. We also use other 3rd party software that converts spooled files into HTML for pushing to the web that looks very clean and you can do text searches on. We try to connect as many Windows Servers using IXA adapters (Windows is here to stay so you better except it and take advantage of them whenever you can) whenever possible. Take it from the dinosaurs, if you do not adapt to change, you will perish.

Posted by: Joe Fierro at October 13, 2005 3:53 PM

It makes sense to keep what you have if it still works or is not financially sound to replace. This is just common sense.

However, if you put racing wheels on a Model-T, it does not become a race car. It is just a souped up antique.

Company presidents do not wake up in the morning and feel the need to make it a better world for coders. I can guarantee they don't insist on buying two computers ... one for gui and one that is otherwise marginally acceptable ... when shopping for new computers. They want one that can do it all with respect to ordinary data processing. And they want to run the IT department without creating a jobs program for programmers and other IT types.

The problem is the old time coder's view of what it means to process data. There is the paleolithic view that 'it was good enough to replace the key punchers, it still is good enough.' There's the heroic view that says 'We can make it work if we patch together enough parts and buy a couple more.' There's the Java junkie view that says 'Code Code Code and more Code'.

Company presidents want to treat IT like a light switch. It just works when it is needed. It does not require old and new light switches just to get some illumination. No candles or kerosene lamps just because they once worked great. No teams of electricians just to keep the lights on. Just plug in a power stick if you need a couple of more outlets.

Posted by: bob2006 at October 14, 2005 5:38 AM

Comment on HelpUsUnderstand posted on
october, 13 2005 12:30 PM

"just perfect"

Posted by: Claudio Cuzzi at October 14, 2005 5:56 AM

Honestly, without putting too fine of a point on it, the older programmers in this field who refuse to reskill in the latest developments of RPG, are killing the machine in the marketplace.

This machine has become synonymous with that crusty old persona. The argument given, "...but they bring business experience to the table..." is overused and obsolete given that machine is not being purchased for newer applications.

As an owner of a software company who offers two major products based on this machine’s latest capabilities - web enablement and 2-way messaging, I fight this battle daily with potential clients who associate this great machine with clumsy legacy applications and legacy staff who revel in the stories of their archeology based "careers."

Companies won't buy the iSeries simply because they don't want the type of crusty staff that must come with it.

This is an industry of change - if you want to make the same thing year after year, you should contemplate making pottery from a kiln.

Posted by: Brett Gibson at October 14, 2005 10:54 AM

I think your company, Adventure Tech "born out of the need to provide services to those venturing into the intimidating jungle of technology", in other words sell products that the crusty old RPGer needs cause they can't figure it out themselves.

Unfortunately web enablement and 2-way messaging technologies are fairly common-place so I would suspect that convincing potentional customers of the iSeries innovation based on those products would fall short.

If your so bent on starting over, then innovate around the old RPG farts that are in your way and lead everyone into the light. Did Bill Gates let anyone stand in his way?!

I would start by not blaming anyone else. BTW, if anyone is to blame for dumbing down the midrange it's IBM. Or from a crusty old fart, it was built that way stupid!

Name a machine that does not run it's legacy? There is not one major machine or architecture today that will still not run it's legacy. So just ignore the legacy and shine on brother!

Good luck and good day!

PS I make damn good pottery too!

Posted by: AwComeOnBeNice at October 14, 2005 12:06 PM

To HelpsUsUnderstand regarding my comment on Objecttype *files with attribute dspf, prtf, icf.

I am only pointing to the metaphor IBM used regarding i/o to devices as being equivalent of something like writing to or reading from a file, a file in the sense of a table holding data records.

This metaphor allows them to use in legacy rpg, cobol a language construction like "read screen", "write screen" the same way as fi. "read customer table". The problem with this concept is you don't store data or retrieve data to a screen as you would to a table. This leads fi. to the confusion that a system administrator would make a daily backup of all objects of type "*file", since he may assume the data is changed daily, as opposite to "*pgm" objects, which only need to be saved after the install of a new release. But *files of type dspf, prtf, icf.. are invariable objects too, so they need not to be backed up daily.

In fact, every other development systems treats device i/o just as it's supposed to be, with a dedicated api to communicate with (the driver/rom code of) that device.

The file notion would be better replaced with today's Object Oriented paradigm, whereby you have an object and a natural verb (method) to operate on that object, like in fi. rc = screenabc.put_title(....).

Posted by: ugeerts at October 14, 2005 2:14 PM

When I continue to read this thread, it comes to mind it seems there are basically 2 camps of posters, the kind who say stuff like "rpg is obsolete, it is 10 years behind all other languages since it didn't pass the transformation to the "visual" and "OO" evolutionary stage)
and the other kind of posters who react furiously to those statements and argue IBM engineers did an excellent job in developing and maintaining the ISeries and RPG as it currently is.

Just an observation, and this bipolar camp attitude seems to attract a lot of posters imho.

Posted by: ugeerts at October 14, 2005 2:27 PM

VHS versus BetaMax.

Betamax was better, smaller, higher quality, yet it lost to VHS.

The best technology does not always win. Often the open and easily accessable technology wins. JVC won the video cassette wars when they openly shared the VHS technology.

Posted by: OneMist8k at October 14, 2005 9:07 PM

Some reactions to comments:

I've seen no furious reactions from anybody here. Also I've seen no posts that hint of "superior" technology of any kind.

For the person that thinks the people that mention GUI here are referring to Windows and are thus Windows zealots, GUI among IBM minions is a code word for web page.

It should be called wannabe GUI but they drop the wannabe.

If I were wrong all these years about a web page being an insufficient substitute for the 5250, much less a real GUI interface, then I wouldn't be reading weekly in all the trade rags about the failure of the web page as interface and attempt to kludge an acceptable solution with AJAX non-HTLM standard workarounds along with many other technologies being introduced.

In any event, roll your own HTML from within code didn't cut and won't cut it. I argue for an XML buffer with keywords to be used by any interface but IBM has sacrificed us to their Websphere empire dream, it's that simple.

They assimilated us because we're the only ones they could. Customers will buy the new Sun server running Unix, Linux, and Windows or the new Linux servers running virtaul Windows without ever considering what IBM has made the AS/400, a toy box for running Linux and Windows.

Thank you, IBM, may I have another.

rd

Posted by: Ralph Daugherty at October 16, 2005 10:43 AM

A lot's said about developing 'web' applications in RPG but what about developing a 'GUI' application (one that doesn't need a web server) in RPG? Or better still, one that can run either in a 'browser' mode or in some kind of GUI client (Java perhaps?).

How realistic is that?

Posted by: John Lewis at October 17, 2005 8:15 AM

who and where will learn RPG in the future?

Posted by: Andrew2005 at October 17, 2005 10:30 AM

who and where will learn RPG in the future?
Excellent question. Who? No doubt will that be industrial archeology students. Where? Well, the class accomodation at IBM isn't what it used to be, so you'll have to do with the toilet. But before you start to shriek, all in all it's not that bad, since free cigarettes will be included.

Hopes this satisfies you question,
signed ibmreadytoteachforthefuture.

Posted by: ugeerts at October 17, 2005 2:37 PM

The question is why learn RPG, and the rest will take care of itself.

People will learn RPG when there are jobs in it, the reason we all learned it. Those jobs went away fast when our manufacturing base went away. That was the core of the Sys/3X world.

We are told that IBM's big market for the AS/400 is the Pacific Rim. Take a wild guess where our manufacturing base went.

As for learning RPG when this J2EE IBM and Oracle failure implodes along with the rest of our economy, I learned RPG in a month at a consulting company and two weeks for CL and operations. And that was typical.

One person from that class was head of AS/SET training at SSA when I worked there as contractor a few years later, another became head of applications development for a billion dollar manufacturing company, all of us were billable and productive when we got out of that class and sent to clients. That's the power of RPG and OS/400 and the great packages that run on it.

I'm not worried about people learning RPG when the time comes. I'm worried about there being any reason to.

rd

Posted by: Ralph Daugherty at October 17, 2005 7:13 PM

Thanks for the Post Scott - I agree with everything you said, but I still believe the ( whatever they are calling it today) is going it the toilet because IBM wants it to go in the toilet. You show examples of what can be done, but you do not mention how hard they are to do if all you have is the free documentation IBM provides. The functionality I want may be there somewhere, but it is in some obscurely named API with examples in C and notes in Swahili for all I can tell. If IBM was serious, they would buy ProData and hire you and Oreilly to do documentation for all the GUI and Net capabilities.

Posted by: Harvey Lewis at October 21, 2005 11:06 AM

RPG it's cool but IBM likes to do harakiri and want to, in a future, joy disapearing the RPG and Iseries

Posted by: miguel at October 21, 2005 3:32 PM

For the person who said ***For the person that thinks the people that mention GUI here are referring to Windows and are thus Windows zealots, GUI among IBM minions is a code word for web page***

Referring to Windows does not imply Windows zealot, but constantly using Windows as a reference tool for "advanced technologies" for the sole purpose of denouncing the iSeries or any other platform as legacy in a negative context regardless of truth, does imply a form of zealousness.

As for GUI equals web page, you maybe referring to IBM’s penchant for all things Java to avoid the Windows issue altogether. However, the vast majority of all computer users will instantly equate GUI with Windows(on any device).

Posted by: Reply at October 24, 2005 12:15 PM

There wasn't an imply, the person said zealot, as in "The very thing you accuse of needing on the Iseries is actually sorely lacking on it but exists in abundance on Windows and other platforms, of which seems to be the rant of many pro-Windows zealots on this topic."

It is easily understood why they would think Windows when IBM web page people say GUI. That was the point of my post.

rd


Posted by: Ralph Daugherty at October 24, 2005 6:42 PM

There are Windows zealots on this blog and no one will believe that GUI means web page, IBM employees or not.

You said: GUI among IBM minions is a code word for web page. So then exactly what does this mean?

Posted by: reply at October 25, 2005 8:58 AM

I've read that analysts are predicting 25%+ growth in iSeries revenue this quarter, following 25% growth in the 3rd quarter, following 10% growth in the 2nd quarter.

Interestingly, this growth consides with IBM's reemphasis in promoting the native runtime environment, improving relationships with 3rd party vendors who develop products for the native environment, and IBM's deemphasis in promoting Java on the iSeries.

Our company is growing too by developing new native Web interfaces for an existing database.

Posted by: Nathan Andelin at October 25, 2005 1:35 PM

Three phrases should be among the most common in our daily usage. They are: Thank you, I am grateful and I appreciate.

Posted by: enzyte at December 13, 2005 3:37 PM

Where did the last active blog go, it's not listed anywhere else?

Posted by: user at December 13, 2005 3:51 PM

I couldn't find it yesterday evening and came in to send feedback to the site about it, but this looks like as good a place as any to do it.

At least from this "last post made to any thread" window on this iseriesnetwork home page, the iSpeak Blog thing can be accessed by clicking on iSpeak Home, which is pretty obscure for openers, or wannabe openers like me.

But then when you get there the threads are listed in creation date descending rather than last post date descending, which if I ran my site that way no one would be able to easily add posts.

The final nail in the coffin is that I can't even find how to go to another page of these threads to get the next dozen or so created however long back in an aimless search for the topic du jour.

Just because I'm stubborn and wanted to see if the thread even existed anymore, I clicked on the top post and started clicking on Previous, and lo and behold I was able to click back through the mists of time and find the Loyalty thread, apparently created in the Big Bang of this blog.

My speculation is that iseriesnetwork is emulating IBM's web site navigation or possibly not wanting to show them up by having a navigable web site where you can find something, or possibly doesn't want to confuse us or give us false expectations that such a thing can exist.

My recommendations are to list the last few topics commented upon in post descending order on the home page. I do this on my home page http://www.justiceforchandra.com with a snippet of SQL in PHP.

Also make the click through to the blog a little more obvious and then list threads in post descending order rather than creation descending order. I do know they have a real forum here, I registered but haven't posted there yet, and maybe we're supposed to be seeing this stuff from an RSS feed and not supposed to be able to navigate it, I don't know, but right now that thread we were posting on yesterday is lost in time.

Now, back to regularly scheduled programming...

rd

Posted by: Ralph Daugherty at December 14, 2005 5:52 AM

-- lost in time... --
Ralph,

Well I hope it is was accidental. I just get the feeling that anything pro-Windows is okay to rant, but challenge those assumptions and all of a sudden the post(s) or thread disappears. I think vulgarity can be censored by filters, even removed by server code(if that was the case). But as a paid subscriber to News I'm not thrilled with this. I was asking some important questions of a blogger who thought he could "laugh" his way through. Unfortunately, either the response was too offensive and/or the whole thing "accidently" disappeared.

In any event, the home team needs to chime in here.


Posted by: AngelInTheMorning at December 14, 2005 7:48 AM

Hi all.
Oops, where has that "loyal customers blog" gone too? Someone called 'enzyte' trew in some meaningless post about three phrases... until you click on the guys name, that reveals his true aim. Is there a webmaster in the house to take care of enzyte's crap -or- is this place deserted ?

Posted by: ugeerts at December 14, 2005 1:23 PM

Ok, since this site is running on autopilot I'll go retrace my steps in the wayback machine and bring that thread up for air.

I noticed that with this blog style discussion that the entire thread prints on one web page, so the blog designers (iSpeak apparently) probably decided that threads should only last so long before they become too big to render on a web page.

In English, that means I guess we're supposed to start a new thread frequently even though you lose the context of the ongoing discussion. What I would do as the admin of my site in a situation like this is to split the thread and start a new one with for example the last few posts that was focused on a particular discussion. But the admin would have to be following it to have a clue when to cut and start a new blog page.

In lieu of that, when it comes back up probably best thing to do is copy the post you're responding to and start a new thread. That's probably all that can be done with this blog stuff.

rd

Posted by: Ralph Daugherty at December 14, 2005 6:47 PM

Folks, it's all too late and too little. The iseries will vanish within 10 years and rightely so. There is no real advantage with as/400 and rpg anymore. Of course, in the seventies the os was quite modern, with it's robust architecture and the consistent design of all things are "objects" (NOT to be confused with objects we know from oo-languages like java). Its just consistent design, nothing more. But no progress since then. Build a system in free java and free linux using state-of-the-art free eclipse and you also have a robust system (java compiles to bytecode, just as an rpg program on os/400 compiles to intermediate code, giving the robustness). I dont ask for regular expressions being used to search files on os/400, that would be too much of course. But hey, why cant i still not choose to search spoolfiles using the case-independent option, which every search-option in whatever crappy software has, but not in the big-bucks as/400. Etc, etc... In the fifties rpg was also quite a modern tool (maybe the first program generator). Nowadays the "modern" rpg (modern meaning early seventies maybe) is a convoluted, complex language with a lot of cruft, no vision (if you like to know the vision behind rpg then go to a musea and look at some of the old punchcard-sorting machines which made ibm big), new "features" which are introduced randomly, etc etc. Its a mess. But, we can make modular code now, wow! Only sad that just 5% in the rpg community now how (and i dont mean syntax) or is interested. As Paul Conte said earlier (i like this guy, he seems to really have a grasp about what software development is all about) its painful to see the rpg community "debate" about the pro's and cons of free-format. No, i dont think there is any hope for the as/400. Its proprietary, it has a bad image, and a developer community which is simply reluctant for change (imagine the popularity of windows if developers kept building dos programs). Ibm gave up on it and is just trying to squeeze as much money out of it as possible (i dont blame them). The problem is mainly with the developers. We had these discussions 10 years ago also. Nothing changed since then. And 10 years ago developers had to learn new skills, demand better tools, and maybe something would have become out of the as/400. But no, most rpg devs are just doing there thing every day, and hoping to reach their pension. Only learning new things when forced. "Rpg is good for business programming". I read this often, but i never read why. Why? Is this only since ILE RPG then? I mean, being restricted to 6 chars for fieldnames is what it takes to be a good business language? No date support... sure, in businnes we dont use dates. But now rpg supports long(er) fieldnames and dates, wow. Oh yes, it has good db access (dBase anyone? No, rpg is really good for nothing. Just for all those rpg "programmers" sticking their heads in the sand. Save those comments for your it managers, maybe they'll buy it. If i need to sort some data in rpg in a special way i need 200 lines of rpg code. Not to mention the poor guy trying to figure out what the code does (dont tell me about docs, thats not the point i'm making here). In java, it takes 5 lines of codes, so you can see at once what it does. Why? Because in java we not only have one-dimensional arrays, but n-dimensional arrays. But not only that, in java you also have sets, bags, vectors, stacks, lists, maps, etc etc. But.... we dont need that as rpg programmers. We can do all we have to do in rpg. Well... you're right... why not program in machinecode, then you can really do whatever you want. I dont expect any descent discussion, i'm tired discussing these kind of things, it doesnt help, rpg programmers just want to do the same thing they did for the last 20 years. Thats it. Yes, business knowledge is important... OF COURSE. But... think about this... you can better write a program thats easy to maintain which does not quite (yet) do what the business wants, then write a crappy program which exactly does what the business wants.

One advice... MOVE ON OF YOU CAN, AS FAST AS YOU CAN!

Posted by: onefrustratedguytryingtogetthehelloutofthisiseriescrap at December 28, 2005 6:59 AM

Ugeerts by another name maybe?
I cannot agree. I am using RPG with CGIDEV2 to write Handheld Web Based apps and they work really well and a lot faster than a lot of the other software found on Handhelds.
I am writing Web pages WITH validation! Where do you see that on the www?
The main difference I see between us old RPG programmers and young PC programmers is Standards in all areas.
We were taught properly and we had to conform to standards. We learnt in the work place. Java, C++ etc is often learnt in College and then applied without any work experience. All the traps that we old programmers avoid (through experience) are eagerly fallen into by many PC programmers.

I am not saying ALL PC programmers are bad. But enough of them are.
Have you ever tried to get a sensible answer from an HTML or Java forum? They are full of themselves and see themselves as Guru's when in reality they are beginners. Sure they know the syntax and hey they can create a pretty Web page, but you can easily poke holes in it.
I believe RPG has a great future. IBM should focus on CGIDEV2 and give us an HTML editor. It would also be nice if the WDSc was as good as Microsofts Visual Studio or, even better if IBM could get together with Microdoft and include some of the stuff from WDSc into Visual Studio.
Take a look at PKS's products. Thats another direction we could take.
If only IBM were looking at things the same way as ASNA and PKS are.
I am learning the Visual Studio but whenever I can I write in RPG, it's easier and I can update a database (acurately) with far less code than I've seen used in Java.
I am not at Scott's level of expertese but I can write amazing things using RPGLE, FREE and CGIDEV2, any RPG programmer can.
Lets all show our companies what can be done and maybe we can save the best machine in the world despite the nay sayers!

Posted by: Mick at December 28, 2005 8:34 AM

No Mike, i'm not Ugeerts.
As a professional, i like to pick the tools that meet my job requirements. Meaning in software development, i need a programming language and tools with which i can create and maintain software meeting business demands agains lowest possible costs during the life cycle of the software (which is 10-20 years maybe). This is a fairly general statement about software development, but true i think. So... does rpg, CGIDEV2 etc deliver this to me. No, not to any extent. Of course, i can create all kinds of apps with these tools. Like i stated, i can make every webapp with RPG and CGIDEV2 as i can with java, JSP etc. But... what about the costs. The long-term costs i mean. Choosing rpg and cgidev2 just because java has a steep learning curve is not a professional argument. And why is this learning curve so steep? Because rpg programmers generally dont learn new methods, techniques and concepts. I learned about OO in the late 80's, before it was popularized in the 90's, playing around with a language like smalltalk. It's fun, learning new things. And you can use OO concepts even when programming in rpg. As a professional developer you want to keep the costs down (again, the long-term costs) and make your work predictable and manageable. This is all about organisation of software, not learning a few new BIF's. What gives rpg me to organize my software? Subroutines, for one, which is an invention of the 50's. And since ILE, we also have procedures and local variables (in general, we make abstractions and use the concept of information hiding to organise software in manageable parts). This was already introduced in the late 60's, in languages like algol. OO is the next level of software organisation, where you package the data and the code which operates on it into one self-contained, reusable package. Of course it also helps when the language is powerful (meaning that you can express a lot in one sentence). Another disadvantage of rpg is that its proprietary. Universal languages like java or C have all kinds of advantages. Downloading some code from internet for example. For java and C there are all kinds of code you can download, if its fairly general. In my practice, on every site each generic function (like a modulus-11 check for example) is custom written. This is ridiculous. But rpg programmers tend to suffer heavily from the not-invented-here syndrome.
The examples you provide are all your own biased view. No standards in the java world? What do you mean? Ever heard of J2EE for example? You can call it a standard. A complete business framework. On most rpg sites its a mess. And its difficulut to compare java with rpg. With rpg you need other standards than with java. For example, a tag in rpg can be only six characters long anf they are global although you can only jump to a tag within the same subroutine. So why are tag names global and not local to a subroutine, must be about one day work to fix this for the one guy still working on the rpg compiler. Anyway, due to this, we need a standard naming scheme for tags, for example the first five letters of the subroutine and then a number. This is a typical rpg standard. Standards needed just to get around the limits of rpg, like short names. Because if you have only six chars then you better have a standard in place to make things a bit understandable. There are many very qualified, professional java programmers using standards. Of course there are also a lot of young java programmers with not much experience. But they will become experienced, just as those old rpg programmers had to learn. You can hardly use this as an argument. Besides, programming java using j2ee or other frameworks is difficult to compare with the rpg environment. And i think an rpg programmer is not qualified enough to make a good comparison. For example, typical is to say that you need less code in java to do db access than in rpg. Well, first of all, number of lines is generally not that interesting. Like the number of keystrokes it takes to program something. It's about maintainability, readability etc etc. But alas, for the sake of argument. If you build from scratch a program in rpg to update one record, its simple and straightforward taking 10 lines of code (i dont count, just for the argument). Do the same in java (from scratch) takes maybe 100 lines of code, because you have to do more, like loading a JDBC driver, creating the connection etc. But, as we all now, we as professional programmers dont deal with systems existing of a little program updating just one record. If this was the case, then programming would be a non-issue. Typing statements in the right order to do something is what even a little child can do. The real problem is dealing with systems consisting of 100.000++ lines of code, doing all kinds of things. Like i said earlier, being a professional means you have to look at the costs in the long term. In rpg, if i have 10 programs and each program updates one record, i have to use that same 10 lines over and over again in each program, which makes 100 lines of code. Worse, because each program has a direct reference to the db fields, when the length of a field changes for example, i have to recompile each program again. In general, if the record def changes, it has a direct impact on all programs using it. Now consider java. I suspect that many rpg programmers starting in java just do the same old trick but in java. Thus, in each program, coding a lot of code just to update that record. And making a mess. And believe me, with a language like java, you can make a big spaghetti-like mess if you dont know what you're doing. In this case you're better off with rpg, where you have simply one source, with all code in it, compiling to exactly one program. But we are professionals of course. And therefore we can make use of the possibilities of an OO language like java. So we create one class (object-type), for example File. In this class we code the necessary plumbing like loading the driver, making the connection etc. And we do this only once. Then we have another class (no, not necessarily a subclass of File, this depends on design, dont overuse subclassing), called Customer, in which we code all the db access functionality like updating the address or whatever, or a function to check the credibility, etc. And we do this only once. Then in all 100+ programs which do something with customer, we use the Customer class. And each time it costs only one line of code, which is also very readable. And if something changes to the record def it only impacts this one class (which maps to one sourcefile). You are more in control and you can more easily manage this change, because of a concept like indirection. The record def is only known to the customer class, not all 100+ programs. So adding a field only impacts this one class. All other programs dont see this change (btw, in java we dont have programs, just classes, but alas). So if you add a zipcode for example, you just have to change the customer class. And you can manage the transition of the rest of the software (the parts that really need to do something with this new zipcode) much more easily, and controllable, because there is no direct impact. Of course, this all depends on design, on how the software is organised. In rpg there are not many choices. Its mainly a matter of entering the statements in right order. Not much thinking about design and how you can make the software managable, flexible, maintainable, extensible, etc. Of course, in java for example, you can also make a complete mess out of it. I think the main problem with rpg programmers is that they dont tend to think in abstractions, which is key in software development. The learned programming using only rpg, which before ILE only supported subroutines, which is a very simple form of code abstraction. They dont understand more abstract things, only what is tangible, like program speed (a program which runs twice as fast is better). Or the time it takes to implement a function. Like making a copy of an existing program, changing a bit and voila, there we are. They dont make choices. For example that it's maybe better in the long run to take two days for that function, and extract code from the program into a third to eliminate duplicate code. They just do it. And dont tell me about deadlines etc. A manager is just used to it that some function takes one day to implement, and changing a fieldlength in a record takes multiple days. Or another simple change takes a week. Another advantage of software design is predictability. In the rpg world, sometimes simple things (from the user's perspective) like changing a field length takes a long time, while other more complex thing take much shorter time. By using modern software techniques like OO then the time it takes to implement something is far more on par with what a user expects. Changing a fieldlength is simple, yes, but it impacts on a lot of programs. Changing a simple function takes a lot of time, because its implemented in a lot of programs which have all to be found and changed, etc.
You get the picture mike? Of course you can do amazing things with rpg. You also can do amazing things with java, basic, c, pascal, magic, as/set or whatever we have these days. But its not about doing amazing things. Its about learning and using some simple programming principles, being a professional, picking the right tools, etc. And if i have to pick the right tools for the job, in my case developing business software, i would pick linux (which is cheap and reliable, has a lot of functionality and broad support), eclipse and java. Of course you also need a db etc etc, but you get my drift. I would not pick the as/400 and rpg. I mean, why?. This forum for example is driven by an as/400. Thats why i sometimes see strange letters instead of the ' character. Why? Dont know, but i suspect it has to do with wrong ebcdic translation possibly due to wrong configuration of the webserver. And its probably fixable. But this is such an example. As/400 is an ebcdic machine, and the rest of the world works with ascii and unicode.
And while i'm at it, green-screen vs gui, yes here we go again. Everyone seems to misunderstand this one. Yes, for data-entry and so, green-screen is better than gui. No, its misformulated. Simple keyboard entry is better then using the mouse for data-entry. Gui is not about mouse and graphics and fancy looks. You can make a character-based keyboard-only application using a gui if you want. You just have more choices with a gui. Of course, transforming a green-screen app to html or some other gui is bull. It gives you the same app, it just looks nicer and you have to use the mouse etc. This doesn't add anything. But this whole discussion is also about ignorance. I dont like the green-screen stuff. Its inflexible, hard to program, using obscure dds keywords etc. If you make a gui for a data-entry app and it isnt productive then blame the guy who built it. Not the underlying technology. Using gui technology you can make every interface you want, and like with java, you can make bad ones too easily if you DONT KNOW WHAT YOU'RE DOING. Things like webfacing dont add anything, they just transform existing green-screens to html, creating a nicer picture. And users then complain its not productive. Then the rpg programmers says "well, you wanted a gui, so now you have one". A little more respect please for the user, who is right, they want a gui because its more productive, but only when its done right of course, suitable for the app at hand. You can make a productive gui for data-entry. But of course, it would act like a green-screen app, using only keystrokes to navigate, not the mouse. You can even use green text in a 80x24 grid with a black background if you want. But with dds you only have one option, using gui technlogy you can do everything you want, including making a mess with it (like i said earlier with java). And no, its not more work with a gui to make a "green-screen" interface. You van make a framework for example in java to build a pure data-entry gui fast(which costs some time to build yes, or download a open-source one, but there is none because all you rpg guys are so full of yourselves). The possibilities are endless. Gui is not about graphics and mouse, like i said earlier. Its a completely different user-interface paradigm. Its about event-driven programming, where the user is in control, not the application. I dont want to elaborate on this. Just educate yourself, thats my advice. Be open-minded, and.... MOVE ON IF YOU CAN!!!


Btw, i dont say that java etc programmers are smarter than rpg guys. It has nothing to do with smartness, its about education and experience (and wrong practices are often hard to unlearn)

Posted by: john at December 29, 2005 6:05 AM

Programming used to be a good and easy way to find a job and earn a good living. Labor costs were not an issue. Large programming departments and consultants were common.

Today, smart managers ask themselves ... "How can we get by without the coders? If we need some, how can we succeed with only one or two?"

If you put an analyst and a coder on a problem, you are putting two people on a one person job.

If you put five coders on a problem individually, you have five distinctly different solutions. In other departments except IT, this end result would be evidence of a low quality environment.

If you put 5 (o even 2) coders on a one coder problem, you have a power struggle.

Rusty, dusty RPG used to be a good and productive language. It still is if you have limited objectives. If you need a modern looking and functional application that supports users as THEY want to be supported, RPG programmers have to use an alphabet soup's worth of technology to even approach the user's goal or talk down the user's expectations. What a waste of money!!!

Making standardized technology work with as few labor inputs as possible is good business. Business programming at the individual company level is dying. Programmer farms at the vendor level or in Bangalore or both is the future. Joe Coder should save his money or go to school, because rants about, for example, modern free form rpg being the ultimate just underscore how important it is to clean house.

Posted by: Bill5000 at December 29, 2005 6:45 AM

Move on if you can... is a slogan I have introduced on this blog site.
Why should I move on? The endless debate about the pro and cons of RPG, JAVA, Visual Studio and so many other development environments probably do not convince anybody to move away from RPG. Each language has his drawbacks, and anything else that's not RPG, belongs to toyland, right or wrong? But if you don't want to be moved by force - a team suddenly dropped in the IT room saying we gonna move this lot to SAP -, convince yourself by looking at job sites and see how many job openings are available for each specialty. You'd be surprised how RPG demand stacks up (not) against the others.

MOVE ON IF YOU CAN.

Posted by: Ugeerts at December 29, 2005 9:56 AM

RE: Move on if you can....

This is an iSeries, AS/400 centric site, ifn you have a problem with that maybe you should....MOVE ON IF YOU CAN...

On the other hand, ifn you can enrich any of us who make a living off of this machine, by all means please don't MOVE ON EVEN IF YOU CAN.

Posted by: ctibodoe at December 29, 2005 1:51 PM

By all means this site is promoted by ibm as a part of their turnaround campaign for the iSeries. Maybe they hoped clients of theirs would tell euphoric stories of the new box they just bought and how much good it did to their users and their business. Or maybe they hoped RPG programmers would fall over each other to praise ibm for the lastest %bifs and free format syntax. But in these times of turnmoil, the speculation of the direction that iSeries will take is about the only exciting part imho.

Posted by: Ugeerts at December 29, 2005 2:47 PM

Turmoil??? Katrina is turmoil. I'm just curious, why is it so important to you to knock anything positive on this site? What's the payoff for you? You probably would have better results posting this stuff on sites centric to something else, where you might snag a frustrated rpg developer looking for something else, instead of tweaking working folks looking for solutions.

But anyway, thanks for your insight.

Posted by: ctibodoe at December 29, 2005 3:02 PM

This argument has been going on since 1983 (if I remember correctly).

The iSeries is the best of all worlds and is getting better. A lot of investment in legacy RPG code is saved while allowing clever boys and girls to code in something more exotic to satisfy their creative leanings.

It really isn't an "either/or" situation.

Posted by: Jinx at December 29, 2005 3:44 PM

Everybody working on as/400 is concerned with it's future, unless he's a year or two away from retirement. Anyone denying that is dillusioning himself. We've seen declining sales year over year since '98, we've seen ibm rebranding the system to iseries and now to i5 and for the first time, sales in 2005 are rising. But then there are naysayers who question wether the raise will hold on next year. The iseries/i5 is certainly in a transformation stage, and what is certain is green screen and drab 132-column reports have no future. But into what will the i5 transform into? To something other platforms have already for years? Websphere versus ASP/PHP? Java and linux on a high priced i5 versus a cheap intel box? BTW, does java en linux run more stable on i5? I don't think so. These newly sold i5 systems, are they 2000/0 cpw (0 interactive) boxes running SAP or Domino or are they still deployed for classic rpg development who need interactive 5250?

In what kind of system should the i5 transform into is a question I'd like to see answered. Opinions welcome.

Posted by: ugeerts at December 29, 2005 4:33 PM

I want some of whatever's John's on.

rd

Posted by: Ralph Daugherty at December 29, 2005 6:39 PM

In 2004, it was reported that iSeries sales fell by 25% in one quarter. In the same quarter of 2005, it was announced that iSeries sales rose by 25%.

This was hailed as a recovery. Actually, if you do the math, it is still a decline.

If you take 25% away from 100, you end up with 75. If you increase 75 by 25%, you end up with 93.75%. This is a 6% decline.

I suppose a slow lingering decline is better than a fast fall. But down is still down.

Posted by: bill at December 30, 2005 5:50 AM

Given what IBM is pushing, sales of iseries have little to do with OS/400 and RPG even when there are sales.

There's a broad mix of technologies, they all have their place, and the AS/400 supports them all well and in OS/400 and RPG provides unique solut