I believe the i5 is the All-Everything Machine that I highlight in my book of the same name. My comments on ISVs and my suggestions for improved i5 Web capabilities and better code porting options in RPG and COBOL are real needs. Yet, they do not diminish the fact that the i5 and its predecessors back to System/38 are still the best systems ever designed. However, in addition to its renewed marketing emphasis, there are a few important things that IBM must do better.
Until I attended COMMON, I felt like a voice crying out in the desert. But, nobody who could help was listening. I felt the need to cry out to do my part to help make my favorite system relevant again. IBM seemed to have forgotten how to highlight the i5 or felt it did not have to because its constituency had done that for the company. Over the past ten years, IBM's AS/400 marketplace was mostly its installed accounts. After COMMON, I get a sense that this has changed.
I saw the eyes of the IBM executives in the Town Hall meeting and afterwards at the COMMON CUDS. I am convinced that this new group of executives have real skin in the game and that they are going to work to bring back the AS/400 (i5) as a key IBM product. I am encouraged again and hope not to be disappointed. From listening to IBM speakers at COMMON and from a few side conversations, many of the things that I wrote about in despair several weeks ago before COMMON may already be in the process of being fixed.
Some may say it is too late. I say it is not too late but it must be done right. It would have been better sooner but if IBM does what it is projecting in hardware, the eServer and the i5 will be very relevant and if IBM does what it said it will do in software (RPG and COBOL), programmers will be immediately in the Web game. IBM will be able to leverage the interactive programming skills of its business application developers in ISV shops and in-house shops.
How about re-enfranchising a few million i5 business application developers and in one fell swoop? Unless it is a dream, IBM is heading to make its RPG and COBOL programmers Web developers with minimal change to their coding paradigm. I am counting on what I heard at COMMON as being the near future, not just a dream.
I do not really want to risk disappointment again. But, I really do believe. I believe that the i5 will be successful again because it uses i5/OS and RPG and COBOL, just as its developers do. I believe that Mark Shearer and Peter Bingaman and George Farr mean what they say. I saw it in their eyes. I see a return to glory in the offing. If it is just a dream, then it is a good dream. The more of us who keep the dream and who believe, the more strength IBM's new i5 executives will have in delivering that dream to our shops. I am encouraged again. But, this time, I expect results.
Let us know what you think about the story; e-mail: Matt Stansberry, News Editor
Brian Kelly retired as a 30-year IBM Midrange Systems Engineer in 1999, having cut his eye teeth in 1969 on the System/3 and later with CCP. While with IBM, he was also a Certified Instructor and a Mid-Atlantic Area Designated Specialist. He has written twenty-three books and numerous magazine articles about current IT topics. His new book was just released titled, The All-Everything Machine.
Recently, Search400 published Brian Kelly's column Tough Love. What do readers have to say about Kelly's views? Here are the responses.
Why should the RPG community have to learn another language?
I summarized this article into two sections. First deals with the Operating System of 5250 vs Internet (Webshphere). Second deals with language, RPG vs. Java.
On the first section, it is interesting to take note that IBM already had a platform that behaves a lot like the Internet; it is called CICS and runs on the mainframes using 3270 as apposed to 5250. The 3270 concept is a Cobol or RPG program is called and it puts the green screen page and the program goes away, just like a java program does on the internet.
On the second section, IBM should have enhanced RPG to interact with HTML just like it does with DDS. Java in its most simplistic form is just another language. Why should the RPG community have to learn another language? Thank goodness for companies like Profound Logic, which developed RPGsp and is what IBM should of done with its native RPG. All I had to teach my RPG programmers was Internet stuff like HTML, XML, etc. and we are coding all of our applications using RPGsp running in Webshphere.
Systems & Programming Supervisor
Collin County Government
McKinney, Texas 75069
The good fight
It is nice to finally read it!
The user interface is the reason the as/400, I mean iSeries, I mean i5 (lotsa names, same crappy user interface) continues to slide into oblivion.
Keep fighting the good fight.
No cost effective solution
I have just read Brian Kelly's story and agree completely. I am a consultant that has been working on the IBM midrange since 1981. Most of my customers are small companies that have been on the IBM midrange for years. I am getting clobbered in the small company arena because I can't recommend cost effective solutions to my customers when they begin to look at Windows solutions.
Many of the issues that Brian raises are exactly what I run into. I have customers that have been using back office solutions for years and then throw away all of that code and go with something they have no experience with because I have no inexpensive way to convert the user interface to a more modern "feel". I know there are ways to get there but the process is too costly for a small company to approve.
I have also been in several situations where we have tried to put a Windows application on the i5 through the attached xSeries box only to have the software vendor tell the customer that they won't support that hardware arrangement which effectively kills the customer's interest in that solution.
The consultant in the field cannot fight this battle alone even though I believe that this is exactly what IBM wants us to do. If IBM does not give us an easy and inexpensive way to integrate our existing applications to the internet and do a better job of selling the i5 advantages to vendors outside of the traditional i5 marketplace then I think we will continue to see the i5 cusotmer base nibbled away until it can no longer sustain itself.
Two specific problems that I see is that:
1.) The IBM solutions usually try to be all things to all people right from the start. This results in the solution being too complex and costly for most customers to implement. IBM should take a page from the Microsoft book and build a limited but cost-effective solution that will get a customers attention and then add more function as the product evolves.
2.) Most i5 software solutions are too costly. Even though a customer may be desperate to stay on the i5 hardware because of reliability, they cannot afford the newer software with all the bells and whistles. The result is that they stay with their "proven" applications for years until one day they wake up and complain that it can't do what the newer applications can do. Then when they come to us to "fix" the application we have no quick and cost effective way of modernizing their "proven" code. IBM says use Websphere and Java (neither of which are quick and cost-effective) or Webfacing (which is still a major investment to transform an entire application).
All too often we have no i5 cost-effective solution to present to the customer.
We are all hosed
I have seldom read anything so depressing but true.
Either us old RPG developers don't know anything about the real world. Or IBM milked the cash cow so long and did nothing that now it is too late.
Either way, it looks like in the long run we are all hosed. He is so right. IBM is not providing the tools the existing community needs.
Where is the RPG solution for the web that includes some reasonable opcode changes and the maintenance of state? Is there any path that the average developer can understand without needing to go back to school?
As an old dog, I will ride out the iseries to the end. For my customers it is by far the least cost solution at the present.
For the future, I have no idea.
From the iSeries war-horse's mouth
Thanks for the great Tough Love article in Search400.com.
I purchased Brian Kelly's book The All-Everything Machine from McPress, and read it cover to cover. It is a wonderful book and it sheds a unique and piercing insight into the past and probable future of not only of All-Everything Machine, but of IBM.
Brian's Wish List topics in your article for The All-Everyining Machine mirrors my experience and my own wish list, but in an articulate and detailed, and in a published way that no else has dared to do.
I believe the All-Everything Machine computer is the result of the brilliance of Dr. Soltis, and his leadership of thousands of talented IBMers in Rochester to achieve his vision.
The IBM language compilers and products such as WebSphere, WebFacing, HATS, etc. of the Toronto labs are, in my opinion as an ex-IBM Senior Systems Engineer and holder of three software patents, really awful and are what is holding back the potential of the All-Everything Machine.
Apparently everybody but IBM can produce the software that IBM must push to survive.
IBM software simply doesn't get it, and will not do the obvious, as Brian so well put it in his new Book and in your article.
My opinions, and Brian's opinions are shared by virtually every thoughtful, and IBM loyal, person that I know, and that is hundreds of talented and experienced people.
Thanks for having the courage to print Brian's article.
Paul H. Harkins