Reader Feedback on EGL: At Least It’s Not Java, But It Ain’t RPG, Either
October 1, 2007 Timothy Prickett Morgan
Well, as you might have expected, Brian Kelly has once again struck a chord with some of the RPG faithful. Several readers who are unhappy about IBM‘s plan to use its Enterprise Generation Language (EGL) as a bridge between legacy RPG applications on i5/OS and OS/400 and the Web took the time to respond, at length, to Kelly’s article that appeared in last week’s issue.
And, if the past is an indicator, then it is also reasonable to assume that Big Blue was none too happy about his criticism of the plan to promote the use of EGL as a strategic development tool at i5/OS and OS/400 shops, too.
You can’t win, you can’t even break even, and you can’t quit. That was how a wise man I know once explained what it is like dealing with tough in-laws. (It was my father-in-law, and he said it sarcastically, and with a big, big grin on his face.) Same situation applies here, as far as I can see, which is why I ran the original EGL story, Kelly’s response, and now this feedback. I would much rather be writing a different kind of a story, one that makes readers happy. I believe in change (evolution, to be more precise), and I believe in possibilities (not probabilities, because people can surprise you), and that is why I work hard on my newsletters every week and try to make sure IT vendors and my readers get the best of my thinking and that of other players in the IT community.
Bringing legacy applications to the Web was hard enough a decade ago, but the desire to take monolithic applications–or even those that have been modularized by the more modern RPG IV compiler–and extend them out as Web services using a services-oriented architecture is a bit trickier than scraping a screen. There are lots of ways of skinning this cat, and just because IBM moves in one direction does not mean that i5/OS and OS/400 shops will follow. Java is a prime example of this.
How well or poorly EGL is received by the i5/OS and OS/400 faithful depends in large measure on the hooks IBM puts in for RPG applications, how expensive it is for developers to get the Rational tools to make use of EGL, and how tough it is to use EGL compared to myriad other ways of taking data and screens inside RPG applications and transforming them into Web applications that can mix and match data from other platforms, if necessary.
But right now, if the feedback we have received is any indicator, IBM has an uphill battle to even get people to hear what EGL is all about and give it a test run. That can, of course, change. There are, for instance, hybrid RPG and Java shops, just as there are hybrid RPG and .NET shops. There can be hybrid RPG and EGL shops that deploy generated Java code and RPG code. A lot depends on how IBM implements this EGL tool and the cost of using it. Lower the economic barriers, and it is surprising how popular something can become.
In any event, here is a sample of some of the feedback IT Jungle received. Some of the comments could not be printed in a family newsletter. Tempers run hot on the issue of native Web programming for RPG applications, and adding EGL to the mix did nothing to cool the issue down.
Couldn’t have said it better if I had written it myself.
It’s been 13 years since Java was introduced on the AS/400 — May 3, 1994.
That’s a piece of trivia that makes the fact that IBM has not helped bring System i shops any closer to being part of the Web in all of those 13 and a half years a lot more than a trivial mistake.
It’s a killer and that is a shame.
Oh man, the whole time I was reading your article I felt almost as if I was writing it. The article reflected my feelings quite well. I am tired of an excellent language with second to none DB access getting closet treatment whenever guests come over.
I am getting sick of IBM’s platform independence Java bell ringing. IBM would do so much better, IMO, if they would simply continue to lock people into the i5 by enhancing RPG! I love the fact that I am locked into this machine with RPG, because it is stable and meant for the enterprise.
Anyway, again, good article. Loved it.
Thank you for taking the time to write this response.
After 13 years of cross platform Java, one would think that IBM’s System i decision makers would bring one home.
You did write the article, as did many others of us.
I agree. I don’t really know much about a mainframe, but when I looked at the EGL description, I did not think I would ever use it.
I am still waiting for RPG to have a native browser interface like it does with the green screen. Until then, it is still an AS/400 green screen application development machine and the Web is consigned to the Intel folks.
I decided to call the AS/400 its current name, what ever that might be, when it really changes for the user. When it goes graphical, then we can call it a System i.
Somewhere the vision of the founders of the System/38 has been lost. That is why IBM has such a hard time making decisions on what to do next. If they took any of their board members and put them in a company where they not only were on the board of directors, but they also had to build the computer applications for the company, and support all those that were using the programs, then perhaps they would begin to get the message.
I am sorry that I could not say what I said in more positive tones.
My article had actually started as a response to TPM’s article because I was agitated at the pompousness of this whole EGL thing. I probably should have held it a few days and toned it down.
IBM is really upset with me about the article. I’ve heard from some of the people who had been keeping me up to date and overall they think I was unfair. I am not sure yet if I was fair or unfair. I was definitely upset.
I have read your article this morning. Your articles are great and I love them because you are the spirit of System i.
As I understand, we have now two IBM’s organizations–one for the SMB market and the other for biggest companies–that means two types of software. In fact, you know that much better than me. 🙂
RPG IV doesn’t need new instructions to go through the Web because RPG programs are MVC. You just need to change the DSPF or add another communication file into *pgm object at the compile time. My point of view is that IBM has no other solution than to add Web services.
Imagine that: At the compile time, you will have somewhere a parameter WEBSEV(*YES) and that’s all!!! You will have Web Services integrated without doing anything!
After that you just need an application server dedicated to Web services and a middleware program for the SMB market with some frameworks to design the Web 2.0 presentations.
I am quite sure we will have that with the V6R1.
I think that the great newsletters at IT Jungle live from ISV that say for years that System i is not modern because they are selling solutions to “modernize.” That’s good and bad for the platform, because managers don’t want to invest even one dollar on a non-modern platform and meanwhile they thought that IBM is “i” like incompetent. If IBM Rochester would announce that we will have SOA integrated, all the editors would write articles for months saying that integrated SOA is a bad solution.
I think that Rochester’s integrated solutions for SMB are going to compete with EGL and other monster for big editors and big companies. EGL runs with WDSc and only 5 percent or so of System i clients are using WDSc. That means another 5 percent out of 5 percent will use EGL.
The day Rochester will integrate SOA to the i5/OS, SOA will become natural and we will have a much better solution than Microsoft client/server. In other platforms, you will need expansive software and an engineer to manage SOA. That will be the third big integration and we will have natural UI.
Tell me Brian, do you think that I am naïve or mad?
Thanks again for your great articles.
Thank you for your response. No, I do not think you are mad. I fear that IBM is so concerned about its solutions being cross platform that they are reluctant to invest in a System i solution. I think you are right on the money. I also think that IBM could sell this notion to the world.
IBM protected RPG from the PC. It was to have announced OS/2 RPG in the mid 1980s, but there was too much fear that the PC with RPG would eat away IBM’s System/36 revenue. Instead, by making RPG a one-platform language, ISVs are moving their wares to Windows where there are more potential clients.
I don’t ever think it is too late to invest and market something new.
It seems to me that there is a lot of risk for an IBM executive to choose to push an SOA/Web-MVC enabled RPG language through all the channels necessary to have IBM bless it. So, we have an IBM that won’t do it for System i because they think there are not enough potential clients (I disagree) and they won’t make RPG cross platform because it would cost too much. RPG is IBM’s only proprietary language besides EGL and EGL is not going to do it for System i shops. You are right–only 5 percent of those that use WDSc. IBM counts differently, perhaps.
But, there is nobody in IBM to do what the System i community needs and what the other platforms would die for. Keeping RPG on System i alone isolates it and unless IBM thought it could take over the world with millions of System i’s, the solution to a pervasive RPG is to build compilers on other platforms. If puny little California Software could do it, why can’t the mighty IBM? Then rename it “The Business Language” and IBM will make the evening news.
Nice article about EGL. Our shop is looking to go down this road sometime in the future. I will forward this link to my boss as an FYI.
Hey, as I was reading this, I had a thought. Why not forget about IBM making RPG Web-enabled. Wouldn’t it be better if some company outside of IBM did that and integrated it into RPG by defining a Web session as a ‘file’ with the SPECIAL device designation?
I remember in the mid-1980s, when I was working as a programmer on a System/36, we bought a product called ACCESS/36, which allowed the S/36 to upload and download text files to and from laptop PCs over a dial up connection. It also allowed us to create voice applications where customers could dial into a number, and interact with the S/36. This was all done using RPG II with a ‘file’ defined as a SPECIAL device. The product included subroutines you could use to have your RPG program interact with the product.
Fast forward to 2007. Couldn’t this same principle be used to allow a product to be developed that had a ‘SPECIAL’ file be used to interact with a ‘presentation’ server that would manage the Web part for RPG? You could have service programs instead of subroutines that would be in a binding directory, rather than using /COPY all over the place.
Just a thought as I read your article.
Thanks for your reply and your kind comments.
The notion of a special device is perfect for an ISV company with the skills to build that device driver for the masses. Again, IBM is the most likely candidate with the skills to interact with the compiler.
Short of that I have seen techniques in which some specialized subroutines get added to the output of the generated RPG module. The special device would be better for sure since the subroutines would be under the covers. Unfortunately, it seems that IBM is not about to revisit the natural browser interface so we do need to move on. Personally, I am learning as much as I can about PHP/MySQL and the Mambo, Drupal, and Joomla content managers as well as IAWEB, HATS, and WebFacing–and another tool that looks intriguing called JACi400.
I think shops need a quick way to take applications to the Web–even though the application itself when moved this way looks a little 5250-ish behind the css lipstick and makeup. It provides a means of getting something up tomorrow if needed — HATS, IAWEB, WebFacing, JACi400–all fit that. Then, you need a tool for new development. For this, there is PHP with the various content managers/templates and free code and for parts of that, there may also be room for EGL. Or something like that.
Of course there are a ton of third-party tools in IBM’s developer roadmap. but in some ways all the choices that exist make it more difficult to be able to select one that will be the right tool over time.
Brian and Tim,
Thanks for your courage and conviction.
I believe that EGL is doomed, along with many System i customers and programmers (and ISVs).
You know Tim and Brian, I am beginning to feel much better now that I have given up all hope!
EGL smeegee-el. I am so sick of IBM not listening to its customer base I could just puke. For all the loyal iSeries customers out there, this is a kick in the groin. In a few years, there will be no need to continue to “develop” RPG. The green screen is way long in the tooth and looks old and tired. Apparently IBM doesn’t even want to give it the Web face lift it desperately needs (and deserves).
I have somebody at IBM today telling me the iPTF Website was never “designed to allow PTFs to be ordered on CD when I have been doing it for years (there is even a check box to do so, but the site has been changed in such a manner that this no longer results in a CD order). The thousands of dollars we spend on Software Subscription apparently isn’t enough to cover the cost of four CDs and the labor to cut them! Just another example of IBM’s ignorance about its own customer base. The only explanation is that IBM wants to run off this business as fast as possible. To that end, things are coming along nicely. That would explain the EGL phenomenon.
Most companies would die for the feedback IBM gets from its customers about what they will buy and not buy.
Why IBM fights the feedback and goes against the grain of customer requests in critical decision after critical decision will be an interesting finding a few years from now when the System i gets its chance for an independent autopsy.
Your article could well be the last straw for me. I have tried to keep an open, objective attitude lest it sour my day-to-day business selling and supporting AS/400 software. Hearing these depressing reports over and over is taking its toll.
I have suffered the slings and arrows of an arrogant IBM for the last 17 years. The brand NON-differentiation of the AS/400, the constant name changes, the push to Java that runs terribly and has no real hooks to RPG, it is really all getting to me. As the staunchest advocate of the AS/400 since I first met one in 1991 and fell in love, IBM has done NOTHING but let me down, and make my life as a DEDICATED AS/400 software vendor HELL. Now, they have concocted a whole new future for us.
Last week, a (formerly) large user group in the Mid-Atlantic area told me they will not have their annual expo next year. They just can’t get the interest or attendance. As a software vendor, we are seeing attrition of the ranks to the highest level ever. IT departments just can’t make the arguments to KEEP the AS/400–maybe partially because knows what the heck it is anymore.
No, we will not be re-writing our applications in EGL. We will not go over to the dark side and start supporting Windoze. We will stand proud as a dedicated resource for those of us who KNOW what we have, appreciate it, and built our professional careers around it. No matter what IBM names it tomorrow.
We will be here until the last RPG shop closes the doors. Damnit.