Ought To Be Committed
June 19, 2017 Timothy Prickett Morgan
Reassurance is a paradoxical phenomenon. If we are secure about something, we don’t look to be reassured about someone’s commitment to that thing; if we have our doubts, reassurance can help calm our nerves and soothe our fears, but ironically, the very act of reassurance can, in a way, undermine our confidence even as it helps reinforce it.
Funny, isn’t it?
It is with this irony in mind that we ponder some recent statements by IBM concerning its commitment to the RPG and COBOL compilers and Rational for i development tools for the IBM i platform. To be specific, back on June 9, Steve Sibley, who used to be director of Power Systems hardware and who has an analogous job in the new Cognitive Systems business at Big Blue, put out a letter to customers to reassure them of IBM’s ongoing commitment to application development tools on IBM i.
Here is the letter, which we found out about thanks to Google Alerts picking it up as it was republished by Data3, a European business partner in the IBM i community. We are the last ones to cast any stones when it comes to typos, but one of them in this short document gave us a bit of a chuckle.
Dear IBM i clients and partners, I would like to take this opportunity to reaffirm the commitment that IBM has made with regard to RPG, COBOL, and the IBM i Development suite of tolls. Recently the team responsible for developing the compilers and tools for IBM i have moved into the IBM i Development organization. This will allow more synergy between the development of the languages and the development of the operating system. IBM Power Systems is committed to the success and ongoing development of IBM i. This includes the ILE RPOG and ILE COBOL compilers and associated development tools. This ongoing commitment has been illustrated with such enhancements as RPG Open Access and Free Format RPG. Each has the ability to interface traditional RPG programs to other devices and technologies and the natural evolution of TPG to free format are significant enhancements supporting client requirements for modernization of their applications. In parallel with our most recent IBM i 7.3 release, new extensions to the RPG and COBOL languages were also announced. We gather requirements for these Application Development environments for IBM i from a number of customer advisory councils and we balance those requirements with our study of industry trends. As the delivery of updates to the compilers and Rational Developer for i unfolds over the next several quarters you will see further evidence of our eagerness to listen, to collaborate, and to deliver the improvements you consider most valuable. IBM Power is committed to your success – to helping you realize business value from your applications and to extend that value with new business opportunities. We thank you for your business and your confidence in the future of one of the world’s most durable and productive platforms for business. Sincerely. Steve
We certainly agree that this is one of the most durable and productive platforms for business in the history of corporate computing.
We are not so sure about application development tolls, but we are not going to read too much into these tools. The newsy bit, which is interesting to us, is that Sibley is saying to the IBM i base that the application tools, which have been handled through IBM’s Toronto software labs for as long as I can remember, have been pulled back into the core IBM i group that is centered at the Rochester labs where the System/3 and its progeny (System/38, System/36, System/34, AS/400, AS/400e, iSeries, System i and IBM i on Power Systems) have long been centered. Sibley did not explain why this was the case, but in general, over the past several years, the middleware associated with the IBM i platform has been pulled from various parts of the former Software Group (which has been disbanded) and pulled into the Power Systems division, where the hardware systems and operating systems that run on them are now managed as a whole. It is difficult to do co-design on a tightly integrated system with people reporting to different groups and aligning with different sets of priorities, and given that one of the key values of the IBM i platform is its completeness and integration of components – that is what the i in IBM i stands for, after all – it makes sense to have the whole stack coming out of the same organization.
There is no question that there is, and has always been, a collaborative relationship between the IBM midrange customers and those who create and maintain the application development tools that are bundled with IBM i. And the many application tool providers who get out there earlier than IBM and take more risks than IBM when it comes to new technologies are also not just collaborating with customers, but are watching the broader market and seeing how other advanced organizations code their applications and backend systems and try to bring these modern approaches to the platform. IBM embraced Java and free form programming, for instance, and brought them to the platform, but others have been ahead of Big Blue on PHP, Ruby, Node.js, and other key elements. It takes a lot of players to make a rich and diverse ecosystem. There is a pretty fat line between being on the bleeding edge and being behind the leading edge, and it is somewhere between 18 and 24 months as best as I can figure. This is about the right pace that IBM i shops can see a new technology, watch it mature, and then see how it is initially integrated into the platform. I have no idea what will come next, but I have no doubt that other tools and other approaches to application development and modernization will be absorbed and integrated into the IBM i platform. This is how it has been for decades, so there is much precedence.
We think it is a good thing that the compilers and tools are being reabsorbed into the operating system development effort, but we wonder if this is masking some sort of cut in investment. That IBM continues to invest, we have no doubt of that. But the fact of investment is not as important as the quantity of that investment, and we know that IBM is looking to cut costs everywhere it can, particularly in the IBM Systems group. Moving the tools back into the stack might be a way to make Power Systems – now Cognitive Systems – look more profitable while also shortening loops and speeding up cycles between the different development teams (including those doing hardware, operating system, database, middleware, and compiler technology). We hope that IBM is not cutting back on the level of investments it is making, but at the same time, there are others who will pick up the pace if it does.
Maybe what we need is open source RPG and COBOL compilers in the GNU compiler stack, maybe more than we need Windows on the Power hardware or any of the other many things that we think we might need to bolster IBM i over the longer haul. IBM has generally frowned upon anyone doing compiler and runtime emulation of its proprietary mainframe and midrange platforms, and it has resorted to lawsuits or threats of them to keep companies from doing it. We are not sure of the legal ramifications of someone making such compilers. (Infinite Software does something to port RPG and COBOL software over, but we are not certain of what techniques it is using these days.) For the sake of helping grow the pool of knowledgeable RPG and COBOL programmers with a twist for the IBM i platform, I think such open source compilers should not only be available, but they should come from Big Blue.
If IBM really wants to speed up how innovation progresses with these programming languages and get the input and support of the community that uses them, I think this is possibly even better than the loosely coupled loops between IBM and the community that has driven compiler development to date. If IBM does do this, by the way, it will be a sign that it is ready to truly open up the IBM i platform – something we have advocated for a long time. But we recognize that such a strategy might only be used against IBM, with third parties creating clone IBM i platforms and pulling the base off Power, weakening the foundations on which the continued operation of this business rests. But that is the nature of a truly open collaboration. IBM can choose to have a larger and more vibrant number of customers who love the platform, or a smaller set who begrudge the heavy investments they are making and who come under heat from those who promote Windows Server and Linux platforms as cheaper alternatives. Either way it will make the same money, but the way people feel about how they are spending the money – and what they get for it – is strikingly different.
We would love to know precisely what IBM’s commitment is, in terms of development resources and roadmaps, because with somewhere north of 120,000 customers still on the platform, it ought to be committed. Time and again, IBM has squandered vast sums trying to chase adjacent markets without taking care of its core customers, and that is why it is in the mess it is in today.