What Is Driving Application Modernization?
August 7, 2017 Timothy Prickett Morgan
What you don’t measure you cannot manage, the old adage says. And the key vendors in the IBM i community, ever chasing that next opportunity, do a pretty good job querying the installed base of customers about what they are up to and why. A better job than Big Blue, which should be the biggest beneficiary of the 125,000-strong IBM i base, if you want to be truthful about it.
We are always on the hunt for any insight into what is going on out there in IBM i Land, and Rocket Software, a provider of application modernization and software change management tools, has commissioned a survey of the IBM i base to get a sense of what is going on with green-screen modernization projects. The results of the survey are published in a report called How Companies Are Approaching Green-Screen Modernization, which you can download at this link. But to get a broader sense of what is going on, we had a chat with Dan Magid, vice president of applications development and DevOps and chief technologist for IBM i solutions at Rocket Software.
As Magid tells The Four Hundred, some of the answers to the questions that Rocket Software posed to IBM i shops are about what you expect, and other responses are a bit perplexing. The survey was performed in March and April of this year by Gatepoint Research on behalf of Rocket Software, and 87 companies participated. Judge for yourself if this was a representative sample or not. All respondents were volunteers – none were solicited using telemarketing, which probably doesn’t work that well these days anyway – from a wide variety of industries, and the distribution of CEO, CFO, and CIO executives taking the survey was low enough – around 8 percent – to be relevant. (A lot of CXOs at large companies have no idea what is going on in their datacenters, in our opinion, but those at midrange companies tend to have a very good idea since, in many cases, the company is their own.) About 15 percent of respondents were vice presidents, 52 percent were directors, and 25 percent were managers, which again does not violate our sensibilities.
The very first question in the survey was: Does your organization currently use older, character-based/green-screen applications? And weirdly, only 89 percent answered yes to this. What on earth were the other 11 percent even taking part in the survey for? In any event, the data culled from the report only includes those who said they were using green-screen apps. Of these, 83 percent of respondents said they were using IBM i or OS/400 as the operating system supporting those green-screen apps, with another 24 percent saying they ran on Windows and another 21 percent saying they ran on the z/OS operating system for IBM mainframes. Another 17 percent said these green-screen apps were on Unix of one sort or another and another 16 percent said they had green-screen apps on Linux platforms. All of these platforms have a character-based mode, and they are not all technically colored green, but you get the idea. The point is, green-screen apps are not just an IBM i thing. (There are probably some DEC VAX and HP 3000 apps still running out there, now that we think on it.)
Perhaps the best question in the survey is the one relating to the challenges that companies are facing as they integrate and extend their green-screen apps. Some 47 percent of those polled said that they would need to re-architect the entire application stack to do modernization on their applications, and 46 percent of customers said they did not have the budget to do this. Another 32 percent said that this would be disruptive, and 29 percent said that the modernization task was too complex and they did not have the skills to take on the job. Magid wants to tease the truth out of this question, which he admits might not have been worded particularly well. (This is ever the problem in doing surveys.)
“Well, 47 percent said they need to re-architect the entire system to completely integrated,” Magid explains. “Does that mean that 53 percent of the people think that they do not need to re-architect the system, or that they don’t think re-architecting is going to be a big challenge? It seems unlikely that they think it will not be a big challenge, so I am presuming is that 53 percent of the people are telling us that they do not need to re-architect the system, which to me is the right idea. We have seen so many customers that go down the road of re-architecting, and it is a very risky path fraught with all kinds of pitfalls and problems. We see more and more customers taking this in stages, and they increasingly do user-based design. They start with a user experience that they want to end up with, but there are many challenges to that green screen. First, there is the green screen itself, which is a problem, but second, it has a very structured workflow that is very tied to how the program operates, and in the Web and mobile environments, we want to work differently. So they start with the user experience they want and then they go back and figure out the architecture.”
This backward approach not only happened at customer sites, but also at myriad IBM i software developers over the past several decades (most of them have been merged and absorbed into the few remaining players, such as Infor and Oracle.) They gutted their code, often porting it from RPG to Java, and they did not substantially change it and ended up with the same look and feel, or worse, something that only ran on a fat Windows client instead of a Web browser.
Companies also make the mistake, says Magic, of spending big bucks on a set of applications from SAP, Infor, or Oracle to replace their homegrown applications, only to realize that they have woefully underestimated the amount of domain knowledge that is encapsulated in their homegrown or heavily modified applications.
“The problem that we see is that the person who says throw it all out and get me SAP or Oracle is usually the CEO,” says Magid. “And this CEO has no idea how this application came about, and oftentimes, they have not been around for the decades that this application has been under development. And on top of that, the IBM i is not a system they know much about. So these CEOs don’t actually know how much work has gone into creating an application that is finely tuned to how that company does business. We often see a disconnect, where the IBM i programmers say that this will not work, but the CEO says they have to do it.”
We hear countless stories about how, in the wake of a merger or acquisition, IBM i apps lose out to an Oracle or SAP suite, and the software actually is not as good at reflecting the actual business, but there are political, rather than technical or economic, reasons for consolidating the apps. The acquiring companies end up destroying some of the value they wanted when they did the deal in the first place, and that is because they do not understand the esoteric nature of people, business, and the applications that encode how they interact.
Application modernization is driven by a number of different factors, and those cited by end users that Rocket Software had surveyed sound familiar. About 60 percent of the shops polled said they wanted to offer easy and intuitive applications that were agnostic in terms of platform and devices. Another 52 percent said they were looking to improve operational efficiency by modernizing their apps, and another 47 percent said they wanted to reduce the cost and time to train end users, many of whom these days have no idea what a green screen is and why it is actually better in some cases. (You don’t get RSI from tabbing, but you do get it from mousing, for instance.) Another 43 percent said it would make users happy, and 14 percent said it would help curtail employee turnover; 17 percent said they needed to integrate various character-based applications on different platforms, citing 5250, 3270, and VT formats.
“I was actually a little bit surprised at how few people were looking at decreasing employee turnover through modernized applications,” says Magid calling out the results of question four in the survey. “We hear from customers that this is one of the big issues for them. Employees don’t like working on green screens, and new employees don’t want to work on them especially. They want to look like an advanced company, one that knows what it is doing technologically. This sentiment may be reflected in the much higher percentage that said they wanted to modernize to increase user satisfaction and reduce time to train and cost.”
Reducing training time and cost is, says Magid, a very big deal at all companies, and modernizing applications so they have Web and mobile interfaces is part and parcel of that. Survey respondents are therefore not surprisingly focused on exposing existing functions as Web services (57 percent) or creating Web user interfaces around IBM i applications (40 percent) or adding mobile applications that interface with IBM i applications (33 percent). Only a relatively small number – 16 percent – of the shops polled said they were replacing RPG applications with those written in other languages. Which is good because such gutting is unnecessary, disruptive, expensive, and risky.