Technologies to study beyond MPE skills
February 22, 2016
As 3000 experts have seen their jobs eliminated, and their employers focus on other platforms, they have faced a challenge. What should they study next to learn marketable skills? One obvious answer is the tools in the community for migration. Some of these open a new world of learning to 3000 veterans. Learning the tools provides an entry to get familiar with new concepts.
However obvious it has seemed to study .NET and Visual Basic, there are many shops planting outside that Windows-box. Open source software is the choice for prospects that reach farther.
Michael Anderson left the Spring, Texas school district six years ago to found his J3K Solutions consulting practice. Even then, when Linux and open source did not dominate IT plans as they now do, Anderson knew Microsoft wouldn't hold its market share.
In 2009 he suggested a good place to start learning beyond MPE were tools like ScreenJet and Marxmeier Software's Eloquence. "Ordina-Denkart's WingSpan, as well as ScreenJet, are both great products," he said. "They are both great models for software design. I have not found anything that compares to them that's within reach of small companies and independent developers." That's a statement on pricing as well as capability.
A caliber of tools like this is not yet available in the Linux/open source market, though. James Byrne, an IT manager at Harte & Lyne, says his company's "progress towards a final departure from the HP 3000 has not been as rapid as we had hoped. The main reason is the primitive nature of the tools in common use by the *nix community. These have improved greatly over the past decade, but they are still nowhere near the effectiveness of efficiency of software I used on the HP 3000 in the 1980s."
Complaints about the "Cognos Products," now owned by Unicom after a five-year adoption by IBM, have legendary status. But the gripes have been about licensing and pricing, not the subtle efficiency those advanced development tools provided. Byrne's company has been using Powerhouse and its cousins since before the products were named as such. Close to 40 years after they were introduced, the tools are still doing a better job for Byrne than open source alternatives.
The systems at Harte & Lyne are completely written in Quick, Quiz and QTP. "They were started before Quasar re-branded itself as Cognos and began calling their products Powerhouse," Byrne said. "The maintenance load of these programs presently is as close to zero as can be imagined. Nonetheless, modifications are still being made after all these years and the speed at which these are completed is absolutely astounding to people used to coding for *nux and Microsoft environments."
Byrne adds that the programs are sometimes creations from the days before the Web existed. "Sometimes we are working on programs that have not been touched since the late 1980s," Byrne said, "and the ease with which that ancient code is understood illuminates just how poorly most code in other languages conveys its meaning."
ScreenJet employs Windows to do its interface transformations, while Eloquence delivers the IMAGE foundations to Linux, HP's Unix and Windows. If tools like these two above existed in the GNU environment — what Byrne calls *nix — then HP 3000 programmers who are finding themselves out of work could study a workable toolset in the GNU environment. Their decades of experience could be tranferred more easily environments beyond Unix and Windows.
It took a village to build the world of open source choices. A very savvy village is required to open-source the expertise in Powerhouse, Eloquence, ScreenJet and others. Consultants have reported their migration work employs such tools. Now the IT teams like the one at Harte & Lyne have moved on to open source. The shift has had its issues, though.
In the late 2000s a product similar to Powerhouse was created in open source, an Informix 4GL clone called AuBit. However, by the time that project was mature enough to consider for our purposes we were already committed to Ruby on Rails (RoR). We adopted the web application approach because we wished to employ generic browsers in place of specific terminal emulation software.
But the price of using RoR has been constant change to its underlying code base with commensurate breakage in our production code. This is due to the very nature of Open Source. Projects such as RoR are staffed by what are essentially a bunch of ill-disciplined hackers. Each contributor is driven as much as by ego and anything else and thus each attempts to place their individual marks on each iteration of the software.
This often occurs at the same time as they are learning what is really needed from the software. The result is an absolute absence of backward compatibility. There is usually an overarching individual that exercises some restraint. But the the mantra tends to be: what is old is useless, new is better, change is for the best, and who cares what breaks.
Compared to the rock-like stability of the HP 3000, writing software based on RoR is akin to building a castle on quicksand: expensive, time consuming, and as likely to fall down and be swallowed as anything else. Thus we are not as far along as we expected.Things may not be quite as bad as that, but I am feeling somewhat put out with that project at the moment.
Open source emerged from a need to reduce costs and tap a waiting array of expertise. Learning these tools is a better next step for the 3000 expert who wants employment or engagements. There's lots of Windows in the world, but lots of Windows expertise, too. Leading a company toward a path of open source solutions is good, groundbreaking work. It might be a novel achievement to list in a resume or a consulting skill set.