If you thought the Modernizing Government Technology Act would mean the end of COBOL code, think again. No, people. The government will continue to look like a technological mashup — think Flintstones meets Jetsons.
Some background: The law provides a $100 million fund called the Technology Modernization Fund (TMF), which agencies can tap for projects. But first they’ve got to create proposals that pass muster with the Technology Modernization Board, which are due the end of June. Criteria for approval derive from last year’s White House report on IT modernization, while the General Services Administration has an office devoted to managing the applications and approvals process.
Oh, and agencies that get money must pay it back from savings.
As Federal News Radio Executive Editor Jason Miller reported in detail, the gatekeepers are looking at projects worth between $2 million and $10 million. I went back and watched the GSA webinar Jason’s story drew from, which was by no means visually slick but still informative. It should have more views than the 376 I counted this morning.
In any case, proposals should emphasize fast delivery of product and payback. They should have a high chance of success, and they should demonstrate use of new technologies and citizen-facing services.
It all sounds good but I see a contradiction.
For decades, “modernization” meant replacing legacy systems with new ones. The IRS spent billions on this goal. After all the “sturm und drang” and smoke cleared, the assembler still stands unperturbed, cranking away on mainframes.
In the context of the MGT Act, and in the apparatus building up to carry it out, “modernization” seems to mean something completely different; to wit, adding new systems using blockchain, artificial intelligence, you-name-it, to deploy systems that feel like Alexa or Uber, but for government services.
So where does that leave the big systems operating the Social Security Administration, IRS, Centers for Medicare and Medicaid Services, Veterans Affairs Department and dozens of others? With millions and millions of lines of COBOL, Assembler, even C.
No one will replace SSA’s code base with a $2 million project, so the savings used to pay back the modernization fund won’t come from a reduction in mainframe or data center operating costs.
By coincidence, I spoke Tuesday with the CEO of a software company focusing on mainframe applications and operating environments. He said “working code is gold,” and there’s good reason such systems operate with such high reliability for so many years.
This executive wondered why the fuss about COBOL in the first place. It’s all just syntax, he said, writing lines to fulfill business rules. In fact, doing what COBOL does in Java can double the volume of code.
That thought, or at least a strain of it, seems to also be in the minds of the authors of the Federal IT Modernization report. In the executive summary on network modernization, they state several reasons for updating the federal network architecture. Among them is to “securely maintain legacy systems.”
The truth is these legacy systems don’t do digital services, at least not how most people envision digital services these days. But you can get legacy systems to enable digital services, and that’s by extracting the data from them and applying it to another logic system, programmed in any language you like.
You can even use open source applications. That’s what several agencies ranging from the IRS to the Air Force logistics operation have done. It’s what banks, airlines, legacy retailers and a host of other industries have also done.
In other words, when you put up a tent for a block(chain) party, you don’t also tear down your house.