VistA modernization must use MUMPS

Once more, the debate about “whether” to migrate VistA from MUMPS has come up again in main stream press.

This always makes me sad because it shows just how fundamentally ignorant people are of what VistA is.

So lets get something straight. If you are not using MUMPS, in some form or fashion, it is not VistA. It is a new software project. New Software projects to develop comprehensive EHR solutions, do not work. Ever. That is called “Big Bang Development” and it is utterly doomed to fail.

In order to create an EHR system you have to grow one. You start with a system that is not comprehensive, you use it anyway, and then it grows into something that is a comprehensive EHR system. You cannot take a comprehensive EHR and assume that you can re-write it, from scratch in another language and that it will work. That is just unfathomable.

This has been tried, several times, and consistently failed.

So the reason that it is not “an option” is that it will fail. Thinking about it as an option is simply madness.

It is very like saying, “We need the Linux Kernel to improve, so we will recode it in Java.. not enough people are trained in low-level C programming” The folly and hubris should seem clearer now perhaps?

You might try drastically reinventing what MUMPS is, like ClearHealth, but you cannot simply “get away from it”.

Another example might be “New York has proven that the street + subway system is effective, the city planners of Venice should adopt that in place of the canal system that they currently use. Obviously New York shows that these modern features are capable of moving far more people…” The reason Venice does not consider a subway system is that it -cannot- work. The city planners there know that, so they never try.

It is ironic that people who say “we should move away from MUMPS” consistently consider those of us who actually understand the architecture and design of the system, and insist that we continue with MUMPS as a kind of “particularly obstinate political faction”, from the link above:

“Is MUMPS the right entity? I think the obvious answer is ‘no,’” Meagher said. “It just happens to have a bunch of very committed people who want to stay in that environment.”

When an engineer says “I can think of no way to achieve near light speeds in our lifetime”, he is not taking an obstinate political position. It is not “pro-light-speed” vs “against-light-speed”. The engineer is taking a position based on what he understands to be achievable.

When I say “VistA must stay with MUMPS”, I do so based on the only relevant evidence’ efforts to move away from MUMPS have consistently, and expensively, failed. I do not like MUMPS at all, but I have a pretty solid understanding of software engineering and you simply do not simply migrate to a new language for a codebase as large as this.

-FT

2 thoughts on “VistA modernization must use MUMPS”

  1. I don’t know enough about VistA to know whether it makes sense to migrate the code away from Mumps. However, I do know that the Healthcare industry in general must move away from Mumps sooner or later. The gap between Healthcare IS and the rest of IS is growing. Do we want to be using this obsolete technology 50 years from now? If not, we must begin to change direction now. I don’t know how to do that, but I think recognizing the problem is a first step. Once a critical mass of Healthcare insiders recognize the problem, we can begin to discuss possible solutions intelligently.

    There is no question that moving the entire industry away from Mumps will take decades. I think of it as being analogous to lead-based paint and asbestos. We have not eliminated either from the environment. But at the same time, when you add to your house, you are not insulating the pipes with asbestos or painting your new bedroom with lead-based paint.

    Yes, of course, we cannot stop getting things done in order to make things better, but if we never try to make things better because it is too hard, they won’t get better.

  2. Your assertion here is that technology is linear, but we know that is not true.

    MUMPS is a powerful “NoSQL” technology. Does the fact that most of the popular modern databases mean that we need to remove SQL like we remove “lead paint and asbestos”. Your analogy leaves no room for multiple competing technology strategies. For instance paper books will probably be around for a long term, because as a technology, they still have benefits over e-books.

    MUMPS is a modern technology that is still developing. The fact that you are not abreast of those new, modern developments means that it is just not your technology stack of choice. Don’t get me wrong, I am not a fan of MUMPS, but it is not because it is “old” technology, that is just not really a valid argument. A much better argument is that it is an “uncommon” technology choice. But in healthcare, it -is- a common choice. So even that does not matter too much…

    -FT

Comments are closed.