You know, they already got Google to bend over the table, why haven’t they taken the logical step and offer a deal - upgrade our IT and we’ll start using lube.
What gives me heartburn is the impossibility that the current DOD acquisition system has any competence whatsoever to accomplish the replacement task.
When I was in college all they had was an IBM 360. So I learned COBOL, RPG, FORTRAN, and Assembler. Oddly, I liked Assembler the most, though I only used COBOL after graduating in…some year that I don’t remember. 
I shouldn’t be surprised. Military procurement for high value equipment is…beyond obscene in it’s development and cost.
They also exaggerate the problems of training. COBOL is a lot easier to learn than one of the C variants. And learning a new application is only as hard as learning a new application, regardless of the language.
I think in those days the parts an application programmer was responsible for were in fewer and obvious places, compared to all the files that get scattered all over a PC/server, to support a modern application.
Not that this is a contribution to the discussion but this issue was highlighted on HBO’s John Oliver show, like two years ago
John Oliver talks about US Nuclear arsenal
Having worked in the field of IT Security, this is not at all that surprising but certainly worth a robust discussion. IT Security on all infrastructure in the US should be a huge issue for all of us, but sadly much or our politics is so focused on the changing wind day to day.
1 Like
Ok, I’ll admit it, I’ve been in the IT field for 40 years. Some of what I’m going to say comes from that background, and I’m still doing technical support on a Mainframe computer that still has some software on it that was written in the 80’s. So take what I have to say here with that in mind.
When we start talking about ‘obsolete’ computer systems, we really need to distinguish between 2 different things; hardware and software. In terms of hardware, the field has indeed moved forward at a blistering pace, and the physical architecture of a computer center is light years away from how things were when I got started. If you look closely, the most important changes have occurred in the areas of networking, storage and distributed processing. This has not only improved the way the jobs that computers do for us get done, but have created entirely new jobs for computers to do. It’s been a revolution, and a truly amazing thing to watch as it’s happened.
While there are new things for computers to do, they are still doing the things we originally created them to do; store transaction data, recall that data when needed, analyze that data over a fairly long time span, and control access to all that data. It’s very easy for people, including the leaders in IT organizations, to get focused on the next new shiny object, and not get involved with systems that are currently working well; this can go on for a long time. As with the infrastructure problems in the country at large, there is no glamour in maintaining what you already have, or think you have. This is one reason why we see some organizations with completely antiquated hardware in some parts of their architecture, while other parts are brand new. They often expect the new systems, often sold to them by vendors as a complete solution to their business needs, to replace the old systems which can then be retired; however, what very often happens is that the new system leaves out some critical function that the business needs to continue to function, and so the old system continues to survive and require expensive maintenance. I know it’s hard to believe this could happen to the Government, but it’s really true!
I think we sometimes overlook the social aspects involved in the evolution of information technology, which is something I’ve thought about a lot over the course of my career, having been part of the flow of it myself. There are trends toward certain languages and certain computing platforms that are completely unrelated to the effectiveness of those languages or the technology of those computing platforms. For anyone with a rudimentary understanding of how computers work, the basic things they do has not changed in all the time we’ve had them; they add and subtract numbers very quickly. Computer languages simplify the process of telling a computer what to do, so we don’t have to solve the many problems of interfacing with hardware and managing memory and complex mathematical algorithms over and over again. Over the wide range of problems we ask computers to solve, the basic truth is that coding is coding. There are certain languages that are marginally better for certain applications, but in general, the guys who write computer languages usually make it possible to utilize all the available functions of a platform. I used to know a guy who had the walls of his office plastered with snippets of code where someone told him ‘you can’t do that in this language’, and he wrote it. What I’m trying to say is that speaking of a language, such as COBOL, as being obsolete, is a misnomer; there’s nothing about a C++ program that works better than COBOL with a current compiler over the broad range of computer tasks. COBOL has simply gone out of style, because it is a mostly Mainframe language, and most people don’t get to write software on a Mainframe platform. But as a business language, it is still easy to learn and easy to read when you go back and have to maintain it, and that saves a lot of money. It just isn’t one of the early languages of the PC platform, where a lot of people learn to code now, so it’s not part of what might be termed the standard education of a modern computer geek.
It took a long time for Windows to be taken seriously in the modern Enterprise computing world, and now that it has matured into something that can work well for many applications, the older Mainframe platforms are spoken of as obsolete, because a lot of management in place now grew up with Windows, and that’s all they know. But Mainframe hardware and operating systems have evolved right along with everything else in the Information Technology universe, and they are still the champions of large volume transaction processing, database access and bulletproof security and reliability. I reload our system a couple of times a year, but it never fails. Never had a virus. Shrugs off unauthorized access attempts. It’s still a go-to platform for large financial services organizations like banks and securities firms, because of all these qualities. Talking about these platforms as obsolete simply makes it easier for our information to be stolen and our lives to be less secure.
I hate to sound like the old man in the corner ranting about how stuff used to be, but I’m trying to talk about how things really are, and some of the modern myths about Information Technology are just that. This is just my effort to correct the record.
2 Likes
A problem of long standing. When I was in graduate school back in the 70s there was a standing IT joke. After giving a group of businessmen a tour of the Cheyenne Mountain command center, the colonel noticed one of the people seemed particularly concerned. “Is something bothering you, sir? The power? The responsibility?” “What I was thinking,” replied the businessman, “was that I wouldn’t trust my payroll to those computers.”
Actually, it isn’t even that deliberate. Before I retired I worked with many of those systems as an outside contractor. Federal IT systems are not only antique, but they also disconnected. Thousands of supplementary programs have been created to shuttle data from one arcane system to another, with customized extraction/translation code comprising millions of lines. Replacing one system means also rewriting integration programs to port data to other systems that will also become obsolete. Truly a 20-year plan with a cost in the trillions.
The upgrades needed are massive and unfathomably complex. My take on it is that no department head wants to take the hit to his budget or start something that will get snarled in bureaucratic horror within 5 years.
The fault belongs to no one party or ideology. Of course, I have a few ideas about how it could be done, but I’m out of the game now
1 Like
Perhaps Rep. Chaffetz could convince the Repukes to fund the upgrades. I wonder if the rest of the world updates computer systems.
2 Likes
LOL That’s a funny story!
Adm Hopper came to speak to my calculus class many years ago. She opened with the “nansecond” story. What an amazing woman!
Ha ha yes! She told us that story too. One of those gigantic room-sized computers was being tested and kept giving inconsistent answers to a set of computations. The whole team was completely flummoxed for hours. Then Grace figured that since the machine ran on tubes (!) maybe one was burned out. So they checked behind the computer and sure enough, a moth had cause a short, burning out a few tubes.
Thus she christened such programming problems as a bug.
Just like the cell phone icon for a call is a hand-held receiver
I’m with you in that corner. I was always a data warehouse architect, not a sys admin, but I have a soft spot for the raw processing power of mainfames and never quite trusted the sales reps who insisted a neural net of PC’s could replace them. When minis became the rage, I remember VAX and PDP systems occasionally going down, but never a mainframe.
I/O lag can make a system seem obsolete, but upgrading storage systems does a lot to extend the life of an older CPU, and some of the fetch and read ahead algorithms are practically magic (to non-hardware guy).
Still know a few COBOL programmers too, although they’re all getting ready to retire.
Your comment, “As with the infrastructure problems in the country at large, there is no glamour in maintaining what you already have, or think you have. This is one reason why we see some organizations with completely antiquated hardware in some parts of their architecture, while other parts are brand new. They often expect the new systems, often sold to them by vendors as a complete solution to their business needs, to replace the old systems which can then be retired; however, what very often happens is that the new system leaves out some critical function that the business needs to continue to function, and so the old system continues to survive and require expensive maintenance. I know it’s hard to believe this could happen to the Government, but it’s really true!”
That has been my experience. I cannot really speak for the language/code side of the house, but from a funding and policy perspective, your comments ring true.
My issue is that I believe in my experience many mid management level folks have gone the route of new and shiny, either because they were sold it or they did not want to invest in the manpower to clean up an old system so that it meets current security requirements and still satisfies the customer. I am not sure the old code vs new code is really the issue. Perhaps it is the current IT security training and certification which relies on the latest and greatest tech spending little time discussing nuances of the history and development (or maybe I have a very narrow perspective and painting with too broad a brush)
“The federal government is years and in some cases decades behind the private sector,” Rep. Jason Chaffetz, R-Utah, chairman of the House Oversight and Government Reform Committee, said in a statement. “Taxpayers deserve a government that leverages technology to serve them, rather than one that deploys insecure, decades-old technology that places their sensitive and personal information at risk.”
Oh, this is rich. Especially when the Republicans have been cutting any funding that would actually go to upgrading the IT infrastructure for the federal government for the past 20+ years.
Yet ole’ Jason is willing to blame it all on the present government. No real surprise there. He doesn’t actually give a damn, and I’m sure that gawd forbid we get a Republican president in 2017, it’s going to stay on the back burner, where the Republicans like it.
1 Like
Nice, well-informed post. Me? I could not have written it, but know enough to appreciate what you’re saying. It could be amplified and edited into a magazine article…
. I sold mainframes back in the early 70s. The cheapest thing in my catalog had a purchase price of apprx $250,000 and today’s PC or laptop could outperform it.
Thank you, that’s very kind.