Many point to Social Security as the whirlpool of social suck that will destroy us but at least some measure of the failings are in the same category of reasons the Pentagon can't find all its bullets (e.g. the computer systems were all designed by mice who never thought of more than hiding cheese). See AP: Social Security's $300M IT project doesn't work
The Pentagon inventory system sounds like a good idea ... but it doesn't have one. Instead each branch of the military has its own. It gets better ... they don't talk to each other. Some projections show the systems will be merged in 2017. In my own crystal ball, I see Tinkerbell coming back to life sooner.
The Social Security has a similar problem but it's larger in some ways as there are fifty-four different computer systems that need to be merged and integrated on one large system. Keep in mind the 'one large system' might be a tightly-linked configuration of smaller computers designed in such a way for reliable failover. That aspect of it for the moment is too deep into design to have any value in consideration.
These problems are symptomatic of computer systems problems across the country (i.e. air traffic controller systems are another screaming example) that were built by people with good intentions but with relatively-poor hardware and design skills that were not sufficient for modern systems. This is why it's been a generic excuse for decades to respond regarding any problem, oh, something went wrong with the computer. In fact, the computer is just fine but an amateurish system surrounding it doesn't work.
Lotho said I could do it. Yes, in fact there's a time when I could have done it. I'd pair up with my friend, Ronald, as we are some dangerous individuals when it comes to big systems. He does the stuff he thinks is cool but I hate and I do it the other way around. Both of us think we work on the coolest stuff in big systems and both of us are right.
The two of us start building teams as we want some people who are very good with networks but aren't so twisted on Unix that they see it as a religious vision. We need sysfrogs who don't get weak-kneed about building big things. I know where to find them as Ron never left them. I remember giving a kid a start about five years ago so now I'd be calling him up to say well, well, young man. Do you think you're ready for fighting dragons.
Yah, yah, that's all bravado. The reality is that the model strikes me as much like a particularly complex bank conversion. Even with the largest conversions with which I was involved, typically they would be done in a year after who knows how much upper-management review. (The time periods on there are very loose. Don't worry about it.)
Side-note: there was a guy in Clifton who would try to cross a street when the light changed but he would change his mind partway across the street and he would go back. He would do this over and over until he was satisfied it really was safe to cross. He was a nice guy and absolutely harmless ... and he would have been perfect in upper management as they work in exactly the same way. Unlike upper management he didn't spend the rest of his time in the relentless solicitation of vendor bribes and free golf trips.
Through the general conversion method, one takes on a single system, studies every damn molecule in it and then does various types of 'mirror magic' to convert it to believe it's now part of a much bigger system than the single standalone system on which it previously lived.
(Ed: it would take fifty-four years at least to complete this conversion.)
This is why you are not on the design team, ye of skinny wallet, as this is a Federal program to convert one of the biggest and most important systems in the country. I'm not bringing one single-threaded team to bear on this, I want fifty-four of them and they will do it in parallel. As each team brings its component up through quality testing, it gets merged over to the big system, it's validated one more time and then move on to what remains.
This isn't even such an innovative form of systems building. How the government managed to find people who couldn't handle it is kind of a mystery ... but the F-35 doesn't work as a jet fighter and they spent a trillion or so on that. See above about the Pentagon not being able to find its bullets.
The Pentagon inventory system sounds like a good idea ... but it doesn't have one. Instead each branch of the military has its own. It gets better ... they don't talk to each other. Some projections show the systems will be merged in 2017. In my own crystal ball, I see Tinkerbell coming back to life sooner.
The Social Security has a similar problem but it's larger in some ways as there are fifty-four different computer systems that need to be merged and integrated on one large system. Keep in mind the 'one large system' might be a tightly-linked configuration of smaller computers designed in such a way for reliable failover. That aspect of it for the moment is too deep into design to have any value in consideration.
These problems are symptomatic of computer systems problems across the country (i.e. air traffic controller systems are another screaming example) that were built by people with good intentions but with relatively-poor hardware and design skills that were not sufficient for modern systems. This is why it's been a generic excuse for decades to respond regarding any problem, oh, something went wrong with the computer. In fact, the computer is just fine but an amateurish system surrounding it doesn't work.
Lotho said I could do it. Yes, in fact there's a time when I could have done it. I'd pair up with my friend, Ronald, as we are some dangerous individuals when it comes to big systems. He does the stuff he thinks is cool but I hate and I do it the other way around. Both of us think we work on the coolest stuff in big systems and both of us are right.
The two of us start building teams as we want some people who are very good with networks but aren't so twisted on Unix that they see it as a religious vision. We need sysfrogs who don't get weak-kneed about building big things. I know where to find them as Ron never left them. I remember giving a kid a start about five years ago so now I'd be calling him up to say well, well, young man. Do you think you're ready for fighting dragons.
Yah, yah, that's all bravado. The reality is that the model strikes me as much like a particularly complex bank conversion. Even with the largest conversions with which I was involved, typically they would be done in a year after who knows how much upper-management review. (The time periods on there are very loose. Don't worry about it.)
Side-note: there was a guy in Clifton who would try to cross a street when the light changed but he would change his mind partway across the street and he would go back. He would do this over and over until he was satisfied it really was safe to cross. He was a nice guy and absolutely harmless ... and he would have been perfect in upper management as they work in exactly the same way. Unlike upper management he didn't spend the rest of his time in the relentless solicitation of vendor bribes and free golf trips.
Through the general conversion method, one takes on a single system, studies every damn molecule in it and then does various types of 'mirror magic' to convert it to believe it's now part of a much bigger system than the single standalone system on which it previously lived.
(Ed: it would take fifty-four years at least to complete this conversion.)
This is why you are not on the design team, ye of skinny wallet, as this is a Federal program to convert one of the biggest and most important systems in the country. I'm not bringing one single-threaded team to bear on this, I want fifty-four of them and they will do it in parallel. As each team brings its component up through quality testing, it gets merged over to the big system, it's validated one more time and then move on to what remains.
This isn't even such an innovative form of systems building. How the government managed to find people who couldn't handle it is kind of a mystery ... but the F-35 doesn't work as a jet fighter and they spent a trillion or so on that. See above about the Pentagon not being able to find its bullets.
No comments:
Post a Comment