Ummm, how many Olympic sized swimming pools is that?

It would be fascinating to know the infrastructure and methods used for storage and to process this volume of data. Presumably, they initially store everything, and then somehow process it to decide what is worth keeping as future potential blackmail material, or occasionally intelligence purposes. The scale of the task is mind boggling.

The singularity, where supercomputers can advance scientific knowledge unaided by humans, is still some way off. However, you are mistaken if you believe there have not been huge advances in scientific knowledge in the last 20 years, or if you believe the rapid pace of advancement would have been possible without the computing power that has become available to support that effort. In earth sciences, medicine, high energy physics, astronomy, meteorology and many other scientific areas, the simulation and information organization capabilities facilitated by state of the art supercomputers have been absolutely crucial.

Sometimes, availability really is critical. In that case I want to take the risk of an automatic restart before the cause is investigated. However, it is important to appreciate that the approach is risky . The restart can cause cascading errors that change a reasonably simple issue into a multi day recovery operation.

You make some good points. However, I believe the march towards the singularity will march inexorably forward for one (highly undesirable) reason: the insatiable appetite of the leaders of nations for power. The populations of those countries will not even be allowed to know much of what is being developed with hundreds of billions of their tax dollars, but technologies that leaders perceive could enhance their ability to dominate the world will be financed. There will be no regulation. If you want to know the state of the art in visual recognition, you should look at military applications: robot soldiers and autonomous drones. For applications of big data (especially its usefulness in widespread blackmailing activities) then, in spite of some initial missteps, look at the pervasive collection of data by the world's "intelligence agencies".

The laws against child pornography should be aimed at protecting children from exploitation, not in making morality statements. Cartoon drawings of children engaging in sex acts certainly indicate people with pretty sick imaginations, but no children are hurt in their creation or consumption. I have seen worse on walls in public washrooms.

I cannot find much detail on this, but it sounds suspiciously like well known techniques for avoiding congestion in complex systems that I learned in queuing theory over 40 years ago.

I guess the website's testing process failed to catch the edge case where someone wanted to navigate past the home page.

Since the source is completely unclear, most posters will blindly assume it is the fault of whichever group is their bête noire. Some favorites will likely be China, North Korea and Russia, but use your imagination folks. There is just as much evidence that it is caused by evil bankers, genetically modified foods, pedophiles or US militarism.

Presumably, someone has been using the infinite improbability drive.

Considering the request to rewrite the record also required the falsification to be secret, how would we know how many times this has been done in the past. We only know that this is the first time such a request has been rejected. I personally suspect it might only have been rejected because of the large number of witnesses of the original hearing.

The knee jerk reaction, of course, is to look for a catch in anything Homeland Security is doing. However, this seems like a really good idea. Finally, they are contributing in a positive way to public safety.

arguments are in the next article

No they're not!

Believe it or not, using 2260 terminals connected to an IBM 360 mainframe running MVT/ASP, I was able to rerun commands anywhere on the screen back in 1973!

A lot of what is in the example reminds me of a 4GL I worked with in the 1980s that had a feature called "computed fields". The idea is extremely nice conceptually, and seems to nest nicely, as well as be easily integrated with other functionality. The main problem is performance. Some pretty smart people worked on the tools, but (with the complexity of real life systems) you tend to end up with the same values being continually recalculated. It turns out that (because of the generality of the functionality, and the inability to predict when values will be updated or queried) it is extremely difficult to suppress these duplicate calculations. A typical application developer will create code that results in values being recalculated thousands of times in a single transaction. The problem is both worse and more intractable than is experienced with computed columns in spreadsheets.