Yes, Win7 x64 Professional. I am using a Samsung 840 (128GB) for the pagefile disk. I have been putting pagefile.sys (and TEMP, TMP) on a separate disk since NT 4.0. I have not run any benchmarks on Windows 7 to see what the improvement is like.

If you are curious about the impact of the pagefile on the OS drive, I would look at disk queue depth and file latency. As long as your queue depth isn't over 2 and your file latency stays in 5-10ms range, you should be fine.

Before I went SSD on the data drive, I was running 7200 SATA drives in RAID1. The write penalty kind of sucked, but having two spindles for reads was nice. The benefit only really showed up during sustained reads, like loading new levels in various games.

To be completely honest, as soon as the system starts hitting the pagefile with any sort of regularity, you need to buy more RAM. Tuning the disk subsystem is not going to get you much benefit. For example with WatchDogs, I started playing it with 6GB of RAM. That was not enough, and the OS was paging at 16MB/s. That was just too much IO, even with a dedicated drive. The drive itself can handle way more than 16MB/s, but the application could not deal with the latency of having to go to disk.

Where did you get the impression that I am not running SSDs?

I was really looking forward to the game and pre-ordered it for PC. My experience has been horrible. I am running an i7-960 (8 cores, 3.20ghz), 12GB of RAM and 2 GeForce 660s in SLI (4GB of total video RAM). I have a dedicated OS drive, a dedicated games drive and a dedicated pagefile drive. By way of background, I run ~1400VMs for a living. The VMs support a number of SaaS applications that are sensitive to transaction latency. I tune applications for performance for a living.

The game runs like crap on my PC, even on medium settings. It reads files from all over the place. It pulls textures out of the temp directory. It pulls data files out of the game directory. Even with over 4GB of FREE (not Available) RAM, it still manages to make the system do a steady 2MB/s of paging.

The game play is horrible. The driving is clunky. The interface scheme was obviously designed for a game pad. The multi-player is embarassing. The net code is crap. With 6 people, there were serious rubber banding issues. That was with a very small slice of the map. It is not like they had to render the entire thing. In a good 50% of the multi-player games I was in, there was at least one invulnerable person. That leads me to believe that the code is obviously pretty easy to exploit.

The game concept was a good one, but the execution was horrible. I have learned my lesson. In this day and age, everything is in beta. Developers are okay with releasing incomplete products and patching them later. I spent my youth couriering warez and getting a free ride. Now that I can afford games, I have been willingly purchasing them to support the studios. I cannot do it anymore. They just release crap products. They are not even worth pirating.

I was more concerned with the Americans who are going to be eating the bananas that, according to the article, are being sent to the States.

File this under donotwant. Unlike most vitamins, A is not water soluble and can build up to toxic levels in the body. Of special importance to Africans is that it also makes the body more susceptible to sun burns.

I was thinking something similar. Now instead of having a bunch of numbers easily accessible to thieves in a compromised POS system, they are simply going to be discarding a bunch of imprints covered in Chinese food waste.

My experience with the technology has been peripheral. The demonstration I saw showed clustering and with that technology, it removed the need to know what to look for. It put the concepts / clusters together for you and let you see, at a high level, basically what the data set was about. This was assuming that the data set was trained on itself. If you had a more focused training set, say one derived from already vetted intelligence, the algorithms could sort through the noise to find conceptually similar result sets.

The above assumes that there are compute and storage resources available to do the initial training. Obviously the entire data set is too large, so it needs to be culled and that is the challenge.

Off topic and not terrorism related, but it would be interesting to see the correlations between big wins in the stock market and communications patterns between individuals leading up to them. Not that the SEC has the stomach for it, nor am I comfortable with the civil liberties implications of such a setup, but it would be cool to let the algos go for a year or two and see what potential prosecutions they bring back.

Isilon + Hadoop + Content Analyst (http://en.wikipedia.org/wiki/Content_Analyst_Company)

The last piece is just what the civilians have access to. It came out of the intelligence community. You can guarantee that the NSA has something way more advanced and/or better optimized at this point. Specifically look at Conceptual Clustering and Categorization.

Their challenge is not going to be pulling the data out of the haystack. It is going to be having enough analysts to sort through the results and enough guidance from on high as to which result sets to review first.

The difference is that the banks are being sued and the underlying data (loan records, underwriting guidelines, securities information, etc.) is being preserved and provided to plaintiffs and defendants. The NSA is refusing to do collection / preservation / discovery at all.

Right. And every child is a special snowflake, so different than their peers and the generations that came before them. Heaven forbid that people who have dedicated their lives to educating children should be allowed to leverage data sets and discuss those data sets with their peers.

Children inherit their values from their parents until they are old enough to develop their own. The community are the educators who are teaching the children. Nerds talk about nerd stuff. Jocks talk about sports. Teachers talk about education. Children have plenty of opportunities to choose various paths. If as a parent you have problems with teachers doing what teachers do, then maybe you should educate your children yourself.

Personally, I went through the public school system and I turned out well enough. I can support myself and my family. I contribute to my community. I was in GATE (Gifted and Talented Education). It was one of those evil programs where educators got together and put together curriculums tailored to "Gifted and Talented" students. (http://www.cde.ca.gov/sp/gt/gt/) There were criteria and tests required to get into it. I was ranked against my peers. Oddly enough, everything turned out alright in the end.

You just assigned me a belief that I do not have. I was simply making an example. Replace "make more money" with "more likely to help an old lady across the street" if it makes you feel better.

Now are you asserting that because there are some incompetent teachers out there that educators should not tune their curriculum to produce more students who are inclined to help old ladies across the street?

You do realize that with a large enough sample size, the impact of singularly incompetent actors will not statistically influence the results, right?

Students have different learning styles. The algebra teacher who you thought was incompetent may have just had a teaching style that was incompatible with your learning style. I struggled with high school math in school and always had to retake it in summer school. The second time through, I got A's and finally absorbed the material. With regards to math, I needed to see the end state. Math was not taught to me in that way. It was taught in isolated steps that were pieced together to solve larger equations. Until I saw how everything fit together, I was unable to conceptualize the individual concepts and failed. Is that the teacher's fault? Is it mine? Or is it simply a fact that one size does not fit all in terms of education?

The records of violence that you are so concerned about are already recorded and are not stopping people from getting firearms. The records of minors are sealed when they turn 18.

Every educator that I have known has acted with positive intent and a genuine desire to make the lives of future generations better. People do not go into education, especially in public schools, because they want to get rich or amass influence and personal power. They do so because they are gluttons for punishment and believe that it is their duty as human beings to make the world a better place.

As a society, we see our data being used against us. Where as the educators are trying to track the effectiveness of their programs, citizens are fearful of the data being mined for nefarious purposes. Some things that come to mind are, increased healthcare premiums / denial of coverage. Denied job opportunities due to invasive background screening. I am sure that the concerns that people have are numerous.

The other side of the equation is compelling though. If the educators are gathering data that showing people who failed or never took geometry end up making 50% less more than students who do pass geometry, they will more than likely look to tailor the curriculum to help students develop the skills and abilities required to pass geometry.

The other issue is monetization of data. Nobody wants to be a product, especially if they are not receiving any benefits. To use the geometry example above, if the data sets are being mined to extrapolate data like, "Students who pass geometry are 50% more likely to purchase a luxury automobile." and that data is then sold to marketers to target Facebook advertising, people are going to be understandably upset.

It all comes down to trust. Even if the educators can prove that their intentions are pure, what about the third parties they engage? What if the third party is initially pure, but then they go bankrupt and the personal data is sold as part of the liquidation of the company? Who is going to control what the fourth party does with it?

Recently there was an article about how the FBI was having problems recruiting competent IT talent due to their zero tolerance policy with marijuana.

Apparently that problem has been solved. All they really need to do is arrest the people who have the skills that they need, and then coerce them into doing the work that needs to be done.

We all know that the prison system is often tapped as a source of unskilled and low skilled labor. Obviously this is just taking that model to a new level. What's next? Mass incarceration of bitTorrent users who will then be forced into the life of skript kiddies in exchange for money on the books at the Club Fed commissary?

Does it need to be that complicated with the signal strength readings? I am not up to speed on cellular technology, but don't cell towers have the equivalent of a MAC address? Surely there has to be some sort of hardware identifier that is visible. We are talking about TCP/IP here....

If so, it would be easy enough to develop a database of legit addresses and do a look up against that list every time a hand-off occurs. The list could be easily paired down by county / state / zip code.

I used to work for an organization where a handful of our users traveled internationally on a fairly regular basis (a few times a year). Because permanent international plans were much too expensive, we "activated" international roaming on an as needed basis. Without fail, every month following an international roaming activation AT&T would fail to restore the account back to its previous plan. The plans always ended up on the most expensive plan AT&T had at the time, completely ignoring the corporate plan that we had with them.