Your exact reasoning is why I made this: Confidence Levels Some people are still surprised that with 6-10 people, an r^2 of 0.9 might not mean what they think it does.

It's also not quite true. Objective-c message passing is quite fast, only 4x the cost of a virtual table call in C++. If you are really interested in what happens behind the scene, see obj-c fast-path

Now, namespaces are still a honking good idea.

It's too bad they didn't use any in the book.

I have used matplotlib for journal plots and actually gave away a copy at a conference I ran so I have to say I really do like the book overall, but if you scan through the pages, you might be turned off.

I'm not worried, given the speed at which my battery drains, this should reduce the attack vector considerably.

There was no such confusion and again, I have to disagree. Perhaps we are considering different starting points, while you may see convention, I think about how the conventions were started. I think there is more leeway than you may give credit for. It took a great deal of artistic skill and thought to come up with exploded views for example:

http://en.wikipedia.org/wiki/File:1942_Nash_Ambassador_X-ray.jpg

Now, of course future works may be considered derivative. In any case, I think in many cases, having constraints actually is quite beneficial to artistic endeavors.

I think this is fundamentally wrong. The symbolism in technical drawings is not realism, it is implied realism. Having tried to do it multiple times, just the abstraction of information into a usable form is as much art as it is skill.

Quoting a song about a name change causing possible attribution error and attributing the wrong name to the author strikes me as the epitome of irony, but YMMV.

Ironically, this song was written by the Four Lads not They might be Giants, which just goes to show how names and attribution are indeed lost to history.

I'm being a bit tongue in cheek here, but maybe you are rather missing the point of Python which is "batteries included." The python standard library contains much more than Python, did K&R include:

• A web server
• An xml parser
• Email parser
• GUI package
• Windows COM interface
• And so on.

It's kind of surprising that Learning python is only 1000 pages, which is not too say that it isn't too wordy.

I completely agree, at some level this seems like an implementation detail. A good example is how The Secret Of Monkey Island(TM) was ported to the iPhone. This had exactly the same problems as Flash. Being a port of an old point-and-click game they had issues with hover-over and they were able to form a solution for these issues.

I thought it strange at first that it used a virtual cursor instead of just tapping on an object on the screen but it actually ended up working better and they were able to use the same engine underneath the hood. But the thing was, you moved the cursor with your finger and your finger didn't obscuring what the virtual cursor was pointing out.

Now, that being said I'm not sure I would want to have two different idioms but saying that it can't be done is just stupid.

I'm in complete agreement with you if the goal is near 100% automation, false positives are maddening and prohibitive as I mentioned with the Sentinel system. And I will also say that our goals were quite different than yours appear to be. We didn't expect very good false positive rates, but we had quite good false negative rates. However, the main aspect was that we did use the sensor fusion date to guide checkpoint interrogations.

You will note that the best scoring criteria for our data fusion was when an individual went between checkpoints, and was interrogated by a trained guard. What I meant by enhanced interrogation at checkpoints is a little more subtle.

As opposed to:

Danger: Phil is in facility X!

consider

Guard at checkpoint:Phil, we noticed that you went into facility X where you don't normally go, could you tell us why?

As the Israeli's will tell you, nothing beats a good, well trained, interrogator, and the more information he has the better. I will agree that our facilities had far fewer people in them than in an airport so what it boils down to is that you can't interrogate everybody so triage will need to be done.

And you are quite right that detecting objects being left behind was quite complicated. We had two particular benefits (1) We were more worried about objects leaving the facility than entering and particular objects at that and (2) we were in complete (almost) control of the facility with regards to checkpoints.

All that being said, the intrusion detection system from EDS simply sucked. Our best success was using it to toggle the camera priorities, i.e. we didn't show alerts we just cycled the cameras to view the hotspots first, again we used it to guide human intervention rather than supplant it.

I agree and disagree at the same time. I worked on a similar stillborn project named AMISS (Advanced Material Information and Security System) fourteen years ago at a government lab designed to protect theft of nuclear material. There was a particular system from EDS called Sentinel used to identify intrusion and was used in places like rail yards. For a particular use case the false positive rate was staggeringly high and users quickly learned to ignore the alerts.

However, when we used our data fusion algorithms to augment the history of a person at a checkpoint (simulated) false positives were okay, they just enhanced the interrogation.

The problem is false negatives which is much harder to quantify. Of course we had access to a scintillator that could identify trace radioactive potassium from the banana you had for lunch...

Absolutely indeed, in fact didn't Valve hire the author of bittorrent to do exactly this to distribute their DRM'd material?

Pensions are a good thing if they are fully funded. This means that the money is put into a pension scheme while the people are still working not by expecting future generations to magically come up with the past 20+ years of missing investment growth. This can add a small overhead but is similar to a 401(k). The problems arise when:

• People live longer

• Governments/Organizations dip into the funds or the funds aren't fully funded

• Wall Street Collapses, although fully funded public pensions tend to perform better.

While biased, there is a good introduction here. Note: I don't have a pension.