While I agree that we have seen technical advances during the past 20 years that have provided previously unseen tools that could be usable for teaching I don't think we should dismiss the old ways altogether.

Some subjects or fields of research if you will, are actually best taught using a chalk and a blackboard. One such an example is mathematics. A well trained lecturer that proves theorems and solves problems on a blackboard beats any powerpoint any day of the week. In fact, math is one of the oldest subjects there is that is taught and the so called didactics (the science of teaching and instruction or pedagogics) behind it has evolved over several hundred years and it is well understood. At least at the university level, high school math is a joke as far as I can tell, at least the math that has been taught there during the past 3-4 decades or so.

I've also been to business schools/universities where the blackboard has been replaced with a whiteboard and the lecturer is using powerpoints and I can tell right away that more efficient ways to ruin teaching are hard to come by! To put it simply, a subject such as math should never ever never ever be taught with powerpoint slides and a whiteboard!

So the bottom line is that don't dismiss the good old ways that have been developed and refined for centuries!

But carefully note that there are a lot of "new" subjects that have not yet found a good (consensus) way to be taught, examples of such subjects involve; computer science, economics, finance, operations management, logistics, ... etc. So it would be interesting to follow how the didactics behind them will evolve over time.

Also note that the average skills among people in math and language have steadily declined during the past 3-4 decades at least in the western world so I would say that at large, the educational system has rather devolved than evolved in spite of computers and the whatnot. I would even dare say that computers and all the gadgets around us have dumbed people down quite a bit. We don't need to be able to read and interpret maps anymore, we don't even need to be able to spell properly as spell checkers take care of that. Fact is that computers and the technical means available do more and more of the thinking for us and we should be careful about it as these means can do more harm than good as we can grow to become too dependent of them.

Maybe it is you that are confused. You are talking about the "2^N" approach to these prefixes as if everyone in the world uses it. Well, I've got a news flash for you; they don't. According to the ISO system, one kilogram is 10^3 grams, one megawatt is 10^6 watts, one gigapascal is 10^9 pascal and so on and this is what is widely accepted. The 2^N convention is not widely accepted. I understand it is reasonable to use it on bytes because the unit itself holds 8 bits (which is 2^3 bits) and not 10 bits. But the bit in and of itself does not have that oddity. One bit is one, just like one atom or one line of code (LOC). As far as I can recall one "k-loc" is 1000 lines of code and not 1024 lines of code.

Since we're talking 'bits' and not 'bytes', isn't it a little bit misleading to say 512kbit/s to denote half a megabit per second? The convention of using the power of twos that are closest to each decimal prefix usually applies to bytes for memory addressing reasons but not on bits as far as I know.

The memory addressing space i.e. set (see mathematical sets ) of all possible memory addresses are always a power of 2. On a 16-bit memory bus the address space is 2^16 = 64 kB. More can be used by using memory banks where you can switch between them using some kind of a binary multiplexer. On a 32-bit system the address space is 2^32 = 4 GB and on a 64-bit system the address space is 2^64 = 16 EB (exabyte). In for example a 16-bit memory bus the address of a certain memory byte is expressed by a sequence of 16 binary digits e.g. 1101001010100101 (0xD2A5 in hexadecimal) which together gives rise to 2^16 possible combinations which is what defines the memory addressing space in a computer. This is why the power of two addressing convention is being used and widely accepted by the majority of computer users as a measure of amount of memory in RAM and hard drive storage.

It's the so called Voight-Kampff test in the Sci-Fi movie Blade Runner to test whether a human is real or a so called "Replicant".

The "profile" of one "best" person is not a statistically reliable indicator of what type of people to look for. From your description this person looks like a statistical outlier; there are many thousands of people out there who "barely gives a crap about the job" and are really lousy at what they do.

If you want to establish more reliable criteria for people to look for, you should establish a population of maybe the 10 best programmers or even the 20-50 best programmers you have run into during your 20 years in the business and look for commonalities among them, their driving forces, their passions and so on.

I have run into a few geniuses in the field where I'm at and I have found only two common factors among them; they are very intelligent and they are good at what they are doing. Apart from that, I have found them to be very different from each other when it comes to personal traits, interests and political opinions. I could not guess that they would have the level of skill that they had. Only in some cases the person was "socially awkward" and/or had an "exotic" taste for clothing where I could figure that either this is a total fool or an utter genius and the latter turned out to be true. In another case one guy was dressed as a "goth" with piercings and makeup, it was easy to dismiss him but after a while I realised that there was more to him than his style and looks.

I generally don't delete files by mistake but there was one day when I did. I had a directory with .rar files that I wanted to remove while doing other things that involved repetitive tasks. So I issued rm *.rar in the directory I was at and removed the files that I wanted. Since the tasks were repetitive I issued commands from the command history and when I was in another directory I mistakenly brought up and entered the rm *.rar command again. Not funny.

I think it would be useful if there were some sort of prefix (like I enter !rm, !chmod, !format instead of rm, chmod, format) that prevents the given command to be either stored in the command history or be prevented from being invoked from the command history again.

For examples of using vocal fry in singing look at the following clip that is "I Don't Wanna Miss a Thing" by Aerosmith:

http://www.youtube.com/watch?v=Vo_0UXRY_rY

pay particular attention at 0:31, 0:39, 0:48, ...

Or look for example att "Angels" by Robbie Williams:

http://www.youtube.com/watch?v=Fl4q2WBCywg

at 1:30. So vocal fry is also used for accentuation and is not merely a trick to compensate for lacking vocal abilities in singing. Vocal fry is also used to activate the vocal cords so as to help building up a vocal technique with a smoother transition between what is called "chest voice" and "head voice" without disconnecting or bursting into falsetto when reaching higher registers or "slamming" into chest voice when transitioning from higher registers into lower.

Well then, there are four ALUs in each core and not two cores. Just because the said performance is higher per core in the next generation CPU doesn't make it right to double the core count the way AMD has done. Hyperthreading or SMT as other CPU manufacturers than Intel call it can allow for more than just two threads per core. The implementation AMD has done is merely an enhanced version of SMT, no more no less.

Sure everything is nice and dandy in theory whereas in practice things are not so shiny and the Bulldozer family has failed miserably to impress even with their line of Opteron 6200 CPUs on server benchmarks. As the per-core licensing scheme is becoming the standard licensing scheme among servers, the Bulldozer family of CPUs looks very unattractive right now.