I'm signed up for almost every coursera MOOC.

I've only officially completed 1, and watched every video for about 30 others, and have downloaded videos 'to watch' for most of the others.

A few things I've found are that

1) Professors seem to like to assign waste of time busy work.

There are lots of classes that require essays or projects where it is essentially a giant waste of the students time. This includes doing videos and presentations for almost any course (a really well taught audio production course wanted every stuent to do a video essentially repeating a subset of the same material he just did. Others have wanted various large scale projects.) Since there would only be 'peer' evaluation of the material, this was all essentially busy work. There are areas where peer evaluation can be useful (some writing with rubrics and such), but mostly it was stuff that wouldn't matter at all from improving learning. Or the amount of learning improved versus the time invested was drastically out of proportion.

The math, science, programming and finance classes tend to 'get it right', only assigning an amount and type of assignment required to understand the material well, not wasting students time.

2) Science, Programming, Finance, Engineering, and Math courses are real courses, courses from Bschool and other sections are often ridiculously simple.

Of course testing and evaluating understanding of computer and science courses is quite easy, but still the quality and type of questions asked in reviews and homework and the type of assignments made sense for the Science/Tech classes; whereas I was sometimes wondering why the other courses had even bother to do a quiz the questions were so ridiculously simple minded.

Kepler is likely a poor choice of analogy. Much more apt is probably religious conversion, or political party conversion, or social movement conversion. Most individuals who have complete reversals of position have neither position based on logic, but on emotion. The individuals will have rationalizations for their emotional decision, but it isn't useful to look at their reasoning, since the 'reason' is generally post hoc rationalization. Usually the actual reason is either a traumatic event (ie death of a family member causing loss or gain of faith; 'betrayal' by a politician) or change in self interest (monetary, sexual mores, regulatory costs/benefits, social welfare costs/benefits).

I used to work for a startup company that created an amazing new block design, lays up like standard masonry and has the beauty of masonry.

http://onestepbuildingsystem.com/what-is-onestep.html

Has an integrated cavity that is filled with concrete and rebar, so is ridiculously strong. And the insulation seals it against water penetration (not as well as the original design which had more internal plastic, but in the water penetration testing it stood up to hurricane force driven water without leakage.)

Also has great sound insulation, has thermal mass to the inside which drops heating and cooling costs significantly, and maintenance is fairly inexpensive.

Not sure what choice you'd use for windows, I recall seeing some that were quite amazing 10 years ago, when I last looked, but I'm sure the market has devised some even cooler stuff since.

No you don't have the implement the functionality, just the API.

Ie if the GPLed api has

int Square_Root(int x);

then my binary compatible test API can just do
int Square_Root(int x) {return x;}

It doesn't have to be functionally equivalent (or even if it is functionally equivalent a far simpler method to implement) just binary compatible to compile against.

Read the LGPL license - at the users option it can be converted to the GPL.

"The user of the driver wouldn't be violating the GPL no matter what. The GPL comes into play upon distribution not use."

I meant that the getting of the driver to the user without distributing the source to the driver would not violate the GPL.

A point/example I meant to raise above would be projects like WINE or Dalvik, that reimplement APIs that are binary compatible with APIs of other OSes. Obviously I can compile a a program against the official windows API and then run it in WINE, and yet the LGPL (or GPL if I compiled the WINE using that clause) would almost certainly not make my software a derivative of the WINE software. This seems directly analagous to the case above where I rewrite a binary compatible API, compile against it, then run the driver with the GPLed kernel.

Or alternatively - a case that is more personally relevant would be individuals creating closed source scripts against the Blender python API. Or closed games that are interpreted by the GPLed blender game engine and use the Blender python API.

It is often asserted (ie by the FSF) that exposed C/C++/Python APIs for GPL software can't be used by non GPLed code unless a specific exception has been added to the license.

However, a non GPLed binary compatible API could be done that the plugin, etc. can be compiled against. Given that it seems like the GPL could not be enforced against driver compiled against a binary compatible API. Ie Alan Coxes recent assertion that Nvidia wasn't allowed to use a certain internal feature of the kernel should be readily made technically irrelevant, since Nvidia could create a stub binary compatible equivalent to the API to compile against, and then the user can install the driver and use it with the GPLed kernel without violating the GPL.

Is there a flaw in this reasoning or do programmers have a way to readily use GPLed APIs as closed source without violating the GPL?

My question is - do you believe that Mr. Jobs is rightfully praised as an inventive genius?

In the popular press, Steve Jobs is often praised as one of history's greatest inventors and as an inventive genius, and I feel it does a disservice to true inventive geniuses (such as Tesla) to praise talented businessmen with modest or little inventing talents as great inventors.

From my reading of the history of Apple and specific Apple product lines - his talents are primarily with business and marketing - with providing little in the way of invention or technical skill.

He of course has his names on numerous Apple patents - but this seems more as a hedge against having patents invalidated by not listing all who 'contributed' to an invention due to the risk of any feedback or comment being viewed as a contribution to the invention however minimal.

Mr. Jobs did have some true inventions to his credit - for instance using plastic cases for the Apple I, insisting on good quality bitmap fonts for the Macintosh.

Most of the inventions that the public attributes to him are primarily based on the inventive and technical talents of others (Ie the Apple line, Macintosh line, iPod, iPhone), were almost all completely developed with almost no significant invention on the behalf of Mr. Jobs (sometimes as with the case of the iPod, the products were concieved of and invented and developed almost completely outside of Apple) .

He certainly contributed by providing good user feedback (ie reducing the delay time for loading and switching for the iPod OS), but such feedback aren't inventive in nature.

While it is true that you can engineer essays to be 'bad' and still score 'good' - the question is - are there natural essays that score good but are actually bad; and good essays that score bad but are actually good.

Every analysis I've seen suggests that these algorithms do have problems with good essays that are highly creative. Essay graders also have difficulties with this kind of essay - giving drastically varied scores.

However there doesn't seem to be much evidence of other issues except when an extremely knowledgable issue deliberately trys to make the algorithm fail. Any student or other individual who can do this probably knows that material well enough to 'get an A' if they were to properly apply what they know so this seems like a non issue.

Probably failing to take into account the value of the targets compromised was the biggest flaw.

Since the average apple user will be far more profitable (apples are a luxury good and thus will have a higher percentage of wealthy users) to compromise than the average pc user, he needed to adjust the numbers downward to take that into account.

We can already tailor antibodies to particular protiens. The issue is that cancer cells, bacteria and viruses are complicated. Their populations change under selective pressure from antibodies so that the protiens that were useful targets beocome useless, etc. While we are finding highly conserved genes to target it is non trivial.

Completely agree - bad ideas are common as dirt, good ideas are extraordinarily rare. Those who say 'ideas are a dime a dozen' are usually wanting to get good ideas for cheap or free.

What is claimed by the article is that the patent should not have been granted because corporations already had hardware on the market that used the particular combination of algorithms recommended in the patent years before the patent was filed. No one had patented it because it was in fact 'patently obvious' since that is what was already being used.

Unfortunately I don't know enough about the technology to know if the claims of the author are accurate.

This wasn't a mistake, it was a deliberate misrepresentation on the behalf of whoever did the editing or ordered the editing done.

When the corporation claims 'oops' when it is obviously deliberate, it leads to distrust. There most definitely should have been firings over this.