No, it refers to controlled experiments run and managed in accordance with established best practice. On the other hand, you ARE guilty of dismissing studies based on conclusions, rather than methodology.

Some of the studies have focused on self-identified electrosensitives.

Now let me talk about MY electro-sensitivity. There is a high-voltage power line that crosses the motorway on the route between my childhood home and where my grandparents used to live. When we went under it, I used to get a funny feeling in the top of my head -- every time, without fail. So someone if my family (I can't remember who) suggested that I shut my eyes and tell them when I felt it. For a year or two, I kept opening my eyes too early and seeing the powerlines before we went under them. So one year I got determined to do it properly. I closed my eyes as soon as we reached the first bend on the motorway and kept them shut for ten minutes or more. No sensation. Ever since then, I have felt nothing whatsoever when passing under the lines. I couldn't even make myself conjure up or relive the sensation.

Not to mention "desk space"...

Yes, but it gets messy and it's often beyond the creator's ability. There's lots of MOOCs out there full of ugly patched over words, or a subtitle annotation saying "Oops! I'm talking nonsense." Editing the videos would be easier if they had been planned and filmed in segments and edited tigether in the first place, but most of them just appear to be video-blogger style pieces to camera wie the visuals/slides/code-sessions recorded live with the content.

They're not -- they're just designed for a very specific course style: the one used in Andrew Ng's Machine Learning course. It works quite well for that partly because he's dealing with a complex subject that means students need a lot of input and explanation before attempting tasks, and partly because of Ng's exemplary skills as a lecturer.

Most programming courses, on the other hand, involve lots of tiny incremental tasks that need minimal explanation, followed by several practise tasks to internalise the logic. Programming MOOCs that I've tried tend to fall into one of two camps: either they give you lots and lots of input in the videos so that you end up forgetting it all before you get to the practise tasks, or they stop the video with a message telling you to go away and try out the exercise in another window and then come back. Neither of these are efficient workflows, but the main platforms in general, and Coursera in particular, all push things in that direction.

I'm not a huge fan of MOOCs, and prefer the model used by w3schools -- webpages with links to tasks in a live coding environment as a pop-up. I wouldn't mind videos being embedded into such pages where it makes sen to do so, but most MOOCs overuse and underprepare their videos.

When someone makes a written mistake in a slidecast/pencast-type video, there's no going back to correct it. To me that renders a lot of MOOCs frustratingly confusing. It also means that all their talk about gathering data "to improve the courses" is nonsense, because it means they've got no way to A/B test individual small changes to the course, so they have no proper comparisons whatsoever.

If it's an emulator, why do they need to authorise? Sounds like a recompilation job to me.

The entire history of human semiotics is one of multiple redundancy. Look at numbers: one, two, three, four, five, six, seven, eight ... we still haven't repeated any vowel sounds, and we've had a fair mix of consonants too. The differences are bigger than necessary for conveying unique information, even taking into account every single part of information theory. why? Because the brain isn't pure maths, and redundant information can often speed things up as two inputs can reinforce each other.

I never could understand how so many people don't recognise that colon is a start-block token. "It's the end of a command that takes a block" is semantic wriggling.

When I started reading, I thought it might be about the way the manufacturers keep releasing new cards to rebalance the game. An episode of Extra Credits on YouTube talked about how they constantly fiddled with the game so that there was always a new potential super-tactic to learn, but after a while it would no longer be quite so super, hence the need yo keep playing and keep learning. The way the guy was waxing lyrical about it, I'm assuming no-one else has an algorithm anywhere near capable of copying them. If you made an AI that copied that part of what they do, there would be several major customers for it -- not just the niche card makers, but also the major MMO makers.

The AI will run the numbers and realise that with Wizards of the Coast's operating profit, the only way to win is not to play...

As a native English speaker, I couldn't tell you the difference between a street and a road. The simplest distinction is pavements for pedestrians, but some roads have them too. What I'm saying is that all these sorts of categorisations are ambiguous.

He asked "did", not "does". I nearly did in 98, but thought better of it.

It's probably down to the way opcodes exploit common functionality between opcodes and use parts of the opcode to trigger certain logical blocks in the chip. That means a lot of the glue logic will have to be the same, if you want the chip to run efficiently.

I agree completely - as I said, test banks are supposed to be databases of the teachers' own questions.

On a similar topic, I'm pretty appalled at the lack of time and attention given to numerical values in a lot of modern maths texts. My lecturers at uni always went out of their way to construct problems that has numbers manageable by mental arithmetic, right up to final step of solving for x -- and typically we weren't asked to do that final step. They took pride in making sure that the task tested our knowledge of specific mathematical concepts, and not on managing to keep all the significant digits throughout. No 3.753579x, but plenty of 5x/4 etc. This was questions just for their classes,mincluding test questions only ever answered by 50-150 students, and yet people writing books to be used by thousands of students across the world year after year can't be bothered to do it properly.

I'm not complaining about the video, just about people's reactions to it. It's an interesting tech demo, but all this hyperbole is as yet unwarranted.