While signalling is no doubt important, I think its influence on undergraduates is oversold. Unless students are a great deal more aware (and watching my sons enter into university gives me no reason to believe so), selection tends to be based on degree requirements, interest, and aspiration. However, I'd agree that my experience was that there was next to no correlation between "cost of providing the course" and the "worth of the course to me on any given metric". And lastly as with regards human capital formation - I believe that if you tested me a few years out of school, most my knowledge was either seemingly useless or forgotten. Yet being able to use the "useless" information to quickly learn relevant material, or being able remember what I've "forgotten" have both proven themselves invaluable over a 30 year career, especially every time my career has taken a right turn, and I've theoretically started from ground zero.

TANSTAAFL. Given that almost every optimization of a highly-developed system is a trade-off, the question could be rephrased to "what do we sacrifice in order to get better battery life?" Personally, since I almost never use laptops unplugged, I'd prefer to see Windows maintain it's focus towards plugged-in usage. The fact that you *can't* optimize for everything is why choice is good.

I think you want the right tool for the job, and depending on parameters of the discussion, either tree or flat might be appropriate. For example: - Want to "encourage" posts to stay on topic - flat - Want to emulate an interesting dinner party, where discussions inevitably devolve into half a dozen separate sub-discussions - tree - Want to keep replies to less than 200? - flat - Want to encourage participation from as many as possible - tree - Want to make it easy for casual/marginally interested browsers - flat etc., etc. Hence I like flat (Stack Overflow is a very nice model) for technical discussions, tree for political discussions, etc.

I'm confused. Isn't he supersaurus implicitly agreeing with you?

To those who want to point out that sometimes (many times) there is no one to blame, realize that you're fighting against one base fact: human beings are not equipped to handle reality. Luckily, our brains have built-in defenses to prevent too much reality from overwhelming and disabling us. (In fact, the clinically depressed often have far more accurate assessments of control and even personal abilities, characteristics and how they are seen by others - unfortunately, the realization tends to make them dysfunctional). Anyway, in summary, you have to expect humans to be human and invent cause and effect even when there is none - it's what keeps us sane. However, just because it's natural doesn't mean that society shouldn't recognize that fact and protect itself from the costs that such misattribution may have. In a case such as described above, I'm not certain why the vast personal costs of the death were allowed to be made explicit to the jury, given that it's immaterial to whether there was malpractice. A lawyers job is to win the case for the client, so I understand why he did his best to feed the same "someone must be to blame" tendency in the jury - but why did the judge allow it?

@Tony Hursh: by and large, "publishers" don't do any of those things. They hire them done. What's the difference? Either they're doing it in-house or they're contracting out, and managing to ensure it's done right. Either way, it's paying out. A paper book is an investment of perhaps $20+K on the part of the publisher with very little guarantee that it will pay out. Jeff's book cost him time with very little return, as is typical for most books. Think about how he'd feel if he'd been the publisher and lost $20-40K... Now, the mechanics for e-books are quite different. There, the problem is that if you aren't already "famous" (for whatever value of famous is relevant to your book), you're probably looking at double digit sales of your e-book for presumably more work (writing the book as you would for the paper version + management) plus whatever it cost you for cover design, line editing, copy-editing, etc. Distribution is irrelevant for ebooks. You can upload a book to the Kindle store, the Nook store, and the iBookstore in 15 minutes. Gaah. I *hope* it's taking you 5-10 times longer and you're ensuring that each type of download works on a variety of platforms. Speaking from personal experience, if you don't buy your own book from each platform and try it on both Mac & PC & tablets, you're risking a horrible experience for your customers. On behalf of e-book readers, please consider some level of professionalism. If your book isn't worth *you* spending a few thousand dollars of your time (and perhaps money), why inflict it on the rest of us? WRT author advances, that's always struck me as being akin to a super-skeevy payday loan scheme -- one that demands the bulk of your paycheck in perpetuity. BIG difference. If your book doesn't pay out, you're not on the hook for the rest of the advance. A publishing deal is a co-investment. The publisher is investing the thousands and you are investing your time. In this world, money is worth a lot more than time. Given that publishers aren't making money hand over fist, I'd say that it likely that the balance is roughly right. (And yes, there are terrible publishers who do rip off authors. There are also authors who take the advance, but never produce a book. Neither define the industry.)

This has nothing to do with whether technical paper books are still viable (although for non-references, I still prefer something I can browse at bed-time), but if you putting out a printed book, the publishers earn their keep. I want most of that money to go to the author, not the publishing middlemen. I have to say, this statement rankles me. The "middleman" is the one responsible for editing, typesetting, covers, cover quotes, publicity, selling into bookstores, accounting, printing, legal, distribution, and, of course, returns. More to the point, the author fronts his time, it's the publisher that fronts the 20-50K to actually get the book out the door and into hands of buyers that is taking the real risk. As a lover of books, you slight the people who make actual books possible.

Why in the name of peter can't I buy the cursed books? it is because they are afraid I will build a bomb with it Luis Robles, you cannot buy the book because Amazon does not have the right to sell it to you. When a book is sold, it's usually sold to a publisher with limited rights, specifically the right to sell it in a given geographical location. It is quite possible, and even likely, that you are trying to purchase the book from the US publisher, who could be sued if he allowed it to be sold (through Amazon) outside the US. As a Canadian, I sympathize. Often Canadian and US rights are bundled, but its not uncommon for them to be divided, and then worse, for the author not to find a Canadian publisher or for the Canadian publisher not to be into e-publishing yet. At that point, there's no legal means for us to buy the e-book. (At least the printed book and can be bought in the US and then physically carried across the border...)

Two points: E-pricing vs. book pricing: This is mostly an Amazon phenomena. Amazon discount printed books hugely. They are often priced at a loss, making the print version pricing very attractive, while producing the illusion that quality books can actually be produced very cheaply. E-books, thanks to the agency agreement, are priced by the publishers and set at about what the publishers need to stay in business (remember, if the book publishers aren't making a decent profit, they'll be closed down by their owners and the money invested in something that does have a decent return. Owners don't keep publishers in business as an act of charity.) Expecting that you can somehow have quality for really cheap is a pipe-dream, although it may be one that we destroy the publishing industry to try to achieve. Bad formatting, etc in e-books: Decently and properly formatting documents is *expensive*. You can't hire recent college graduates who desperate to work with books at tiny wages to do decent electronic formatting. Which mean e-book prices need to be higher to support the expense. However, the publishers are no dummies. They know that people go ballistic over current e-book prices let alone one's increased to pay for the cost of decent formatting. Moreover, consumers get mad about pricing far more than they get mad at shoddy formatting. So, in the end, publishers give the public as close to what the public wants without actually going bankrupt, and we all get badly formatted books. I'd be more sympathetic if the publishers were rolling in money, but outside of extremely isolated sectors, they aren't. And that's while they're paying their employees as little as they can get away with. Anybody who expects that you can cut the amount of money going to publishers by 2/3rds and not have it make a fundamental change to the quality of books is dreaming in technicolor. Costs of Goods are significant, but electronic distribution is a *long* way from cost-free, as anyone in the business can tell you.

I think other commenters are pointing out the true problem. The presenter has produced a tremendous tool for an extremely domain specific problem. Unfortunately, the vast majority of real programs have so many degrees of freedom that you couldn't possibly generalize a system like this in any meaningful way. Even worse, the outputs of many programs don't lend themselves to easy visualization. When I am working on a program, a chance in code *could* effect 30-40 different output files, change the network communication of 2-3 different connections, etc. There simply isn't a way for a human to visualize that all at once. The aphorism is "make things as simple as possible, but no simpler". The trouble with programming at any real scale is that *it isn't simple*, and nothing can make it simple. My experience is that many general systems designed to make life easier, make simple tasks really simple, moderate tasks slightly easier, and complex tasks anywhere from a little bit to a lot harder. I've used good tools for complex problems, but they were all extremely specific to the project. I have yet to find good global tools that make general programming fundamentally easier.