Current system: professor gives lectures to a group of students, and the students can ask for clarification as he continues. Students requiring more in-depth explanation can utilize the professor's office hours. Students are given small activities to do on their own initiative to help them discover holes in their learning and to test their progress.

Your system: if a student doesn't understand a lecture, he's screwed. Classtime problems don't help him at all because -- according to you -- if he doesn't already know the material he should be failed for not watching the lectures. Office hours are completely redundant. Classes separate into two groups -- people who understood the lectures and therefore don't need class time (why have the professor?) and people who can't learn the material by watching youtube and are therefore failed (why have the professor?).

We've about mined this out -- our experiences have been wildly different -- but I wanted to say that I was in a ferocious temper earlier and probably came on a bit too strong ("borderline dishonest" and aggressive crap like that). I apologize if I sounded harsh and appreciate your patience in discussing this civilly.

So what I see is Linux big where there's big iron. Our clusters, like I said, wouldn't exist without Linux, and as you say the server market is huge (one step shy of homogeneous). I see desktop problems frequently, though -- regarding software upgrades (and the more I think about it, the more I think this is a deep problem) even having to add a repo for a PPA is an order of magnitude more complex than simply installing new software without thinking about it.

Such things aren't unheard of in MS land -- the DirectX issues, as you point out, and so forth. But all of those decisions have been met with huge resistance, and I would posit that 99% of the software on the average Windows desktop user's machine can be upgraded simply by getting the new version for a good ten years afterwards, whereas in Linux if you want feature upgrades your options are all going to require a little reading. Again, project cars.

None of this is really meant as a criticism of Linux, just my own opinion that it's not really competitive in the desktop space. NTTAWWT, because it can fly in the right hands -- it's not really needed in the desktop space. Hell, the megacorps want everyone to have thin clients again, anyway. *grin*

(sigh) Okay, we can do point by point.

Consumers -- the people using the software -- are comparing desktops to desktops. It doesn't matter why your brakes failed.

I've literally never, not once, in my experience with my own machines, my co-workers' machines, or my friends' machines seen a Windows box be brought down by a video driver upgrade -- ATI, nVidia, whatever. But everyone I know who uses Linux (myself included) has found themselves doing "lynx google.com" after a video driver upgrade.

This is a more complex issue. See, I don't have to upgrade Windows to get newer versions of software. Modulo backporting trickery, that's not true of Ubuntu -- if I want a newer GCC or whatever I either a) build from source, destabilizing my packages or b) do a release upgrade, breaking my machine.

This is completely false. I have contractual obligations not to talk about the hardware too much (and yes, I see where that's leading *rolls eyes*) but the machines are almost completely heterogeneous -- no two share the same processor, there are a variety of both ATI and nVidia cards, etc. Failures were rampant. We suspect it was because of a regression with the new dm's compatibility with LDAP'd network shared home directories. Complete showstopper with no permanent fix that doesn't cripple parts of the system.

Funny, the guys running Windows upstairs don't need a service history (to a rough approximation). Are you listening to yourself? You're seriously suggesting it's okay to have to document getting a consumer machine running. We document install procedures on supercluster nodes with exotic hardware. That shouldn't be necessary for desktops with commodity hardware.

Well, yeah, that completely agrees with what I said, because Microsoft has been paying thousands of developers for decades -- there are billions of dollars and hundreds of thousands of man-years of experience invested in Windows. Some of that is on the same 'fun' aspects of development that FOSS has taken care of in the development of Linux. Others are covered by the sheer competence of the people drawn to Linux development. But a lot of it isn't covered at all, and Canonical isn't even a drop in that bucket. The last 10% takes 50% or more of the work, and Canonical doesn't have anything like the funds to pay for that -- it's borderline dishonest to suggest otherwise.

His anti-Linux arguments are correct.

I do research at a large university, and do a bit of support with the IT guys (we've got too many brains and not enough elbow grease). The recent Ubuntu upgrades hosed about ten systems to the point where they wouldn't boot. I remember a Windows XP (IIRC) update years ago that had a chance of making a tiny percent of computers not boot, and people here went absolutely wild. But we're talking about something like a 70% failure rate with the Ubuntu upgrades. In addition, "backing up /home and reinstalling" is not the simple process you make it out to be; you're taking on every problem you ever had getting the box to run correctly.

Ubuntu will never have the funding to pay people to fix the "last mile" of problems, because those aren't sexy. New interfaces, funky new object models, all that stuff is fun -- but no one is going to volunteer to fix the interaction issues and corner cases without a competitive salary, and it's going to take a lot of those guys to polish it. Never happen without some kind of fundamental shift in the market (a couple gens of Windows 8 style products might do it).

In my field (HPC) we use Linux because it blows Windows out of the water when you need to tune up a serious distributed cluster. As a result I'm familiar and comfortable with Linux and can really make it scream, so I use it on most of my boxes, and there's a lot about it that I really like -- compare it to a mechanic's project car. But I keep a Windows 7 machine at home for games, and while every Linux box I've had has required forum searches and post-update repairs, the Win7 machine has remained stable through a (gigantic and no doubt inefficient) list of updates -- just like that mechanic doesn't drive that project car to work, because he knows he'll strand himself eventually.

This is really interesting stuff -- thanks for the info!

A "standard 101-key keyboard" doesn't have lower-case and upper-case letters. 26 base letters, 26 uppercase, 10 numbers + 16 punctuation, 26 function keys, and 26 control keys can all be done on 5 chorded buttons (using 27-31 as prefixes).

I'm not really feeling the outrage, here. So you can tell I have kids and drink a lot? Shit, my wife could tell you that.

(Disclaimer: don't actually have kids or drink, and we're not married yet)

And, immediately, there will be a market for insurers that don't.

RAIDs are not substitutes for backups -- only modifications of uptime.

You're catching a lot of flak for this, but I agree with you 100%. You are correct that this is an uncomfortable trend, and I think absurdity should be able to stand on its own. Good call.

Don't get me wrong, bonch, I'd slap a disclaimer on every one of your posts with a link to your history *grin* but to each their own, and for the moment we're in the same boat.

Just wanted to toss a little rationality in here.

How, then, do you account for systems which behave differently under observation? As a simple and recently high-profile example, what about the Quantum Zeno effect? If observation does not motivate quantum coupling, then the results are inexplicable, and still so under the model of "sub-Plank discrete higher-dimensional fundamental particles". I suppose what I'm saying is the problems are orthogonal, except that the observation problems can be used as a starting point to further develop a theory of quantum interactions -- the one we have, coincidentally -- whereas making gross assumptions about the invalidity of current field and entropic models appears to add nothing to the discussion.

Also, in every single one of your responses to every one of the people attempting discussion of this point, you have responded with derision, insults, condescension, and disdain for the life's work of hundreds of PhD holders, each of which would love nothing more than to be the one to publish a paper formulating a model obviating the need for current quantum "weirdness." In light of that, I must confess to some curiosity regarding your own credentials, since you appear to hold everyone else's in such disdain.

You've stated over and over that "every observation can be easier explained" if you assume that "at small scales time can go in either or both directions."

I am asking you to elaborate. What energy is expended by this transition? If time is moving in "reverse" or in "both directions" how does that affect entropy? What particles have this property, what determines which ones may "travel through time," and what bounds the distance they may travel? How does that remove wavicles? Just as an extremely simple example, how does that explain the results of the double-slit experiment?

In your scenario, the guy who sells hardware without the "circuitry complex enough that you need a Scanning Electron Microscope" and the "kill switch" and the etc. etc. is gonig to be the successful one, because he'll save a ton of money not trying to play walls and ladders with hackers.

If Big Media kills capitalism enough through legislation that such hardware is made in not a single place in the US, then it will be made elsewhere. If they then dictate that the US will not do business elsewhere, then the US will cease to be relevant in the global economy and will be replaced by someone that doesn't waste money on putting increasingly-complicated locks on (only one of) their doors.

Like with all the previous technological advancements, the cat is out of the bag. The people who learn how to work with the new environment will, as always, outpace those who spend all their time chasing cats.

This is a gross misunderstanding of the Schrodinger's cat thought experiment, and something of a fallacious presentation of it.

I don't think there was ever any doubt that a cat locked in a box for a sufficient length of time would expire. That is neither in doubt nor interesting.

The formulation deals with the status of a cat in a box present with some measuring apparatus capable of detecting decay of some isotope, linked to a sealed capsule of some poison, in a sealed container with a cat. Supposing the isotope has a roughly 50% chance of decaying in the next five minutes, and iff it decays the poison is released (killing the cat), after five minutes is the cat alive or dead?

The "collapse the waveform pseudo-science b***s***" here is simply translating the simultaneous probabilistic states into a single actual one. The reason this is relevant is in quantum mechanics there are real, measurable effects that occur as a result of the probabilistic waveform that differ from the effects of the collapsed state -- once you know whether the cat is alive or dead, in other words, you have a fundamentally different system than before it was observed.

No, the statement from Apple with the list "suggestions" for not infringing their design patents was in California. The statement makes the argument, further, that elements such as thinness and rounded corners in a portable device are not "functional" and hence patentable.

Apple also tried to ban the 10.1N, so they clearly thought that was not enough.