But what is more important? A device that does everything you could possibly think of for 6 hours, using up power faster than it would have to? Or a device that uses power efficiently, does everything you want it to do, and that won't let you down when you need it?

"Previous-gen OS", you said? Compare the power and resource management capabilities of Symbian and Android. Which one is more modern?

Again, some manufacturers don't get it. If a phone's battery will last for six hours of usage when you put every single resource you can think of on it, the device has been badly designed. A phone should, at the very least, last for the whole day under the heaviest usage pattern. If the one you designed doesn't, you should remove features, add battery or improve the power management system. This is the main feature of Symbian: you can drive your system with a less powerful CPU while having the same level of responsiveness, even if you keep the CPU clocked way down for most of the time. This is a point where the EPOC kernel is leagues ahead of Linux or Mach.

You mean "iOS and Android-based smartphones already last less than a day under any kind of usage", right? Because I rarely see Blackberry and Symbian users complaining about their battery life, and when they do it's something like "I need to top off the battery on this phone every night, because if I don't I'll have to control my usage on the second day". My Symbian smartphone gets five days out of its battery under regular usage, maybe three if I go wild with the radios and games. A phone whose battery doesn't last through the day has a clear design flaw, and should be kicked out of the market ASAP.

The way I see it, 12 hours of battery on your phone is as ridiculous as 30 minutes of battery on your laptop. The whole idea of a cell phone is to keep you available all day long, just like the idea of a laptop is to have computing power on the go. If you have to top off your phone after a few hours, it's doing something wrong, just like it's just plain wrong if you can't even read your emails on your laptop without being plugged.

It's completely possible to have a phone that can do everything* you said and still have a battery that won't let you down, while keeping a thin profile (the Nokia N8 is a great example). The manufacturer just needs to understand that a 1GHz CPU is overkill for our current battery technology, and that an OS should save the battery as much as possible. That's it. Some manufacturers get that (Nokia, RIM), while others don't (Samsung, HTC, Apple).

*=excluding the parts that no phone does yet, obviously.

Slackware gets this right on their -current rolling branch (or at least they used to get it right). Keep the system as stable as possible, and give users the most recent user-facing software. Heck, they are still using LILO: if it ain't broke, don't fix it!

Notice that the release notes for the 96.43.19 version, released a week ago, includes "support for X.org xserver 1.8 & 1.9". Yes, xserver 1.8 wasn't supported until a week ago. It was released in April, seven months ago. In other words, if you wanted to run a current xorg for over half of this year with a Geforce 4 and the binary blob, you were out of luck; and their official position (which they fortunately reverted) was "we still support those cards, but only on xserver1.7.999". This is not "remarkable support" at all.

This is true for every console, IMHO, and specially for the ones that sell the most in their generation. Look at the PS2: it had thousands of crapware/shovelware titles, and a few definitely great games. Same for the PS1, SNES, etc. Maybe the percentage of great games is lower on the Wii, but you can't seriously believe that there are more good games than bad games in any console.

Actually, "bobolhando" is a joke site. And, if you go to the "source" on the blog (the Twitter feed of the blog's main author), you will see:

Amigos, trollar o Guardian, quem já conseguiu? http://bit.ly/aiVTeW Texto do @Thiago__Caetano no @bobolhando sim

Roughly translated to:

Friends, trolling The Guardian: who has done it? http://bit.ly/aiVTeW @Thiago__Caetano's text at @bobolhando did

The interaction length decreases when the energy increases (or: solar neutrinos, with a MeV energy, can cross light-years of lead, while extragalactic neutrinos, with energies up to 10^10 GeV, are completely blocked by just a couple of km of rock). If the LHC managed to create a 7TeV beam, that energy would be enough to lose some percent of the beam due to the interaction with the Earth. "Some percent" here is somewhere between 1% and 10%, assuming a distance of the order of 10000km from Cern to FNAL. I can't find my masters thesis to give you a more exact number, but be sure: a ~TeV beam wouldn't travel through the Earth and end completely unspoiled. The interaction cross-section grows with something like E^2, and the attenuation of the beam grows with exp(cross-section). The Earth isn't completely opaque to 7 TeV neutrinos, but the attenuation surely does happen.

Besides, the most important point is that pointing a 7 TeV beam from Cern to FNAL wouldn't do anything to confirm or deny this effect. You'd need a much higher energy beam, which cannot be built now; and even if it could, the attenuation in this energy range is very high. Specially for electron antineutrinos (which are exactly the signal they measure), due to a coherent "annihilation" with the electrons. There are other possibilities to (re)confirm this effect using Cern's accelerators, but a Cern-FNAL beam would be useless.

That's not a measurement of stability. We know there have to be four neutrinos (if this result is correct, of course) because, in oscillations, you get something that depends on the difference of mass between two different neutrino states. If there are three neutrinos, there are two independent mass differences (neutrino2-neutrino1 and neutrino3-neutrino2; if you know those two, you know what is neutrino3-neutrino1). All experiments except this one (and LSND) are compatible with this. Now, MiniBooNE saw a mass difference that's completely different from the others. Imagine the previous measured differences were 1 and 10; they measure 1000. That's impossible to fit with only three neutrinos, and shows that there has to be another one. Notice that we'll never detect it: it has to be sterile, or in other words it can't interact with mater. That's a very clear limit from the LEP (the big Cern experiment before the LHC).

Neutrino decay isn't 100% excluded as a secondary effect; but the indirect result of the existence of a fourth neutrino has nothing to do with it.

At 7 TeV you'd have some attenuation of the beam (or some percentage of neutrinos absorbed between the creation and detection); and to confirm the same effect at a distance 20000 times bigger, you'd need ~20000 times more energy than MiniBooNE, so that would be 160 TeV (or "quite much more than the LHC"). What the LHC could do is to make a lower energy antineutrino beam and direct it to other European lab, with an energy such that the distance to the lab divided by the beam energy is around the same as MiniBooNE. That would give us a third confirmation of the same effect (after LSND and MiniBooNE), with much higher statistics.

Actually, it is. Kubuntu gets sidelined by the development of the Gnome/Unity Ubuntu, where most tools are developed, and ends up looking, feeling, and working like a mess.

Yeah, what about that? A distro that is just (K?)Ubuntu with a couple of extra repositories and different default settings is basically the same thing as (K?)Ubuntu, but with different default settings and applications? Who would have imagined?

Sorry, but any experience with RH5/Slack 1.x is meaningless in deciding about distros today. You can't decide against Windows 7 because Windows 3.1 sucked, or against OS X because System 7 had no real multitasking. And I've fed the troll too much already to keep going, but your post is pretty funny.

Skills needed to maintain a good webpage != skills needed to maintain a good distro.

Putting it in another way, if the site is awful and they still manage to stay alive and relevant for 17 years, they must be doing something right.

So you test your script offline? You know, exactly like you test the changes you will do through a GUI in an offline server before going to the live one?

I can factor any prime number very easily, if I know it's prime before starting. And it's fast. It takes only the time needed to write "1" and the number itself.

If you are going to do a bunch of administrative tasks in a row, it's more effective to just "su -" than to prefix 30 commands with "sudo".