I read this as a VERY carefully worded line that rather than saying "the NSA is actually pretty reasonable" really says "if you think what we're doing in the US is bad, you should see what we're doing overseas." It practically comes out and says that they're doing all of those things "outside" the US borders. He also implies that all of the metadata collection that is done domestically is just fine.

Based on this, I would suspect that some program that the NSA agrees costs more that the intelligence gathered is worth is going to be cut, but overall nothing is going to change.

Sometimes (I'll admit this is the exception rather than the rule) simplifying provides a clearer picture of something specific that the original person stated and that the person wants to address.

Substituting "consensual" for "as long as no one is coerced" is equivalent in my book, and otherwise there is nothing in your response to imply that GP's statement is an inaccurate portrayal of Stallman's views. If anything, using the word "consensual" was perhaps being too generous towards Stallman, as it could be argued that "as long as no one is coerced" is actually a LOWER bar to meet than "consensual."

Based on a cursory search of the internet, it appears that while RMS pays lip-service to the real power issues that exist in adult-child relationships, I don't think he really understands the depth to which power in such relationships is centered around the adult. Furthermore, it appears that he doesn't realize how young of children are abused. RMS seems to really be thinking people who are at the boundary of adulthood, whereas many of the abused are so far from adulthood that there is no question that they CANNOT provide any meaningful form of consent. It's not clear to me if Stallman believes that young children could ever meet his standard of not being coerced.

Both GP and I are addressing a specific portion of Stallman's statement because that part is cloaked in a number of areas on which reasonable individuals could respectfully disagree. That specific portion shows either extreme ignorance or a frightful lack of judgement. It is difficult to tell which.

My experience has been that X is broken on all fronts, and, while I don't understand the internals, I completely believe that something new is required to really fix them.

X is overkill for local display because it has to carry along a bunch of logic to pretend like it's displaying over a network. This results in significantly degraded local performance compared to what would be possible with a more optimized solution. A lot of workarounds have been put in so that this isn't as bad as it was, but it's still less than ideal.

X is bad for modern remote display because it doesn't tolerate intermittent connections. If your client-server connection goes down all your applications die, which makes it useless for viewing remote applications from a laptop over wireless.

Even in cases where the connection is persistent, X is dog-slow over high-latency connections. Firefox rendering over X from a server a time zone away is unusable. My understanding is that this is because, at least for naive client implementations, too much communication between the client and server is required beyond simple display data.

Based on the above, I can think of NO case where X11 provides a GOOD solution to modern display needs. The experts appear to agree with me. My naive expectation would be that the right solution is to design a new protocol that works wicked-fast locally and that has hooks to allow its output to be displayed (using a separate protocol) over an intermittent network connection. Wayland at an extremely high level seems to match my naive expectation.

Let me preface this by saying that, as you would expect, I have done no detailed research or cost analysis of rural or low-income phone subsidies.

That said, I can come up with a number of reasons why rural and low-income phone subsidies could turn out to be a good idea, and could end up being a net positive (or at least a net neutral) economically.

For rural subsidies, the most obvious expense saved is fire prevention. Subsidizing phone service for rural areas can ensure that both forest and house fires in rural areas are reported faster, resulting in reduced loss of property and reduced government costs associated with fighting those fires. This alone may be enough to justify subsidizing rural phone service. There are likely other advantages to subsidizing rural phone service as well.

For low-income subsidies, the case is a little harder to make. The best argument I can make is that low-income subsidies may result in higher employment levels. It can be awfully hard to get (and sometimes to keep) a job if you don't have a phone. The government gets to collect income taxes from those who work, so they have a financial interest in creating maximum employment. Again, there are likely other advantages to low-income phone subsidies.

In both cases, I think everyone can agree that it decreases the value of these programs when individuals exploit the system in ways it was not intended, particularly if those people break the rules when doing so.

Realistically, many times programs like these (particularly the low-income one) really just provide an incentive for people to do things that they should already be doing. Since the government is unable to tell people "you MUST pay for phone service," (grumble... individual mandate... grumble) the subsidies may be a cost-effective way to encourage people to make decisions that they might not otherwise make (like paying for phone service) that ultimately save the government money in other areas.

Again, this is a complex area with a lot of interconnected and non-obvious consequences. It could go either way. Without a good study it's impossible to know. Finding out would require a study of the cost-effectiveness of rural and low-income phone subsidies which would, predictably, also likely have to be paid for by the government. :-)

While I generally agree with what you are saying, I feel that your statement here needs correction.

The constitution both allows laws and bans laws. The concept of enumerated powers in the Tenth Amendment says that the constitution gives certain powers to the national government, and that all other powers are "reserved to the States respectively, or to the people." The Supreme Court has given the federal government a tremendous amount of leeway in interpretation of the Commerce Clause, which is the constitutional basis of the modern federal government.

It's possible that those in the Tea Party really would like to go back to a government where the Commerce Clause has a substantially less expansive role, but that would require DRASTIC changes to the federal government. Agencies that we rely on every day (think USDA and FDA) would have to be shuddered, or their roles drastically diminished, as they would no longer be able to regulate commerce within a state. The role played by those agencies would have to be performed at a state level. State tax rates would have to rise dramatically to cover their significantly expanded role in monitoring intrastate commerce. Furthermore, this would produce a tremendous amount of jurisdictional challenges as we would have to re-litigate what qualified as inter-state and what didn't, and even once the rules were established a lot of court cases would revolve around determining what side of the inter-state/intra-state line a particular behavior fell on.

As you noted, the constitution also contains a number of important safeguards that prevent certain types of laws being passed, both by the federal government and by the states (via the Fourteenth Amendment.)

Timing... for developers.

You want to get your 64-bit processor out the door so that people who make apps that might benefit from more than 4gb of memory can start to write their apps for 64-bit BEFORE you actually start shipping phones with more than 4gb of memory. This allows them time to convert to 64-bit without being rushed into it. It also gives your OS developers time to get the 64-bit OS out the door. If the 64-bit OS isn't ready when you ship the product, you release with a 32-bit OS and you just don't advertise the 64-bit feature. (Or you say "64-bit ready" or something like that and promise the next OS release will bring 64-bit to existing phones.

In short, as a consumer, you don't care... yet. You want the 64-bit in a year or two when you have 8 gigs of memory in your phone. In order to have applications for that 8-gig phone, you want Apple to release a 64-bit phone now, so that developers will be ready with 64-bit applications to put on that 8-gig phone.

The other aspect here is that most architectures tend to clean things up when they move to 64-bit, and ARM is no exception. Some of those architectural changes that come with 64-bit will be more valuable sooner, and could translate to performance boosts right now on some applications that switch to the 64-bit architecture.

The problem is that now, with this ruling in hand, anyone who texted someone near when they were in an accident is likely to get sued. They will then have to prove that they did not know that the person was driving. This is a civil suit, and I believe preponderance of evidence is the standard, not reasonable doubt.

But, what if the sender *should* have known. A very good example of this is above, where someone pointed out that if my roommate recently left to go to the store, and I realize that we forgot to put Orange Juice on the list, I could be liable for texting them to that effect. A lawyer would argue that I *should* have known they were driving if not enough time has passed for them to actually reach the store.

Basically, this is a great big gift to lawyers in New Jersey. If someone texted a driver involved in an accident, they will now be sued along with the driver. On the other hand, if I texted someone around the time that they were in an accident, I would now need to call a defense lawyer.

Can I invent a scenario where the sender who is not the driver *should* be liable? Yes, I can.
If I am a judge, should I mention that hypothetical in a ruling unless I already have that case before me? NO. Why? because even insinuating that such a situation exists is just inviting lawyers to vigorously try to determine just what the limits are for that type of liability, at the expense of what I'm sure will be many innocent individuals. Sometimes as a judge what you don't say or rule on is as important as what you do. This is a case where ruling on something you didn't have to does more harm than good.

In this case, the judge should have said something like "If I, for the sake of argument, accept the Plaintiff's assertion that defendant could be liable, they would still fail in THIS case because... ... therefore I do not have to rule on the more general question at this time."

Did you even read the rest of my comment after the first line?

All of the cases you describe are cases where my "Plaintiff pays" is the same as "loser pays."

For the cases where they differ, your only response to this was "I don't have any assets." However, some of us do, and would like to keep them. I think it is fair to set the balance of risks against the person bringing suit, thus discouraging frivolous lawsuits.

Basically, if a defendant is accused of stealing your stand mixer, and the Plaintiff sues them about it. Assume legal costs are $5000, which is incredibly low.

US system, Plaintiff wins: Plaintiff out $4800, Defendant out $5200
US system, Defendant wins: Plaintiff out $5000, Defendant out $5000

Loser Pays, Plaintiff wins: Plaintiff up $200, Defendant out $10000
Loser Pays, Defendant wins: Plaintiff out $10000, Defendant even

Plaintiff Pays, Plaintiff wins: Plaintiff out $4800, Defendant out $5000
Plaintiff Pays, Defendant wins: Plaintiff out $10000, Defendant even

Plaintiff Pays is the only system that actively discourages frivolous lawsuits by making the Plaintiff lose in most frivolous cases. In both the US and Plaintiff Pays systems, they can convert to loser pays but this has to be litigated.

Loser pays has serious problems because then if a massive corporation pays a lawyer a million dollars to sue someone for a thousand dollars, the defendant is in effect actually being sued for a million dollars.

The rule should be what I would call "Plaintiff pays."

1. If the Plaintiff wins, both parties pay their own legal costs, unless the court decides to award the Plaintiff's legal fees as part of damages.
2. If the Plaintiff loses, the Plaintiff has to pay both parties' legal costs.

The result of the above combination is to encourage Plaintiffs to settle out of court, which is really ideal. It disproportionately discourages suing large corporations (as they would have larger defense legal bills), but at least gives truly innocent defendants back the ability to choose to defend themselves.

In the context of phones, the MediaTek has all high-performance cores. Not high performance by desktop standards, but relatively high performance. To be clearer: In the context of general-purpose consumer computing (which includes phone CPUs) there is currently little advantage to having more than 4 cores that have essentially the same power and performance characteristics.

The advantage of big.LITTLE is that you pair a set of high-performance/high-power CPUs with a set of low-performance/low-power CPUs, and in theory get the best of both worlds. It's not clear if this is true in practice, unfortunately.

It's not quite that simple.

All things being equal, if I could build 1 core with 2x the performance at 2x the active power usage, I would build that rather than building 2 cores at 1x performance and 1x power usage each. The single core would offer more system performance and as a result less power usage for the same workload.

The problem is, we've gotten really good at making cores, and now building a machine with 2x performance will take more like (hypothetically) 4x active power usage. This is true across the industry, from servers all the way to mobile phones. Two cores with 1x performance each is 2x system power for 1.9x system performance in some workloads and only 1x system performance in others.

The problem is, the more cores you have, the less likely you are to get additional performance from them. Right now, 4 cores is about the most that PC games are able to utilize, and I would suspect that for phones the same is true. 8 high-performance cores really is overkill for a phone. 4 high-performance and 4 low-performance cores (bit.LITTLE) is a completely different thing, and can make sense because you can get the high-performance cores into even lower power modes.

In the short-term, I actually think the QUALCOMM guy is generally right. 8 high-performance cores just doesn't make sense right now, as hardly any software that runs on a phone will be able to make use of that many cores. The high-end phone market turns out new products so fast that we need to be closer to having software support for 8 cores before we start wasting limited resources by adding cores that software won't be able to use. I'm fairly confident that there are better uses for that die area than for additional cores.

This really is it, but you don't even have to go that far. Suppose I am a politician that is accused of a minor crime. Instead of committing that crime, I was doing something socially unacceptable (say, playing violent videogames) and the prosecutor knows it.

In this situation I am left with the choice of revealing something I don't want revealed or confessing to something I didn't do. This is an unacceptable situation for the court to be able to put an individual in. The 5th amendment right against self-incrimination is what protects against this. It allows a person to simply refuse to answer, and tells the court that it can't assume that silence represents a confession.

Now, does the right to remain silent disproportionally benefit the guilty? OF COURSE IT DOES. That is "FAIL3" and nothing will get around it. If you assume that the police and the judges are good people (essentially requirement 5) then you can assume that they will only arrest people reasonably expected to be guilty and release them if they are innocent. Under those rules any rights given to the accused would then disproportionally benefit the guilty.

The MeMO Pad Smart 10 and the transformer T300 are both 720p screens rather than the 1080p on the nook.

Then again, the Tegra3 in the MeMO is better than the OMAP in the nook, so it isn't a clear-cut decision. If the MeMO was a 1080p, it would probably be a slam dunk for ASUS. However, I don't think ASUS wants to cannibalize their high-end tablet sales, which such a tablet might do.

The nook HD+ now looks pretty interesting. I like the google apps, but could live with the nook launcher (or install a new one off google play?)

$270 for a 9-inch tablet with access to google apps is pretty compelling.

I think you have the right idea about when DRM is okay. DRM is okay when the user is renting access to something, be it music, movies, books, or other works. In this case, the DRM simply automates the process of returning the item once the rental term has completed. It can produce some system incompatibility and some Fair Use restrictions, but it represents at least a fair deal for the consumer.

Personally, I don't think there is any case where DRM really makes sense for purchased content. But, I actually am okay with watermarking purchased products, which some people consider to be DRM. This does not restrict the user's rights at all, it only allows findings of fact if a user does something they aren't allowed to do. However, I think the user should be notified if it is being applied.