Your minimal-government-intervention solution is for the government to force the incumbent ISPs to lease their privately-owned infrastructure to their competitors, and at government-regulated prices? I am interested to see how you justify that as a lower regulatory burden than forbidding the prioritization of packets based on origin.

Most cable companies pass those user fees on to customers explicitly, "Regulatory Recovery Fee." Obviously, all expenses of the cable company are eventually paid by subscribers, but they choose to account for those right-of-way fees in the same way as they account for the Universal Connectivity Fee and State Sales/911 Tax, as though it's not a part of their doing business.

They can deliver you 10 MB/s even while they throttle the connection between you and Netflix to 2 MB/s or less. This is, in fact, what was done during the "negotiations." Bandwidth is throttled upstream of the client link, so the client, if he tried, could run a "speedtest" in parallel with his crappy, stuttering video, that would show healthy, full-bandwidth connection to other upstream sites. He could, if he tried, see a perfectly fluid Hulu video in one window, next to a crappy, stuttering Netflix video in another. The client has no way of knowing whether that's because Netflix's servers are overloaded, Netflix's ISP is overloaded, or if Verizon is throttling Netflix: they all look the same to the end viewer. This also makes it essentially impossible to determine fraud (aside from the fact that your contract with your ISP does not - can not - guarantee you a bandwidth to any particular service.

That's fine for the client ISP and the server ISP, but their packets will traverse an unknown number of intermediates whose networks are completely out of the control of server, server's ISP, client, and client's ISP. In fact, one of the internet design principles is that the physical network is unreliable and subject to congestion. This is unavoidable: traffic grows to fill the available bandwidth, and during times of peak demand, every network can be congested

This is why the internet has different protocols. Compare Email and VOIP: for VOIP it is essential that each packet be delivered, in order, and as quickly as possible. Delays of even 100ms create audible distortion. For email, your message is still understandable if it's delivered 5 minutes late. SMTP has that resiliency built-in: if it fails to deliver a packet right now, it will keep trying, periodically, for a day. Eventually, it will find a time when the network is uncontested and it can transfer that terabyte attachment. If your VOIP packet has to wait around until SMTP has delivered that attachment, you're going to think someone has hung up on you.

I know, now that everything is a browser plug-in, that it's easy to forget that HTTP is only one of many protocols on the internet. It's completely appropriate prioritize RTP/RTCP (VoIP) over SMTP or FTP. The problem is when an internet middleman decides it should be able to prioritize YouTube's HTTP over Hulu's HTTP, just because YouTube has paid a ransom.

No, that's not ok. Email is already a best-effort service without guaranteed delivery. If the S&R team actually needs a particular piece of information delivered immediately, they should choose a service that is optimized for that purpose. It's not the job of every internet middleman between here and Beijing to rank the moral value of each IP packet or source.

Note that this is different from an ISP determining that an email source is "spam" and blacklisting that source.

Aside from the technical fact that the client only finds the ads in the web page text, it is (again) not appropriate for the internet middlemen to determine whether the client is more interested in images from doubleclick or images from slashdot. If the client chooses to prioritize which images it requests, that's a completely different question. The point of net neutrality is that, within a recognized communication stream, people who transfer the data should not look at the data to determine whether or how quickly to forward it. The post office accepts your letter, looks at the postage you've paid, and delivers it. It will deliver my Priority Mail envelopes faster than my Media Mail envelopes, but it will not deliver Netflix Media Mail envelopes faster than my Media Mail envelopes.

Legitimate packet prioritization is based on the service, not on the vendor. "Quality of Service" is already part of TCP/IP, and lets routers know how to balance latency and throughput. For example, it is more important to deliver VOIP or streaming video packets on time than to deliver SYN/ACK packets quickly. You should let VOIP packets skip ahead of SYN packets or FTP packets. However, it is not appropriate to let Verizon VOIP packets skip ahead of Nextiva VOIP packets, just because Verizon has paid for that prioritization.

Common carrier rules mean that the carrier can't discriminate among its clients, they can still distinguish between "First class" and "Book rate" services.

Here's the problem: the person who takes your order is not the person who delivers your order. There needs to be some way for server A to identify you to server B. Possible solutions:

Assign a number to customer, and expect customer to answer to that number. Problems: depersonalizing, customers forget their numbers, "thirteen" sounds like "thirty"

Let the customer assign an identifier for his order, providing some illusion of personal service. Problems: customer identifier may be confusing, customers may get annoyed if server A does not use the mystical spelling customer has in mind, servers may spend more time massaging the identifier than actually preparing product, server B may not pronounce the same identifier as server A recorded

Photograph customers for product delivery. Problems: privacy fanatics, bad pictures, servers turn incidental photodocumentation into DMV-like picture-taking ritual

In sum, there is no good way to make high-volume service look like personal service. People pretty quickly see through efforts to disguise it. While many people are willing to play along, occasional servers and customers will both manipulate these systems for their personal amusement. Misspelling your name is the barista equivalent of building paper-clip animals. Giving a fake name is the equivalent of painting your stapler purple. Try not to get bent out of shape when they call you "Susquehannah" instead of "Susquanna" or "Todd" instead of "Tom": they aren't trying to annoy you; they aren't trying to learn your name; They aren't likely to remember you next time; the content of your name is irrelevant to the process; and and effort to "get it right" only delays people around you.

If those claims have merit, then they should be sent to the NIH's Office of Research Integrity (since the research was NIH funded). NIH will do a proper, 3rd party investigation. Look at the original data and find out whether the supposedly copied bands are copies or just similar in appearance. Find other potential instances. Try to find out whether it's the PI, staff, or trainee(s) behind any manipulation. The ORI has the power to impose actual sanctions, as opposed to just innuendo. Making anonymous accusations on pubpeer is not the way to improve science.

Because that kind of arglebargle obfuscates their message. Just because the slithy toves do gyre and gimble in the wabe, does not necessarily make the borogroves all mimsy.

There's a place for poetry, and a place for clearly stated information.

I disagree. The macroscopic phenomenon of sound comes from vast numbers of atoms acting in aggregate, and their effect dissipates rapidly as the initial energy is spread across more and more atoms. That can't happen at the quantum level. These folks suggest that, at a small enough level, the interaction becomes quantized, such that "sound" energy might transfer from one atom to exactly one other atom. ie, that the "billiard ball" model of atoms bouncing off each other can be reduced to a quantal exchange of energy very much like fluorescence resonance transfer.

Clearly, not a good way to listen to the latest Katy Perry song (if there is a good way to listen to the latest Katy Perry song), and pretty clearly not the ordinary definition of sound as a subjective phenomenon. If you're a physicist, trying to explain your study of quantal energy exchange among atoms to the lay press, "sound" is probably a pretty good metaphor.

In terms of headcount, the US has the 3rd largest military, behind China and India. (North Korea is 4th) The US military employs 70% more people than the Russian military.

In terms of spending, the US has no close competition. The US spends 3.5 times as much as the next largest spender (China), and accounts, by itself, for more than a third of global military spending.

China, by dint of having 20% of the world's population and 18% of the gross world product, is an enormous investor in everything. In contrast, Iceland, despite being a "developed" country highly dependent on geothermal energy, is one of the smallest global producers of renewable energy (53% = 80 PJ, vs 20%=440 PJ in Germany or US 12%=2,300 PJ).

You have to keep in mind whether the important number is absolute investment, per capita investment, or fractional investment.

The navy does not expect its submarines to operate at a profit. This is partly because they know that the market for nuclear missile-generated craters is fickle, so their sales are going to vary dramatically from year to year, include whole decades at a stretch where they may not deliver even a single warhead. It is partly because their other principle cargo, national influence, is very hard to value objectively. Most companies carry this product as "goodwill," and serious accountants completely disregard it in valuations.

The whole business model of nuclear submarines is a sham. A ponzy scheme foisted off on a credulous public awed by technology and investor story time, run by directors spending other people's money, but guaranteed to collect their own luxurious salaries regardless of whether the business ever turns a profit. 50 years without delivering a single megaton warhead...you'd think investors would wake up.

That article talks almost entirely about sulfur, which is only one aspect of pollution, and arguably less important for vessels/vehicles that spend their time hundreds or thousands of miles from populated areas.

They compare a car with annual use of 15,000km, carrying approximately 100 kg of cargo (1000 ton-miles), with a ship that travels 200,000 km carrying 150,000,000 kg of cargo (20,000,000,000 ton-miles). ie, based on equivalent use, one massive container ship is equal to 20 million cars. If that container ship produces sulfur equivalent to 50 million cars, despite using fuel with 2000 times more sulfur than terrestrial diesel, then I'd say they're doing a damn fine job of pollution control.

That's a bit of hyperbole, isn't it? Yes, google is rolling out to relatively high-density neighborhoods, but none of these are "flooded" with ISP options. At best, they have one cable option, one FIOS option, and one DSL option. To the best of my knowledge, there is nowhere in this country that you can choose between two cable providers.

You may also be forgetting that, when the incumbent ISPs were themselves startups, they didn't offer much in the way of rural service, either. In fact, most of them had to be paid by the government (and are still being paid by the government) to extend service outside of the most profitable neighborhoods.

Of course, that was 20-50 years ago, so you can be forgiven for imagining that Comcast launched in 1969 to 100,000,000 homes scattered across 80,000 square miles. I, for one, am happy to see anything vaguely resembling a new entrant in communication services. If they can only roll out to one city block, or even just one apartment complex, and provide better, faster or cheaper service than the legacy behemoths, I'll be happy to see them succeed.