1997-1998-ish...

Except the article's point is both stupid and naive. It's approximately the same as shutting off your city water supply because you're pissed at where they installed bike lanes. That'll show that pesky government!!

Might as well throw this out there too: GStreamer's first release was 16 years ago on Halloween - 0.1.0 "gscreamer".

I would hope so. Given our rather significant issues with excess atmospheric CO2 right now, the last thing we need is to dig up old polystyrene and create more. That might make sense once we've reigned in total CO2 and/or can do a full capture of what these worms output, but for now I say let sleeping foam dogs lie. (though absolutely keep working on the science/tech in the meantime)

It's the loss between the absolute physical heat energy from the combustion source and the effective electrical output of the entire plant.

http://www.eia.gov/tools/faqs/faq.cfm?id=107&t=3

This shows that for 2013, the average heat efficiency of a US coal-fired plant is 32.62%, while that of a natural gas plant is 42.92%.

The fact that nuclear, coal, and oil are all similar implies that the actual steam conversion cycle is the primary limit here (as opposed to the summary which claims that the steam cycle is *another* 50% loss, which it is not). That begs the question in my mind though: how do natural gas generators beat that, and can whatever tricks those are be applied back to coal/gas/nuclear?

2Gbps over fiber is a speed I associate with GPON, which is a fiber loop that connects dozens of endpoints on a single fiber, just like the existing copper cable system. The 2Gbps figure is then derated for overhead, and finally split between all the users on that loop.

Google on the other hand is apparently doing home-run fiber from each house to a central location, where it can aggregate the bandwidth into ludicrously fast switches and hand it off to 100Gbps etc backhauls. That means that (with guessed but plausible numbers) instead of e.g. 50 houses sharing each 2Gbps for an average of 40Mbps with Comcast, you would have 1000 houses sharing 100Gbps for 100Mbps average with Google. Yeah, the "peak theoretical" is higher, but the actual effective available bandwidth is very different.

Then there's the fact that with a home-run fiber to each house, Google can easily upgrade their aggregation equipment and backhaul links in order to boost total shared bandwidth, without having to go out in trucks and mess with fibers again. Comcast OTOH would have to go around and split all their GPON loops in half and hope they can get those new sub-loops run back to their agg points. Heck, there's nothing stopping Google from upgrading the transceivers at each end of the fiber for a given house to make use of more advanced optical techniques, because the fiber isn't shared.

Consider this: you are not required to actually mark any candidate/yesno for each given race currently. Mandating that people make the effort to at least show up to vote or mail in a ballot forces them to at least *consider* what's going on and the ability to cast for a given candidate/yesno, but there's absolutely nothing forcing them to actually cast for each race. If they don't feel sufficiently informed, they just don't mark anything. The point IMO is that they were forced to at least *think* about make a decision between A/B[/C]/dont[care/know].

Yeah, I supose that's true. But not only does it not make me feel any better about how secure it isn't actually, but it makes it absolutely impossible to actually select a password that can be remembered. See my response above http://news.slashdot.org/comments.pl?sid=6909609&cid=48991127

Well, not exactly. The GW2 client remembers the password for you, thus I don't have any reasons to remember a random collection of 4 words. Except when you are forced to reinstall it, or install it on another machine, you suddenly need the password again.

The result is that I now have an email I sent to myself, in a folder, which very clearly states "GW2 password is 'aaa bbb ccc ddd'". It's in a Gmail [apps] account at least (so as per the article it's reasonably secure), but it's really no different than writing my password on a post-it on my monitor from an *actual* security standpoint.

Thus in trying to "improve" security, they force me to have a very infrequently used password that there's absolutely no chance I will ever remember, so I have to store it in an alternate location. Either that or pretty much every time I [re]install the client I have to "forget password", at which point either they're relying on absolutely nothing more than my email account's security, or they randomly require that I send in some kind of identification and wait 24+hrs like my wife had to a month ago.

FAIL.

I signed up for a Citi credit card about a year ago, then found out after the fact that not only do they allow short basic passwords, but they MANDATE them. You cannot have any special character at *all* in your password. I called them on this and they told me that they had just made the change in order to "improve security". Even better, the change happened as I was initially setting up my account, so the first form I filled out let me put in a proper password because it hadn't been crippled yet, then the actual login page kicked me out after that saying my password was invalid. I had to call them up and fight through getting my password reset, then hope that the password I created through the form that still didn't check their new rules would actually let me log in.

There's got to be a way to report these outright failures to some kind of regulatory body, and force them to fix these things. I'm just worried that there might not *be* a regulatory body for this....

On the other extreme, I found myself having to "generate a password" for Guild Wars 2, who take http://xkcd.com/936/ as gospel and created a 4-word passphrase for me. Compound this with the fact that they kick out "any password used by you or anybody else *ever*" as a password change, which makes it absolutely clear that they store all passwords in plaintext, and I'm not really impressed with those jokers either.

I would hope that they're dropping the archaic cursive style of writing because it's just that: archaic. OTOH, ceasing to teach kids how to write in a legible block "font" would be mind-blowingly stupid. No matter what people need to be able to write, but they don't have to write "fancy".

(Not to mention I can't actually manage to *read* most people's cursive writing, no matter what era they were taught it in.)

Of course. I never said I'd write one, just *how* I would. Anyway, there are already tools out there to do this, benchmark the card at the same time, and can read some of the more SD-specific functionality out as well (e.g. check the *actual* manuf string vs the case).

If you think you're getting a card for 1/5th the price, you're probably getting 1/5th the card. I have personal experience with cards that claim to be 8GB but only have 1GB of actual flash in them. I won't touch on the malware issue, but before you actually try to make use of the cards you need to find a way to very exhaustively exercise the entire card. I haven't looked for such a program but I hear they're pretty easy to find. If I were writing one I would put a pseudo-random sequence across the entire advertised size of the card, then read it back and confirm that the same pseudo-random sequence comes back. The sequence should be longer than the card, or at the very least not repeat on something like a 1GB boundary. I suspect a common trick in these cards is to simply drop the upper address bits, so you'll read the same contents off e.g. the 2nd GB as you will from the 1st, and all the others.

http://pflog.net/dsl_overhead/

Looks like they're counting ATM frames, not your IP traffic.

First off, I assert that whether Verizon is actively throttling packets, or simply not providing sufficient peering to get to Netflix, they are committing fraud by advertising high speeds and not delivering them.

However, to *really* convince people, more rigorous experiment has to be performed: find a VPN (or set up your own with a colo) that's connected as closely to Verizon as possible, as close to their peering with Netflix as possible. That way the route between Verizon and your VPN/colo is as similar as possible to the Verizon<>Netflix route. You can then measure Netflix bandwidth to your VPN/colo, and the resulting full-path bandwidth.

I *strongly* suspect you'll see the exact same behavior, but by doing that you've proven beyond a shadow of a doubt that Verizon is absolutely to blame. It still doesn't separate the packet-throttling scenario from the insufficient-peering scenario, because even though your Verizon ingress point is ideally the same router, Netflix is *supposed* to peer to that router through dedicated lines (e.g. trunked 10G to the next room over where Netflix's router is).

Of course, since Netflix has offered to both purchase and install the 10G cards and wires on their own dime, that scenario is absolutely no different than packet-throttling. Except that in order to do packet throttling, Verizon had to spend *more* money on hardware than they would have to just add more capacity. Now *there's* a bit of research to do: $ to throttle vs $ to add capacity.....