Everything you've stated is true, but typically the ISP does not show the cap in the advertising and just compares everything in regards to upload and download bandwidth.
Not in the UK - most of the ISPs have the cap in the headline advertising.
It depends on your ISP. As I said, iff they are advertising the cap, instead of saying "unlimited" then I see no problem. And indeed, this is what _most_ of the ISPs do here in the UK (certainly the smaller ISPs pretty much never say "unlimited" and advertise a cap.) Some of the ISPs that fall into the "big, cheap and crap" category do still sell dirt cheap "unlimited" accounts with hidden caps, but this is increasingly uncommon.
And a cap that can be consumed in 1-2 days of moderate usage is a fraudulent advertisement when you claim multiple megabits of capacity.
No, no it isn't. If the advertisement says "unlimited" then I would agree that capping it is fraudulent (unfortunately the ASA disagrees on this point); but my ISP says "up to 24Mbps capped at 15GB/month during peak hours" (and defines those "peak hours"), and there is nothing fraudulent about that: I'm absolutely happy with that advertising because I bought the account knowing full well what the cap was and how much it would cost me to increase the cap if necessary. This is good for me - I'm not a heavy downloader, I don't use more than 15GB/month during those peak hours, I don't want to be subsidising the people who thing that they have some kind of a god given right to max out their connection 24 hours a day.
Not that my ISP has caps. Nor that I'd ever sign on with an ISP that has them.
But that's because I'm not interested in getting conned.
I don't see how you can claim an ISP capping people at the advertised limit a "con" - if you can't read the terms of the contract you're signing then that's your own stupid fault, not an attempt to con you. Now, if the ISP is genuinely misadvertising and not explaining the caps in their literature, then that would be a con, but since you said you'd never sign on with an ISP that had caps, you presumably mean you'd know about them _before_ signing, so no, it wouldn't be a con.
Now if the ISPs were to advertise something like "260GB per month at up to 5mbit speed", then they'd be producing fair ads. Ditto if they sold different cap tiering levels and bundles without requiring you to pony up for a higher speed to get a higher cap.
It took me a while to figure it out too. There are two types of Twitter users. Some use it to randomly spam their friends, and get replies almost immediately. It's like a group conversation among whoever is listening at the moment. Kind of like newsgroups except it's only for your friends (because no one else cares) for people who aren't bothered by the short limit.
So, uh, kinda like a really limited version of facebook?
Look, I pay for a class of service (5Mbit down, 640Kbit up.) Deliver that level of service. Period.
Your "throttling" attempts and "bandwidth caps" are nothing more than trying to steal back what I've already paid for.
No. If you buy an Internet connection that is sold as "5Mbps down, 640Kbps up maximum with 60GB/month cap" then that's what you pay for - the ISP is well within their rights to throttle down your connection (or even disconnect you entirely) after you've hit the 60GB cap, since that's what you agreed to.
The issue is ISPs selling "5Mbps down/ 640Kbps up, unlimited" services and then imposing a "fair usage" cap - the contract says "unlimited", so it should damned well be unlimited. Unfortunately, in the UK at least, the advertising standards association ruled, many years ago, that the term "unlimited" didn't actually have to mean unlimited. This is a problem, partly because the public expect "unlimited" to mean exactly that, and partly because the "fair usage" caps vary wildly between ISPs - for example, Orange's GPRS "fair usage cap" used to be 10MB/month, at the same time as 3G MNOs were offering caps in the 1-5GB/month region and most ADSL connections came with caps somewhere in the 30-200GB/month region; Basically, if ISPs are allowed to use the term "unlimited" for connections that aren't unlimited then the customer has no way to know what they are actually buying.
My personal opinion is that "unlimited" connections are pretty much unfeasable at the usual price-points - home internet connections are made cheap and fast by having a burst rate far in excess of what the network could sustain as a continuous rate from all customers at the same time. For example, FTTC tops out at a burst rate of 80Mbps - building out a network that can cope with all the customers hammering it at 80Mbps at the same time would massively push the prices up. I (who uses about 20GB/month) am not interested in the price of my internet connection increasing to subsidise the cost of handling the customers who leave bittorrent going 24/7 - if they want masses of bandwidth, they can pay for it themselves. So IMHO there *should* be some kind of data cap that is set based on how much you're willing to pay.
How do I compete with the mass market pricing of the Telco or cable service?
If your only competetor is blocking access to most parts of the internet that people want to use, then it seems to me that you don't need to compete on price at all - you are providing a service people want, your competition aren't providing a service people want, therefore as long as your product is _affordable_ to your customers, it can be more expensive than the competition.
As I see it, I can't see how Verizon can make this work - is their customer base really big enough for the vast majority of websites not to say "well fuck 'em then" and live with losing that part of their audiance, rather than handing over cash? Hell, even if all the US websites paid (which seems unlikely) and the rest of the world didn't, that'd still leave Verizon's customers with a next to useless internet connection.
I certainly agree that I'd like criminal charges applied to decision makers at companies like this. BUT...
50,000 gallons isn't that much water. It was a 10,000 gallon-per-day spill. That's garden-hose territory. The fact that they were made to clean it up and pay fines seems reasonable to me - I'm not sure jail time is warranted here.
I can't see how they can claim it is neither negligent nor deliberate - as far as I can see the only reason you wouldn't consider this to be negligent is if it was done intentionally.
So whilst it may not be a serious spill, I would say that they need to be punished either for doing it deliberately (for which there is no excuse), or for trying to weasel out of it instead of just admitting that an accident had happened and cleaning it up.
If it turned out to not be deliberate, and they didn't try to weasel out of their responsibilities then I might agree with you, but companies need to be punished if they try to avoid being held accountable for their own actions, even those actions that are accidents.
That's the most obnoxious open source cop out there is.
It really isn't a cop out. Lets compare open source and closed source shall we:
With closed source code you are generally paying someone. This covers the development costs of the software. In theory, you could ask them to make a change, and they could do so because you are paying them for this.
With open source software you are generally getting it for free. You're not paying for anyone to do the development for you. If you really want to have some change implemented, you are free to pay someone to do it for you.
In this regard, there really is no difference between open and closed source software - in both cases, if you want a change to be made then you really need to fund a developer to do the job.
Of course, in reality, things are often a little bit different. Certainly, I've had a lot more luck working with open source projects to get bugs fixed and features implemented *for free* than I have convincing the likes of Apple and Microsoft to fix bugs in their closed source software which they are being paid for.
So why is the "fix it yourself" comment often heard in the Free Software communities, so much more of a cop out as the "we're not interested in fixing that" you get from proprietary software companies? Why would you expect someone to fix this stuff for free, when conversely the commercial world is often not even interested in fixing this stuff when you pay them?
Free software has a major advantage over commercial stuff: you get the opportunity to get things fixed by hiring someone to do it, if they are important enough to you - you can't do that with proprietary software.
I have been corresponding with my MP about this and in her last letter she indicated that circumventing Cameron's porn filter would become illegal. I asked her for urgent clarification of this point as it would appear to outlaw many vital technologies, including VPNs.
I just hope it was a mistake on her part, otherwise privacy will be criminalized.
ISTR that social services had already made a comment on the radio that they would use whether you'd turned the filtering off to determine your fitness as a parent. Unfortunately I can't find a recording of the interview.:(
What is needed is regulation of the insurance companys to reduce the scope of their discrimination.
Discrimination, at least insofar as it relates to risk, is pretty much the entire point of an insurance company.
No, it isn't.
For the customer, excessive discrimination is a bad thing: if the insurance company can determine your risk with 100% accuracy then you are _guaranteed_ to be paying over what it would've cost to self-insure. So the higher the insurer's accuracy WRT discrimination, the worse it is for the consumer.
On the other hand, assessing risk accurately is very good for the insurance company, because out-pricing the customers who are most likely to need insurance and retaining those who don't need it is good for their profit margins.
CA would help the NSA in this way unless they were put under immense pressure
Like do as we say or spend the next twenty years behind bars?
Yes, but still, if this was happening on a mass scale we would've heard about it by now because its trivial to detect. So the conclusion is that either its not happening, or is only happening in specific targetted cases.
That doesn't solve the insurance problem. The insurance company would demand disclosure of the DNA and use it to deny coverage or charge more.
That goes for any diagnostic tool that would spot illnesses in their earliest stages though doesn't it? What is needed is regulation of the insurance companys to reduce the scope of their discrimination.
If you don't have it, it's much harder for the insurance company to demand it
Or they will simply demand you get sequenced before giving you insurance...
Problem there is even if the current government doesn't allow 3rd party access to the DNA database, would the next congress/administration share the same vision?
Indeed. I would be happy for DNA to be sequenced at birth, analysed for any conditions that need immediate treatment, and then the only copy given to the parents for safe keeping (not that I think the general public are especially good at doing the "safe keeping" thing, unfortunately). Keeping all the data in a database is problematic for exactly the reason you state.
How long until someone decides its ok for 3rd party researchers to access the data for 'medical study'?
For *actual* medical studies, this is pretty good - if you can provide the anonymised DNA sequence along side anonymised medical records then that gives the bioinformatics people a hell of a lot of data they can statistically analyse - that's the kind of thing that would probably greatly advance knowledge about what conditions are caused by what genetic sequences. The problem with this, of course, comes back to the fact that in order to generate this anonymised data, you probably need a non-anonymised database of it all somewhere, which could be abused.
If no compelling medical issue requires sequencing in a newborn, it is invasive and coercive to conduct it.
What if the baby might have some condition that is better treated as early as possible? Does that not count as a "compelling medical issue"?
Are blood samples not already taken from newborns anyway? If so, genetic testing doesn't seem any more "invasive".
However, I agree that indeed my concern is that the data will be kept on file and used for purposes other than medical treatment - I like the idea of the medical profession having lots of genetic data on file; I don't like the idea of the government (and by extension, police, security services, insurance companies, etc.) having that data.
I'd wager that the fundamental flaw in HTTPS is that the government has the private keys direct from the CAs. The protocol is flawed in the key management (as most are).
It's not that simple. When you get your certificate signed, you never give the CA your private key - you give them a certificate signing request instead. So by compromising a CA, the NSA can't get access to the key required to do offline decryption of an SSL session.
They *can* get the CA's own key, of course. And that would allow them to forge a new certificate that claims to be yours. They can use that certificate to perform an online man in the middle attack of your SSL sessions, and anyone validating the certificate by checking it is signed by the CA wouldn't be able to tell the difference. However, the real cert and the forged cert would be different, which means there's a reasonable chance that someone's going to spot this if its done on a large scale. For example, as far as I know, Chrome checks that the certificates Google's websites are presenting it with are actually Google's certificates, even if they appear to be correctly signed by the CA - if it sees a fake cert it reports back to Google. (This is how Google has caught compromised CAs before)
As we've seen before, if a CA is caught handing out forged certificates, their life is pretty much over - all the browser vendors immediately revoke their CA certificates. So given how easy it is to spot the forged certificate when this kind of attack is pulled on a large scale, and how bad things would go for the CA when someone spotted it, it seems unlikly that a CA would help the NSA in this way unless they were put under immense pressure.
Presumably you are talking about your experience in the US, because that kind of shit would never be allowed in the EU. You have a right to food, accommodation and compensation. I don't know why US consumers put up with being treated that way.
Whether its allowed or not is irrelevant. 2 years ago I was flying with Thompson Airlines on a Crystal holiday (Thompson and Crystal are the same company, but they operate under different brands). For the return trip, in order to save themselves some money they had decided to use a single transfer for a number of widely spaced flights - that meant we were taken to Chambery airport 4 hours before we were due to depart - that's long before the checkin desks are even open, so we couldn't even check in our luggage. Then, severe weather lead to the closure of the airport for some time.
EC directive 261/2004 requires that air carriers provide free food and meals after only a 2 hour delay, so about 2 hours after our expected departure time, they announced that the plane was landing, that there would be a rapid turn-around and waved us through to the departure area. Of course once we went through security, it rapidly became clear that the Crystal reps had lied to us and the plane was, in fact, not there. However, there were no Crystal or Thompson reps in the departure area and when airport staff phoned them they were informed that the reps were "too busy" to come and speak to us. Eventually, 3 hours later, we actually got to see a rep, because one of the other passengers got pissed off and ran back through security.
We asked for the food and drink that we should have received, and the rep told us “Thomson is aware that we have a legal obligation to provide food and drink but we are choosing not to do so”. Eventually they did agree to provide drinks but refused to provide any food. We eventually took off 11 hours after we were supposed to, having sat on the cold floor of the airport for 10 hours (because Chambery is small and there weren't enough seats). The air crew had been informed that we had all received food and drink on the ground and therefore didn't require any during the flight, so most of the passengers had to go without food until we landed at Bristol.
Upon returning to the UK, Crystal disclaimed all responsibility, stating that we had to complain directly to Thompson (even though they are the same damned company). Both Thompson and Crystal refused to apologise or compensate us, essentially claiming they had done nothing wrong. This was despite them knowing that we were a large group who booked a group holiday every year. ABTA, IATA and the CAA all refused to do anything about their flagrant disregard for the law.
So yeah, it may "not be allowed", but it still happens - the airlines ignore the law and the regulators don't punish them for it. Needless to say, Crystal/Thompson will be losing out on over ten thousand pounds of business every year, since our group will never book with them again - all because they decided to save a couple of hundred pounds by refusing to meet their legal obligations and offering utterly shite customer service both at the time of the incident and in the months afterwards.
All I was saying is that, while you CAN have a different port when you're connecting to a different address/port, there's no need to (when the internal address/port is the same). Since there's no need to, and it's easier to use a single mapping table (mapping internal IP/port to an external port, regardless of what it's connecting to), I'd think that would be the solution most (cheap) routers would use (the kind that also would only have one external IP address).
A lot of the cheap routers run Linux, which certainly doesn't do as you describe:)
Do you still want no say in operation when safety measure are cut to protect your 'investment'?
I would like people with some understanding of the job at hand to have a say. However, the vast vast majority has _no_ understanding of this stuff, react to what the papers/environmentalists/whoever say, and their voices drown out those who know what they're talking about. So my only conclusion is that no, maybe the public shouldn't have a say - it should be down to the experts. The trick there, unfortunately, is how to ensure that decisions are made by neutral experts rather than people with a vested financial interest in one thing or another (most of the MPs have their fingers in lots of the pies that they have to vote on... and they don't declare this of abstain from votes when they have a conflict of interest).
Why is Chinese government* even allowed to operate a nuclear plant in the UK?
Because consecutive governments have chosen not to make a decision on the future of British energy because they know that *all* the choices are unpopular. By the time crunch time had been and gone and a decision was long overdue, the inaction of the government had basically ensured that the UK had no nuclear industry who could feasably commission nuclear power stations. So now we're stuck having to bring in the nuclear industry from other countries who *didn't* put the whole nuclear power idea on hold for decades and therefore still have some expertise in this area.
As it stands, things ahve been left so long that there is no way the UK is going to be able to commission new power stations in time to avoid a shortfall in generation capacity and there is next to no working UK expertise WRT building nuclear power stations anymore, so what else can we do but bring in foreign corporations to help out?
I'd think the obvious solution to how to map port numbers for NAT would be to map each host/port pair to a single random external port, regardless of the address it's sending/receiving to/from. That way you have just one lookup table for the mapping each way. IF the NAT router does that, then all you do is connect to a common server, which notes your external IP address/port and passes that on to the other party (and vice versa).
This is a misunderstanding of how a lot of NATs work.
A connection is identified by a tuple of (protocol, source IP, source port, destination IP, destination port). The source IP and source port are going to be rewritten by the NAT. The NAT will maintain a table of connections, which will map:
The "translated source IP" is always going to get rewritten to the router's external address (* this is not entirely true - see below), so that bit is relatively easy - you can poke an external server and that can pass on the "translated source IP" to the peer.
The "translated source port" is arbitrary - the NAT may decide to leave it as it is, or it can pick any port it wants. And there doesn't have to be any correllation at all between any two connections. I.e. if you connect to host A, destination port B with original source port C, and then connect to host Y, destination port Z with original source port C, there is no guarantee the translated source port will be the same in both cases. So you can't just poke an external server and have it pass on the translated source port to your peer, because the translated source port you're going to use when talking to the peer could be different.
In practice, many NATs try to avoid changing the source port at all unless there is some other connection identified with the same (protocol, translated source IP, translated source port, destination IP, destination port) - then it has to change the source port. This means that you end up with things mostly working, and then randomly breaking for no apparent reason (i.e. when another machine on the network is trying to connect to the same host and has happened to pick the same ports). You can mitigate this to some extent by trying to randomise ports and addresses as much as possible, but its never going to completely go away, and furthermore there is no guarantee that the NAT will even try to keep the port the same.
(* routers can have multiple public IP addresses, so there's also no guarantee that they will always give you the same translated source IP either. Although this is rarer).
I guess this wouldn't be such a problem if there was a standardised way of doing NAT... but there isn't - there are a variety of systems employed in the real-world, and to have a robust solution you have to be able to cope with all of them. Anyway, most of this is covered by the STUN protocol, which is a method of trying to determine all of the information you need to tunnel through a NAT; and the STUN RFC specifically says it is a best effort attempt and can't be made reliable.
You'd still need to get through the firewall if the router is doing stateful blocking (which is easy to get around - there ought to be a connect_to low level call to do that, but easy enough to do a listen/connect on both sides).
Actually, stateful firewalls are pretty much a doddle to tunnel through - you essentially do exactly the same as you would for NAT, but it completely eliminates all the guesswork since you know that the addresses aren't going to change arbitrarilly. This is why peer-to-peer stuff over IPv6 works well, even when you have stateful firewalls - you just need an unfirewalled server somewhere to mediate the transaction between the peers, and you don't need to be concerned with anything in the middle modifying the traffic in ill-defin
They have these gantries on the M42 and M6 near where I live. You can see the first one from over a mile away, and they're spaced at approx 1/4 mile intervals. You can often see two ahead at least.
Fine on straight roads - the M4 through south wales has things called "bends" which obscure some of the gantries until you're a bit too close to lose 20mph of speed without hitting the brakes. As I said, all they need to do is drop the speed limit in 10mph increments instead of 20mph increments and it'd be fine.
I hate those cameras so much. I wouldn't mind them so much if the road was permanently 50, but 50-70-50-70-50-70 with cameras in between is just pure ass-hole.
Variable speed limits *do* work to improve traffic flow, and you probably do need cameras to enforce them, so I don't have a big problem with them. But changes to the limit need to be done in small enough increments that people don't have to slam on the brakes to avoid going too fast. My only minor issue with average speed cameras in a 50 limit is that the trucks seem to ignore them - trucks are generally not great at overtaking; they creep past at 52mph, and then pull in before they've actually finished overtaking. Twice I've been run off the road by trucks trying to do 52mph in a camera'd 50 zone whilst I've been religiously doing dead on 50 to avoid being ticketted.
The other issue is: what do you do when the limits are blatently stupid. I've been cruising down the M25 in the early hours of the morning before, not a thing in sight and suddenly the variable speed limit signs come on telling me to slow down to 20mph - that seems downright dangerous to me, since the road is completely clear - if someone decided to ignore the limit and come up behind me at 70mph, that'd be a serious accident waiting to happen. In that case I decided I was going to ignore the limit and do 40mph, fully expecting to have to argue my case for safety in court, luckilly I didn't get ticketted.
While everybody is writing about the water, the real issue is the spent-fuel-rod pool. If that thing is not secured very soon, Tokyo becoming uninhabitable within a very short time is a real possibility. There is so much radiation in there, it is staggering. The pool is inadequately cooled. The pool is damaged enough that even a minor earthquake could prevent cooling it more and a fire starting in there would both be impossible to put out and starting by itself very fast.
Is there really much chance of a fire? The material in the spent fuel pools has been cooling for at least 2 years already...
Personal prediction: TEPCO will go on blundering about, and eventually they will get a nuclear catastrophe that makes all others so far look like a summer breeze. After that, Japan will not play a role in the world for a few thousand years or longer, because for all intents and purposes it will not really be there anymore.
I don't think it's going to be that serious, *but* I do wonder WTF anyone is still allowing TEPCO to run the show, since they have repeatedly been shown to be covering up at every turn. At the very least the upper management should all be removed, even if they need to keep the engineers.
I can only assume that the iOS app is similar, but the Android app uploads not just your phone number (which is scraped without your explicit permission), but also your call history every time you log in.
Let me repeat that: Facebook uploads your entire call history every time you open their Android app.
I switched to the Atrium app a few months ago, having got fed up with the intrusive ads on the official app and knowing that I could never upgrade the official app because they had massively expanded the permissions it required (there was no way I was going to give it permission to do things like see what apps were running, etc). I've been pretty please with it so far - a couple of slightly niggley bugs, but on the whole its good.
Sounds like a wise choice. I know Facebook isn't the only app employees might waste your time on, with their phones. Every time I see someone's employee fiddling with their phone -- usually while not providing customer service, but always while stealing paid time -- I wonder how they manage to remain employed.
Flexibility runs both ways - if you're going to be a dick and prevent employees from taking the occasional 5 minute break (because it's "stealing" from you) then they're not going to be inclined to do anything over and above their contract either. Don't expect someone to stay late to clear up some problem (because that would be you "stealing" from them) if you're never going to return the favour.
FWIW, an old employer of mine started doing the kind of shit you're suggesting - I got a massive bollocking for ending up 5 minutes late due to traffic one morning... the previous night I had stayed 2 hours late to finish some work. Needless to say, I never stayed late again, and left the company relatively soon afterwards... in fact, most of my colleagues also got pissed off with them and quit - they lost 75% of their technical staff in a 2 month period.
I thought it was just a case of getting both parties to attempt to send packets to each other on an agreed UDP port - that's how I thought Skype did it anyway.
Depends on the NAT. There is absolutely no guarantee that the NAT won't alter your source port (which means the traffic from the other end that is directed at that port won't get to you). This stuff isn't too bad if one end is behind a NAT, but it becomes unreliable once you have both ends behind a NAT because there's just no way to guarantee the NAT will behave how you need it to behave.
What kind of grade A moron would make a car brake to match a speed limit? It'll just turn the engine off to slowly reduce speed.
The same grade-A idiots that make you brake by reducing the speed limit too sharply to meet by engine braking alone and then whacks a speed camera in to catch anyone who didn't slam on the brakes.
An example I drive with reasonable frequency - the variable speed limit on the west-bound M4 near Newport, Wales. When they decide to reduce the limit to 50, the first 50 sign you see is too close to slow from 70 before passing it without braking. And every other gantry has a set of speed cameras in it, so you've basically got to hit the brakes on a motorway to avoid getting a NIP. This could be easilly solved by making the first sign a 60, and the next one a 50 to give you plenty of slowing down time.
Slashdot Mirror
Everything you've stated is true, but typically the ISP does not show the cap in the advertising and just compares everything in regards to upload and download bandwidth.
Not in the UK - most of the ISPs have the cap in the headline advertising.
But they don't advertise the cap, do they?
It depends on your ISP. As I said, iff they are advertising the cap, instead of saying "unlimited" then I see no problem. And indeed, this is what _most_ of the ISPs do here in the UK (certainly the smaller ISPs pretty much never say "unlimited" and advertise a cap.) Some of the ISPs that fall into the "big, cheap and crap" category do still sell dirt cheap "unlimited" accounts with hidden caps, but this is increasingly uncommon.
And a cap that can be consumed in 1-2 days of moderate usage is a fraudulent advertisement when you claim multiple megabits of capacity.
No, no it isn't. If the advertisement says "unlimited" then I would agree that capping it is fraudulent (unfortunately the ASA disagrees on this point); but my ISP says "up to 24Mbps capped at 15GB/month during peak hours" (and defines those "peak hours"), and there is nothing fraudulent about that: I'm absolutely happy with that advertising because I bought the account knowing full well what the cap was and how much it would cost me to increase the cap if necessary. This is good for me - I'm not a heavy downloader, I don't use more than 15GB/month during those peak hours, I don't want to be subsidising the people who thing that they have some kind of a god given right to max out their connection 24 hours a day.
Not that my ISP has caps. Nor that I'd ever sign on with an ISP that has them.
But that's because I'm not interested in getting conned.
I don't see how you can claim an ISP capping people at the advertised limit a "con" - if you can't read the terms of the contract you're signing then that's your own stupid fault, not an attempt to con you. Now, if the ISP is genuinely misadvertising and not explaining the caps in their literature, then that would be a con, but since you said you'd never sign on with an ISP that had caps, you presumably mean you'd know about them _before_ signing, so no, it wouldn't be a con.
Now if the ISPs were to advertise something like "260GB per month at up to 5mbit speed", then they'd be producing fair ads. Ditto if they sold different cap tiering levels and bundles without requiring you to pony up for a higher speed to get a higher cap.
Oh look, that's exactly what they do.
It took me a while to figure it out too. There are two types of Twitter users. Some use it to randomly spam their friends, and get replies almost immediately. It's like a group conversation among whoever is listening at the moment. Kind of like newsgroups except it's only for your friends (because no one else cares) for people who aren't bothered by the short limit.
So, uh, kinda like a really limited version of facebook?
Look, I pay for a class of service (5Mbit down, 640Kbit up.) Deliver that level of service. Period.
Your "throttling" attempts and "bandwidth caps" are nothing more than trying to steal back what I've already paid for.
No. If you buy an Internet connection that is sold as "5Mbps down, 640Kbps up maximum with 60GB/month cap" then that's what you pay for - the ISP is well within their rights to throttle down your connection (or even disconnect you entirely) after you've hit the 60GB cap, since that's what you agreed to.
The issue is ISPs selling "5Mbps down/ 640Kbps up, unlimited" services and then imposing a "fair usage" cap - the contract says "unlimited", so it should damned well be unlimited. Unfortunately, in the UK at least, the advertising standards association ruled, many years ago, that the term "unlimited" didn't actually have to mean unlimited. This is a problem, partly because the public expect "unlimited" to mean exactly that, and partly because the "fair usage" caps vary wildly between ISPs - for example, Orange's GPRS "fair usage cap" used to be 10MB/month, at the same time as 3G MNOs were offering caps in the 1-5GB/month region and most ADSL connections came with caps somewhere in the 30-200GB/month region; Basically, if ISPs are allowed to use the term "unlimited" for connections that aren't unlimited then the customer has no way to know what they are actually buying.
My personal opinion is that "unlimited" connections are pretty much unfeasable at the usual price-points - home internet connections are made cheap and fast by having a burst rate far in excess of what the network could sustain as a continuous rate from all customers at the same time. For example, FTTC tops out at a burst rate of 80Mbps - building out a network that can cope with all the customers hammering it at 80Mbps at the same time would massively push the prices up. I (who uses about 20GB/month) am not interested in the price of my internet connection increasing to subsidise the cost of handling the customers who leave bittorrent going 24/7 - if they want masses of bandwidth, they can pay for it themselves. So IMHO there *should* be some kind of data cap that is set based on how much you're willing to pay.
How do I compete with the mass market pricing of the Telco or cable service?
If your only competetor is blocking access to most parts of the internet that people want to use, then it seems to me that you don't need to compete on price at all - you are providing a service people want, your competition aren't providing a service people want, therefore as long as your product is _affordable_ to your customers, it can be more expensive than the competition.
As I see it, I can't see how Verizon can make this work - is their customer base really big enough for the vast majority of websites not to say "well fuck 'em then" and live with losing that part of their audiance, rather than handing over cash? Hell, even if all the US websites paid (which seems unlikely) and the rest of the world didn't, that'd still leave Verizon's customers with a next to useless internet connection.
I certainly agree that I'd like criminal charges applied to decision makers at companies like this. BUT...
50,000 gallons isn't that much water. It was a 10,000 gallon-per-day spill. That's garden-hose territory. The fact that they were made to clean it up and pay fines seems reasonable to me - I'm not sure jail time is warranted here.
I can't see how they can claim it is neither negligent nor deliberate - as far as I can see the only reason you wouldn't consider this to be negligent is if it was done intentionally.
So whilst it may not be a serious spill, I would say that they need to be punished either for doing it deliberately (for which there is no excuse), or for trying to weasel out of it instead of just admitting that an accident had happened and cleaning it up.
If it turned out to not be deliberate, and they didn't try to weasel out of their responsibilities then I might agree with you, but companies need to be punished if they try to avoid being held accountable for their own actions, even those actions that are accidents.
That's the most obnoxious open source cop out there is.
It really isn't a cop out. Lets compare open source and closed source shall we:
With closed source code you are generally paying someone. This covers the development costs of the software. In theory, you could ask them to make a change, and they could do so because you are paying them for this.
With open source software you are generally getting it for free. You're not paying for anyone to do the development for you. If you really want to have some change implemented, you are free to pay someone to do it for you.
In this regard, there really is no difference between open and closed source software - in both cases, if you want a change to be made then you really need to fund a developer to do the job.
Of course, in reality, things are often a little bit different. Certainly, I've had a lot more luck working with open source projects to get bugs fixed and features implemented *for free* than I have convincing the likes of Apple and Microsoft to fix bugs in their closed source software which they are being paid for.
So why is the "fix it yourself" comment often heard in the Free Software communities, so much more of a cop out as the "we're not interested in fixing that" you get from proprietary software companies? Why would you expect someone to fix this stuff for free, when conversely the commercial world is often not even interested in fixing this stuff when you pay them?
Free software has a major advantage over commercial stuff: you get the opportunity to get things fixed by hiring someone to do it, if they are important enough to you - you can't do that with proprietary software.
I have been corresponding with my MP about this and in her last letter she indicated that circumventing Cameron's porn filter would become illegal. I asked her for urgent clarification of this point as it would appear to outlaw many vital technologies, including VPNs.
I just hope it was a mistake on her part, otherwise privacy will be criminalized.
ISTR that social services had already made a comment on the radio that they would use whether you'd turned the filtering off to determine your fitness as a parent. Unfortunately I can't find a recording of the interview. :(
What is needed is regulation of the insurance companys to reduce the scope of their discrimination.
Discrimination, at least insofar as it relates to risk, is pretty much the entire point of an insurance company.
No, it isn't.
For the customer, excessive discrimination is a bad thing: if the insurance company can determine your risk with 100% accuracy then you are _guaranteed_ to be paying over what it would've cost to self-insure. So the higher the insurer's accuracy WRT discrimination, the worse it is for the consumer.
On the other hand, assessing risk accurately is very good for the insurance company, because out-pricing the customers who are most likely to need insurance and retaining those who don't need it is good for their profit margins.
Like do as we say or spend the next twenty years behind bars?
Yes, but still, if this was happening on a mass scale we would've heard about it by now because its trivial to detect. So the conclusion is that either its not happening, or is only happening in specific targetted cases.
That doesn't solve the insurance problem. The insurance company would demand disclosure of the DNA and use it to deny coverage or charge more.
That goes for any diagnostic tool that would spot illnesses in their earliest stages though doesn't it? What is needed is regulation of the insurance companys to reduce the scope of their discrimination.
If you don't have it, it's much harder for the insurance company to demand it
Or they will simply demand you get sequenced before giving you insurance...
Problem there is even if the current government doesn't allow 3rd party access to the DNA database, would the next congress/administration share the same vision?
Indeed. I would be happy for DNA to be sequenced at birth, analysed for any conditions that need immediate treatment, and then the only copy given to the parents for safe keeping (not that I think the general public are especially good at doing the "safe keeping" thing, unfortunately). Keeping all the data in a database is problematic for exactly the reason you state.
How long until someone decides its ok for 3rd party researchers to access the data for 'medical study'?
For *actual* medical studies, this is pretty good - if you can provide the anonymised DNA sequence along side anonymised medical records then that gives the bioinformatics people a hell of a lot of data they can statistically analyse - that's the kind of thing that would probably greatly advance knowledge about what conditions are caused by what genetic sequences. The problem with this, of course, comes back to the fact that in order to generate this anonymised data, you probably need a non-anonymised database of it all somewhere, which could be abused.
If no compelling medical issue requires sequencing in a newborn, it is invasive and coercive to conduct it.
What if the baby might have some condition that is better treated as early as possible? Does that not count as a "compelling medical issue"?
Are blood samples not already taken from newborns anyway? If so, genetic testing doesn't seem any more "invasive".
However, I agree that indeed my concern is that the data will be kept on file and used for purposes other than medical treatment - I like the idea of the medical profession having lots of genetic data on file; I don't like the idea of the government (and by extension, police, security services, insurance companies, etc.) having that data.
I'd wager that the fundamental flaw in HTTPS is that the government has the private keys direct from the CAs. The protocol is flawed in the key management (as most are).
It's not that simple. When you get your certificate signed, you never give the CA your private key - you give them a certificate signing request instead. So by compromising a CA, the NSA can't get access to the key required to do offline decryption of an SSL session.
They *can* get the CA's own key, of course. And that would allow them to forge a new certificate that claims to be yours. They can use that certificate to perform an online man in the middle attack of your SSL sessions, and anyone validating the certificate by checking it is signed by the CA wouldn't be able to tell the difference. However, the real cert and the forged cert would be different, which means there's a reasonable chance that someone's going to spot this if its done on a large scale. For example, as far as I know, Chrome checks that the certificates Google's websites are presenting it with are actually Google's certificates, even if they appear to be correctly signed by the CA - if it sees a fake cert it reports back to Google. (This is how Google has caught compromised CAs before)
As we've seen before, if a CA is caught handing out forged certificates, their life is pretty much over - all the browser vendors immediately revoke their CA certificates. So given how easy it is to spot the forged certificate when this kind of attack is pulled on a large scale, and how bad things would go for the CA when someone spotted it, it seems unlikly that a CA would help the NSA in this way unless they were put under immense pressure.
Presumably you are talking about your experience in the US, because that kind of shit would never be allowed in the EU. You have a right to food, accommodation and compensation. I don't know why US consumers put up with being treated that way.
Whether its allowed or not is irrelevant. 2 years ago I was flying with Thompson Airlines on a Crystal holiday (Thompson and Crystal are the same company, but they operate under different brands). For the return trip, in order to save themselves some money they had decided to use a single transfer for a number of widely spaced flights - that meant we were taken to Chambery airport 4 hours before we were due to depart - that's long before the checkin desks are even open, so we couldn't even check in our luggage. Then, severe weather lead to the closure of the airport for some time.
EC directive 261/2004 requires that air carriers provide free food and meals after only a 2 hour delay, so about 2 hours after our expected departure time, they announced that the plane was landing, that there would be a rapid turn-around and waved us through to the departure area. Of course once we went through security, it rapidly became clear that the Crystal reps had lied to us and the plane was, in fact, not there. However, there were no Crystal or Thompson reps in the departure area and when airport staff phoned them they were informed that the reps were "too busy" to come and speak to us. Eventually, 3 hours later, we actually got to see a rep, because one of the other passengers got pissed off and ran back through security.
We asked for the food and drink that we should have received, and the rep told us “Thomson is aware that we have a legal obligation to provide food and drink but we are choosing not to do so”. Eventually they did agree to provide drinks but refused to provide any food. We eventually took off 11 hours after we were supposed to, having sat on the cold floor of the airport for 10 hours (because Chambery is small and there weren't enough seats). The air crew had been informed that we had all received food and drink on the ground and therefore didn't require any during the flight, so most of the passengers had to go without food until we landed at Bristol.
Upon returning to the UK, Crystal disclaimed all responsibility, stating that we had to complain directly to Thompson (even though they are the same damned company). Both Thompson and Crystal refused to apologise or compensate us, essentially claiming they had done nothing wrong. This was despite them knowing that we were a large group who booked a group holiday every year. ABTA, IATA and the CAA all refused to do anything about their flagrant disregard for the law.
So yeah, it may "not be allowed", but it still happens - the airlines ignore the law and the regulators don't punish them for it. Needless to say, Crystal/Thompson will be losing out on over ten thousand pounds of business every year, since our group will never book with them again - all because they decided to save a couple of hundred pounds by refusing to meet their legal obligations and offering utterly shite customer service both at the time of the incident and in the months afterwards.
All I was saying is that, while you CAN have a different port when you're connecting to a different address/port, there's no need to (when the internal address/port is the same). Since there's no need to, and it's easier to use a single mapping table (mapping internal IP/port to an external port, regardless of what it's connecting to), I'd think that would be the solution most (cheap) routers would use (the kind that also would only have one external IP address).
A lot of the cheap routers run Linux, which certainly doesn't do as you describe :)
Do you still want no say in operation when safety measure are cut to protect your 'investment'?
I would like people with some understanding of the job at hand to have a say. However, the vast vast majority has _no_ understanding of this stuff, react to what the papers/environmentalists/whoever say, and their voices drown out those who know what they're talking about. So my only conclusion is that no, maybe the public shouldn't have a say - it should be down to the experts. The trick there, unfortunately, is how to ensure that decisions are made by neutral experts rather than people with a vested financial interest in one thing or another (most of the MPs have their fingers in lots of the pies that they have to vote on... and they don't declare this of abstain from votes when they have a conflict of interest).
Why is Chinese government* even allowed to operate a nuclear plant in the UK?
Because consecutive governments have chosen not to make a decision on the future of British energy because they know that *all* the choices are unpopular. By the time crunch time had been and gone and a decision was long overdue, the inaction of the government had basically ensured that the UK had no nuclear industry who could feasably commission nuclear power stations. So now we're stuck having to bring in the nuclear industry from other countries who *didn't* put the whole nuclear power idea on hold for decades and therefore still have some expertise in this area.
As it stands, things ahve been left so long that there is no way the UK is going to be able to commission new power stations in time to avoid a shortfall in generation capacity and there is next to no working UK expertise WRT building nuclear power stations anymore, so what else can we do but bring in foreign corporations to help out?
I'd think the obvious solution to how to map port numbers for NAT would be to map each host/port pair to a single random external port, regardless of the address it's sending/receiving to/from. That way you have just one lookup table for the mapping each way. IF the NAT router does that, then all you do is connect to a common server, which notes your external IP address/port and passes that on to the other party (and vice versa).
This is a misunderstanding of how a lot of NATs work.
A connection is identified by a tuple of (protocol, source IP, source port, destination IP, destination port). The source IP and source port are going to be rewritten by the NAT. The NAT will maintain a table of connections, which will map:
(protocol, original source IP, original source port, destination IP, destination port) <-> (protocol, translated source IP, translated source port, destination IP, destination port)
The "translated source IP" is always going to get rewritten to the router's external address (* this is not entirely true - see below), so that bit is relatively easy - you can poke an external server and that can pass on the "translated source IP" to the peer.
The "translated source port" is arbitrary - the NAT may decide to leave it as it is, or it can pick any port it wants. And there doesn't have to be any correllation at all between any two connections. I.e. if you connect to host A, destination port B with original source port C, and then connect to host Y, destination port Z with original source port C, there is no guarantee the translated source port will be the same in both cases. So you can't just poke an external server and have it pass on the translated source port to your peer, because the translated source port you're going to use when talking to the peer could be different.
In practice, many NATs try to avoid changing the source port at all unless there is some other connection identified with the same (protocol, translated source IP, translated source port, destination IP, destination port) - then it has to change the source port. This means that you end up with things mostly working, and then randomly breaking for no apparent reason (i.e. when another machine on the network is trying to connect to the same host and has happened to pick the same ports). You can mitigate this to some extent by trying to randomise ports and addresses as much as possible, but its never going to completely go away, and furthermore there is no guarantee that the NAT will even try to keep the port the same.
(* routers can have multiple public IP addresses, so there's also no guarantee that they will always give you the same translated source IP either. Although this is rarer).
I guess this wouldn't be such a problem if there was a standardised way of doing NAT... but there isn't - there are a variety of systems employed in the real-world, and to have a robust solution you have to be able to cope with all of them. Anyway, most of this is covered by the STUN protocol, which is a method of trying to determine all of the information you need to tunnel through a NAT; and the STUN RFC specifically says it is a best effort attempt and can't be made reliable.
You'd still need to get through the firewall if the router is doing stateful blocking (which is easy to get around - there ought to be a connect_to low level call to do that, but easy enough to do a listen/connect on both sides).
Actually, stateful firewalls are pretty much a doddle to tunnel through - you essentially do exactly the same as you would for NAT, but it completely eliminates all the guesswork since you know that the addresses aren't going to change arbitrarilly. This is why peer-to-peer stuff over IPv6 works well, even when you have stateful firewalls - you just need an unfirewalled server somewhere to mediate the transaction between the peers, and you don't need to be concerned with anything in the middle modifying the traffic in ill-defin
They have these gantries on the M42 and M6 near where I live. You can see the first one from over a mile away, and they're spaced at approx 1/4 mile intervals. You can often see two ahead at least.
Fine on straight roads - the M4 through south wales has things called "bends" which obscure some of the gantries until you're a bit too close to lose 20mph of speed without hitting the brakes. As I said, all they need to do is drop the speed limit in 10mph increments instead of 20mph increments and it'd be fine.
I hate those cameras so much. I wouldn't mind them so much if the road was permanently 50, but 50-70-50-70-50-70 with cameras in between is just pure ass-hole.
Variable speed limits *do* work to improve traffic flow, and you probably do need cameras to enforce them, so I don't have a big problem with them. But changes to the limit need to be done in small enough increments that people don't have to slam on the brakes to avoid going too fast. My only minor issue with average speed cameras in a 50 limit is that the trucks seem to ignore them - trucks are generally not great at overtaking; they creep past at 52mph, and then pull in before they've actually finished overtaking. Twice I've been run off the road by trucks trying to do 52mph in a camera'd 50 zone whilst I've been religiously doing dead on 50 to avoid being ticketted.
The other issue is: what do you do when the limits are blatently stupid. I've been cruising down the M25 in the early hours of the morning before, not a thing in sight and suddenly the variable speed limit signs come on telling me to slow down to 20mph - that seems downright dangerous to me, since the road is completely clear - if someone decided to ignore the limit and come up behind me at 70mph, that'd be a serious accident waiting to happen. In that case I decided I was going to ignore the limit and do 40mph, fully expecting to have to argue my case for safety in court, luckilly I didn't get ticketted.
While everybody is writing about the water, the real issue is the spent-fuel-rod pool. If that thing is not secured very soon, Tokyo becoming uninhabitable within a very short time is a real possibility. There is so much radiation in there, it is staggering. The pool is inadequately cooled. The pool is damaged enough that even a minor earthquake could prevent cooling it more and a fire starting in there would both be impossible to put out and starting by itself very fast.
Is there really much chance of a fire? The material in the spent fuel pools has been cooling for at least 2 years already...
Personal prediction: TEPCO will go on blundering about, and eventually they will get a nuclear catastrophe that makes all others so far look like a summer breeze. After that, Japan will not play a role in the world for a few thousand years or longer, because for all intents and purposes it will not really be there anymore.
I don't think it's going to be that serious, *but* I do wonder WTF anyone is still allowing TEPCO to run the show, since they have repeatedly been shown to be covering up at every turn. At the very least the upper management should all be removed, even if they need to keep the engineers.
I can only assume that the iOS app is similar, but the Android app uploads not just your phone number (which is scraped without your explicit permission), but also your call history every time you log in.
Let me repeat that: Facebook uploads your entire call history every time you open their Android app.
I switched to the Atrium app a few months ago, having got fed up with the intrusive ads on the official app and knowing that I could never upgrade the official app because they had massively expanded the permissions it required (there was no way I was going to give it permission to do things like see what apps were running, etc). I've been pretty please with it so far - a couple of slightly niggley bugs, but on the whole its good.
Sounds like a wise choice. I know Facebook isn't the only app employees might waste your time on, with their phones. Every time I see someone's employee fiddling with their phone -- usually while not providing customer service, but always while stealing paid time -- I wonder how they manage to remain employed.
Flexibility runs both ways - if you're going to be a dick and prevent employees from taking the occasional 5 minute break (because it's "stealing" from you) then they're not going to be inclined to do anything over and above their contract either. Don't expect someone to stay late to clear up some problem (because that would be you "stealing" from them) if you're never going to return the favour.
FWIW, an old employer of mine started doing the kind of shit you're suggesting - I got a massive bollocking for ending up 5 minutes late due to traffic one morning... the previous night I had stayed 2 hours late to finish some work. Needless to say, I never stayed late again, and left the company relatively soon afterwards... in fact, most of my colleagues also got pissed off with them and quit - they lost 75% of their technical staff in a 2 month period.
Punching holes in NAT is a fucking nightmare
I thought it was just a case of getting both parties to attempt to send packets to each other on an agreed UDP port - that's how I thought Skype did it anyway.
Depends on the NAT. There is absolutely no guarantee that the NAT won't alter your source port (which means the traffic from the other end that is directed at that port won't get to you). This stuff isn't too bad if one end is behind a NAT, but it becomes unreliable once you have both ends behind a NAT because there's just no way to guarantee the NAT will behave how you need it to behave.
What kind of grade A moron would make a car brake to match a speed limit? It'll just turn the engine off to slowly reduce speed.
The same grade-A idiots that make you brake by reducing the speed limit too sharply to meet by engine braking alone and then whacks a speed camera in to catch anyone who didn't slam on the brakes.
An example I drive with reasonable frequency - the variable speed limit on the west-bound M4 near Newport, Wales. When they decide to reduce the limit to 50, the first 50 sign you see is too close to slow from 70 before passing it without braking. And every other gantry has a set of speed cameras in it, so you've basically got to hit the brakes on a motorway to avoid getting a NIP. This could be easilly solved by making the first sign a 60, and the next one a 50 to give you plenty of slowing down time.