Thanks. It looks a bit better than some of the options out there, but it is still missing email and documents (word processing, spreadsheets, presentations). So, it is hardly a google apps replacement, but certainly it is going in the right direction.
cant pass legislation to protect itself from school shootings
I live in a country where school shootings are, as someone said above, a statistic anomaly and yet, it's not because of laws that protect us from them but because society, as such, does provide much less cause for them. Any law in the world won't be enough. It's much harder than that.
Yup. MAYBE there is a law somewhere that would help with that, but it isn't like you can just make murder illegal and solve the problem. I'd also say that the US is not unique in having problems like this. Sure, the EU doesn't have so many guns floating around, but look at all the issues France was having with conflict between various religious communities lately.
These sorts of issues are cultural in nature. For starters, people need to not have a laundry list of issues they're willing to kill anybody for disagreeing with them on.
Name a feature you want and it exists...What you want to do is archive, delete, or send to spam on a single keystroke?...I suppose you're looking for a good webmail client?...But I'm sure you could find a good webmail client that is FOSS if you wanted.
So, obviously you've never looked for them. I have. The best options right now are Roundcube and Squirrelmail, or the less-FOSS Zimbra. None of them let you archive/delete/spam email with a single keystroke, and I don't think any of them support tag-based email either. That function in Gmail lets me blast through an inbox in about a minute or two, has an offline cached client for Android, and works in a browser.
A proper email client donkey stomps gmails webclient and always has.
And it won't work on a Chromebook or a mobile device with only a browser.
The vast majority of mail that arrives at my email accounts is automatically sorted. I can receive hundreds of mails in a day and know what I got that matters in about 5 seconds...And that is entirely independent of the server.
If you're doing it on a client, then it is useless when you're not using that particular client. That's the whole point of the cloud - you're not tied to one client. MAYBE I could get by with a curses-based email client over a terminal, but giving up a GUI seems like a poor move anytime after around 1990.
My email is all sorted as well, typically in more than one way since I'm using tags. Stuff I follow goes in the inbox, stuff I browse more by group doesn't go into the inbox.
How about sense-and-avoid in combination with ADS-B? This article [gcn.com] suggests that people are working in that direction.
Perhaps for long-haul, larger UAS platforms (like freight haulers, or long-mission mapping systems and whatnot). But do you really think that a contractor who uses a 3-pound plastic quadcopter to checkout the top of a residential chimney for 90 seconds a couple of times a week needs an ADS-B enabled platform? It's just craziness.
There is no reason that ADS-B has to cost more than $30. Just have the government bless a reference platform instead of having everybody invent their own.
ADS-B is a GPS and a modem. Guess what you'll find on every SoC in every cellphone sold today?
I have access to my home movie and music library anywhere, can remote into my home systems whenever I want from my phone, and can host any file I want on line without having to give it to a third party.
There is a lot more to the cloud than a page full of links behind.htaccess or whatever.
I'd love to self-host, but I don't see any FOSS options that are equivalent to the likes of Gmail or Google Docs or Google Music. There are some web-based email applications, but they're pretty weak. I've yet to find one that lets me archive/delete/spam an email with a single keystroke.
ownCloud 8 on my Raspberry Pi is working just fine for me.
If only. It is lacking most of the features of Gmail/Google Docs/Google Play/Google Music.
I'd really love to have open-source alternatives to the cloud. The problem is that the best anybody seems to come up with are X11 apps plus some kind of dropbox synchronizer or something. If it doesn't work entirely from a browser, then it is a non-starter.
The certificate and rules sound mostly good. A private pilot's license isn't a commercial license, it's fairly easy to get, but ensures that you know the 'rules of the air' like a person with a driver's license presumably knows the rules of the road.
The rules of the air are useless 400' above the earth - they're only designed to allow pilots to operate around other aircraft, and these drones won't operate there, at least not for now. Frankly, if they ever do operate autonomously at altitude the rules of the air that exist today will be worthless anyway.
A pilot's license is also not easy to get. For starters, you generally need to not be terrified of flying in a small plane, which by itself is something probably half the US population would fail to satisfy. Mostly it is about stick and rudder skills that are useless in a drone, and which also not everybody has. It is also very expensive - at a minimum you need about 40 hours of flight time, much of it accompanied by an instructor. Planes rent for about $120/hr in most places, if you want something from the 70s.
while self-driving cars are allowed on the road for testing, unless they're on a closed track they still need a rated human driver available to take over if something goes wrong, etc...
These drones aren't being allowed "on the road" - they're being restricted to below 400' and in fairly uninhabited areas. You're legally allowed to operate a fully-autonomous vehicle in your back yard without any human inside as long as it doesn't go out onto the roads. Otherwise you'd need a license to buy a Roomba.
If Amazon wanted to test landing a drone at KSFO then certainly requiring an instrument-rated pilot at the helm would make a lot more sense - they're going to be in controlled airspace talking to ATC and essentially operating under IFR in an environment where a mistake might kill hundreds of people. This isn't that.
Currently, 14 CFR 91.113(b) imposes a requirement on all aircraft operations that, during flight, “vigilance shall be maintained by each person operating an aircraft so as to see and avoid other aircraft.” This see-and-avoid requirement is at the heart of the FAA's regulatory structure mitigating the risk of aircraft colliding in midair. As such, in crafting this proposed rule, the FAA sought a standard under which the small UAS operator would have the ability to see and avoid other aircraft similar to that of a manned-aircraft pilot.
This really strikes me as the wrong way to go about this. Sure, changing this principle would require completely rewriting the rules for ALL forms of air travel. However, I think that see and avoid already works poorly in practice and will become even more untenable once drone technology really takes off. There are far better ways for aircraft to avoid each other, and sooner or later we'll need to come to grips with the fact that you can build a device capable of broadcasting its position to nearby aircraft for the cost of a smartphone (you need a GPS and a radio, which virtually every phone in the US has had for over a decade due to 911 compliance rules - they had them long before they had touchscreens). There is no reason that clearances shouldn't be communicated via modems talking to computers and followed by computers to ensure that airways remain free of conflicts. Humans just mess things up. If you want to have them so that when they get disoriented they can override the computer and crash the plane to give the investigators more work, just find someplace else to fly.:)
The FAA doesn't care about the rabbit, it cares about the people. Which it's equally likely to hit if it crashes.
They're testing the drones in a low-density area. They could have just said "avoid flying over people" and left it at that.
A private pilot's license isn't that high of a bar, and it's pretty much the lowest bar the FAA has. It just ensures that the operator knows the 'rules of the air'. Sure, some of the knowledge is useless, like some of the stuff in my driver's test I'm never going to use. Same with the medical certificate, because if the drone operator croaks, it might crash before they can get another operator there. Remember, prototype. It's easier to relax restrictions than it is to crank them up.
Which of those rules of the air are relevant to flying a drone flying under 500'? About the only rule I can think of is the one that tells pilots not to fly below 500', which they're intentionally violating. Knowing which way to turn if you spot a crop-duster heading towards you isn't going to be a big help when you're not displaying navigation lights and the crop-duster has no way to know which way you're headed and won't see you anyway.
The problem is that this kind of thing stifles innovation, which means that all the R&D ends up moving overseas and the US will end up being perpetually behind in what is eventually likely to be a technology that completely replaces all aviation today (civil or military). Is that in the US interest? I'm not saying that Amazon should be flying two-ton drones over major cities without any oversight, but this is about testing concepts in the middle of nowhere.
It probably makes more sense to ensure that the people programming the drones understand the rules of flying, and then only if they're actually going to fly in conflict with other aircraft. In that situation the rules would probably need to be changed anyway, since "see and avoid" isn't going to work when you can't see the drone.
If they had different requirements ("sedentary medical certification" for example), then that would represent a heapload of additional work for them, cost for the taxpayer, and, as this is an EXPERIMENTAL PROGRAM, potentially wasted effort.
They could have different requirements like "just keep it under 400 feet and away from populated areas and do whatever you want as long as OSHA doesn't have a problem with it." I don't see how that would cost taxpayers money, or be any less safe.
equally likely? wtf. you mean to tell me there is as many bunny rabbits that are customers of amazon and living in densely populated cities of millions as there are people? try again.
The FAA didn't give Amazon permission to operate drones in cities. It gave them permission to operate drones in the middle of nowhere, where there probably are more bunnies than people walking around.
Let me get this. The FAA doesn't consider it safe for Amazon to play around with a drone outdoors in a rural area under 400' altitude unless:
1. It is in VFR conditions. 2. The specific design is pre-approved by the FAA. 3. The drone operation holds a private pilot license. 4. The drone operator holds a medical certificate.
Wouldn't want to have the operator die of a heart attack, and then have the drone go out of control and hit a rabbit. Wouldn't want the pilot sitting in a chair and looking at a monitor to have medical conditions that cause issues with balance and vertigo. Better make sure that the pilot staring at the monitor 12" from his face has good vision.
I get that they want to keep these things close to earth and away from airports. I don't get why you need to be able to glide a plane onto the runway during an engine failure in the landing pattern when you're probably flying a drone that is incapable of gliding at all and which is multi-engine besides.
Flying drones and flying planes are completely different skillsets. The FAA really needs to get away from making drone piloting an add-on to a private pilot license.
..and many companies burn through their capital on their 3rd attempt at failure. Failure isn't the goal. Forward progress is the goal. Recognizing failure or impasse quickly and cutting losses is the goal.
Isn't that basically the whole point of the article? Google tries new things, and abandons them if they don't work out.
You learn no more from failure than you learn from success. There are many ways to fail and few ways to succeed, thus it is better to learn what to do than what not to do.
Sure, but the point is that you often can't do the one without the other. Fear of failure tends to cause companies to just stick with what they already do well. That means they basically aren't learning anything at all.
Of course companies should fund the projects that they think are most likely to become profitable. They'll still fail at some of them, and willingness to embrace that increases the odds that they'll come up with something truly innovative.
A big problem with software project management is that software projects are rarely equivalent to building houses/bridges/etc. You can't just look at the square footage and guess how long it will take to build.
I agree that quality management is a value-add in any industry. However, that doesn't mean that you can manage a sports team the same way as a construction team or an orchestra or an art gallery or an ERP implementation team or a video game studio.
I do not agree in the slightest that your ability to be safe (get out of town in an emergency) should be determined solely by your ability to pay. Maybe we should also let the firefighters decide they don't really want to work today, so if you have a fire we will only work if you pay us extra.
Are you saying that individual employees should be compelled under force of law to perform a job against their will?
The $10 cabs that are already 100% saturated will still be around to not give you a cheap ride out of town even if the option to ride uber for $100 exists.
When you cap prices, it inevitably leads to shortages one way or another. The only time it makes sense is in monopoly situations. The whole point of services like Uber is to get rid of the monopolies. There is no reason there can't be 25 other similar competing services, which means that prices will generally reflect market conditions.
And in a true disaster situation, the national guard should be getting everybody out free of charge anyway. I'm referring to situations that raise prices where the national guard isn't giving everybody free rides. Clearly in those situations Uber couldn't charge a lot of money, since who would pay $200 for a ride the government would give them for free?
The user sees the license so he knows that the person has a valid taxi license. Having a valid taxi license means things like he has not been convicted of a crime.
And that is necessary for taxis, since as a passenger all you see is a random yellow-painted car pull up which might or might not be a legally-licensed taxi. Also, it is important that they haven't been convicted of a crime, since the driver could shoot you and dump you in a ditch and nobody would know you had even gotten into the car.
With a reservation system like Uber the car that shows up is pre-identified, so you don't need any kind of further validation. Likewise, your trip was recorded in advance, so it would be impossible for the driver to do something to you without being detected, which is a deterrent to crime. Sure, he could do something to you anyway, but anybody driving a car could run you over on the sidewalk too, and we can't regulate everything.
There is no area of the economy that has a 10x (let alone 50x) fluctation in price during the course of a day. Most areas would not see a 10x fluctuation in price over a decade. And when, by law, 3/4 of the cabs are on the streets at all times, there is no need for surge pricing.
I'm sure plane tickets fluctuate in value by factors this large. People can easily pay 10x more/less for the same ticket depending on when they buy it.
I don't get the problem - why wouldn't you want more people driving people around when there is more demand for drivers? If there is an emergency and I have to get out of town, I'd rather pay $100 for a ride that gets me there than try to hail one of the already-in-use cabs for a $10 ride.
How many of those regulations really offer any value in a reservation system like Uber? Why have the license in a lighted frame when you can just display it to the user on their phone or whatever? Why does the user need to see the license anyway?
Many of those regulations herald back to the day when enforcement of the law and recordkeeping was difficult, so they had to try to keep tabs on what was going on and deputize the citizenry.
I see variable fares as a benefit. Why not have reserve drivers who are willing to step in when things get busy for a higher rate? If somebody was paying $500/hr I'd be happy to get in my car and give people a lift, but I'm not going to do it for $10/hr. That is how virtually every other area in the economy works.
Rootkitted? I think that's unlikely, even the NSA isn't omnipotent. But do I think they are monitoring Tor nodes from their own nodes, probably. They'd only rootkit you if they wanted specific data from a specific person-of-interest, I think.
Well, it was already divulged that they root sysadmins to get credentials to log into boxes even when the sysadmins themselves are of no interest to them otherwise. (Ie, ISIS has a website hosted by AWS, so they find some random Amazon employee who VPNs in from home and steal their keys or such.)
I have no idea if they're rootkitting tor nodes, but it seems like a fairly obvious way to circumvent the tor network. If you have root on most of the nodes, then you can trivially follow most of the traffic.
The big thing with the NSA's hacking efforts is that it is largely automated. If they stick a list of tor IPs into a database, they'll all get hacked automatically, and then managed automatically. If somebody does a security update and only 3 of their 5 backdoors are left intact somebody will be notified to step in and open two more or whatever. It is basically what you'd get if you combined anonymous with a competent sysadmin team and a bunch of security researchers and then a bazillion SMEs to make sense of the extracted data. It probably costs as much for the NSA to hack into another PC as it costs Amazon to spool up another virtual machine.
How on earth does Apple Pay have more simplicity than a credit card? Here's how it works with a credit card:
1. Touch card or even whole wallet on reader. 2. Done!
And for more expensive transactions (over 20GBP, soon to be 30):
1. Insert card. 2. Enter PIN. 3. Done.
It doesn't get much simpler than the first one, really. I don't even have to extract my card.
The problem is that it doesn't work this way in the US. The way it works here is:
1. Usually swipe card. MAYBE use a contactless system. 2. Sign piece of paper. 3. When you leave the merchant keeps a copy of all the data required to impersonate you in future transactions.
A big part of Google Wallet and Apple Pay is getting rid of #3. In Europe chip and pin protects you against #3 for the most part - the merchant doesn't get all the credentials necessary for future transactions as some of them stay in the chip.
I like the looks of Apple Pay, and think it's a great move forward but even as an Apple fan, it seems bizarre for Apple to move forward on their own payment standard rather than the industry creating one.
Scenario 1 - Apple creates new system. If it takes off, Apple makes more money.
Scenario 2 - industry creates a new system. If it takes off, industry makes the same money they have always made. Industry is mostly run by MBAs who don't realize that if somebody else does scenario 1 they might become obsolete.
Can you see why scenario 2 not happening isn't as bizarre as you might think it should be? We're also talking about banking - hardly the most progressive industry on the planet. We still use paper checks in the US and think that taking a few days to do an electronic funds transfer is normal.
So in other words - UberX in NYC obeys insurance and safety laws, while in other states it manages to be cheaper by skipping around them?
They claim they have insurance elsewhere. They're just not licensed with a particular government agency.
I think the safety law issue is a red herring. The reason you need to license cabs is that they make trips without any kind of recording of the event. If I get in a cab and pick you up on the curb, I could kill you and leave you in a gutter someplace and nobody would know you were ever in my cab. Systems that involve reservations are inherently more secure, because there is traceability.
I think that these kinds of systems need to be regulated differently. Applying the same rules as for taxis that can be hailed just drives up costs without providing any real safety improvement.
I hate saying this, but do you "really" need to access your e-mail on more than one machine these days, with the ubiquity of phones and tablets? If you're away from your "home machine", then use the tablet/phone. Android DOES have e-mail clients that support IMAP and gnupg
Most of the IMAP clients I've seen aren't terrific about offline access - certainly not in comparison to Gmail. They also don't handle tagging well (the same email being in 10 different "folders" at the same time). I also use multiple computers at home, and some are based on ChromeOS.
I'd like to have my cake and eat it too, but there aren't a lot of threat models that thunderbird+gpg protects against that Gmail doesn't. It certainly won't stop the NSA from snooping on your email if they care to.
Sure it will stop the NSA, they can't break pgp/gpg.
They can extract the key from my PC though. The one I'm typing this on happens to run a tor relay node. How likely do you think it is that it isn't rootkitted, despite religious application of patches and generally following best practices for linux?
And you might want to replace your expired gpg pubkey on slashdot with your current one: the one with Key ID: 55EC123A Key fingerprint = 3665 3E11 22C0 8BCE A16D 1529 08C1 70DE 55EC 123A
Thanks. I'd forgotten that Slashdot even tracks such things.:)
I'd love to see a decent FOSS webmail application that supports encryption. The only options that exist right now are pretty weak compared to something like GMail.
Then use Gmail over IMAP with a proper e-mail client that does support encryption. GMail has had IMAP support for over 7 years.
I migrated away from X11 clients and IMAP for a reason. I used to GPG sign everything I sent. The problem is that when you start using multiple computers/operating-systems it becomes a real PITA. It wasn't like I was using encryption anyway, since 95% of the people I communicate with don't use it themselves.
I'd like to have my cake and eat it too, but there aren't a lot of threat models that thunderbird+gpg protects against that Gmail doesn't. It certainly won't stop the NSA from snooping on your email if they care to.
How do you store and access those keys safely and securely from a web browser? That's what most of the existing webmail plugin solutions for pgp/gpg/smime do - they just provide a local keystore and make that available to the js methods to do the work.
Oh, I get that. However, there are a few options:
1. Store it on the server. That doesn't protect you against server compromise, but it still protects you against a lot of stuff, and there is no reason the server can't be as secure as your client would otherwise be. Keep in mind that webmail does not necessarily mean 3rd party provided.
2. HTML5 storage plus javascript. This can in theory be about as secure as a local client, but my big concern here is having some way to ensure that the javascript isn't tampered with. In practice I think this is no better than #1. Google can write the best local storage system in the world and make it airtight, but if somebody compromises gmail.com they'll just send you a modified javascript file when you browse the site which will send them a copy of your keys when you unlock them.
I'd really like to see a good FOSS webmail client so that I can host my own.
Slashdot Mirror
Try Cozy then.
Thanks. It looks a bit better than some of the options out there, but it is still missing email and documents (word processing, spreadsheets, presentations). So, it is hardly a google apps replacement, but certainly it is going in the right direction.
cant pass legislation to protect itself from school shootings
I live in a country where school shootings are, as someone said above, a statistic anomaly and yet, it's not because of laws that protect us from them but because society, as such, does provide much less cause for them. Any law in the world won't be enough. It's much harder than that.
Yup. MAYBE there is a law somewhere that would help with that, but it isn't like you can just make murder illegal and solve the problem. I'd also say that the US is not unique in having problems like this. Sure, the EU doesn't have so many guns floating around, but look at all the issues France was having with conflict between various religious communities lately.
These sorts of issues are cultural in nature. For starters, people need to not have a laundry list of issues they're willing to kill anybody for disagreeing with them on.
Name a feature you want and it exists...What you want to do is archive, delete, or send to spam on a single keystroke?...I suppose you're looking for a good webmail client?...But I'm sure you could find a good webmail client that is FOSS if you wanted.
So, obviously you've never looked for them. I have. The best options right now are Roundcube and Squirrelmail, or the less-FOSS Zimbra. None of them let you archive/delete/spam email with a single keystroke, and I don't think any of them support tag-based email either. That function in Gmail lets me blast through an inbox in about a minute or two, has an offline cached client for Android, and works in a browser.
A proper email client donkey stomps gmails webclient and always has.
And it won't work on a Chromebook or a mobile device with only a browser.
The vast majority of mail that arrives at my email accounts is automatically sorted. I can receive hundreds of mails in a day and know what I got that matters in about 5 seconds...And that is entirely independent of the server.
If you're doing it on a client, then it is useless when you're not using that particular client. That's the whole point of the cloud - you're not tied to one client. MAYBE I could get by with a curses-based email client over a terminal, but giving up a GUI seems like a poor move anytime after around 1990.
My email is all sorted as well, typically in more than one way since I'm using tags. Stuff I follow goes in the inbox, stuff I browse more by group doesn't go into the inbox.
How about sense-and-avoid in combination with ADS-B? This article [gcn.com] suggests that people are working in that direction.
Perhaps for long-haul, larger UAS platforms (like freight haulers, or long-mission mapping systems and whatnot). But do you really think that a contractor who uses a 3-pound plastic quadcopter to checkout the top of a residential chimney for 90 seconds a couple of times a week needs an ADS-B enabled platform? It's just craziness.
There is no reason that ADS-B has to cost more than $30. Just have the government bless a reference platform instead of having everybody invent their own.
ADS-B is a GPS and a modem. Guess what you'll find on every SoC in every cellphone sold today?
I have access to my home movie and music library anywhere, can remote into my home systems whenever I want from my phone, and can host any file I want on line without having to give it to a third party.
There is a lot more to the cloud than a page full of links behind .htaccess or whatever.
I'd love to self-host, but I don't see any FOSS options that are equivalent to the likes of Gmail or Google Docs or Google Music. There are some web-based email applications, but they're pretty weak. I've yet to find one that lets me archive/delete/spam an email with a single keystroke.
ownCloud 8 on my Raspberry Pi is working just fine for me.
If only. It is lacking most of the features of Gmail/Google Docs/Google Play/Google Music.
I'd really love to have open-source alternatives to the cloud. The problem is that the best anybody seems to come up with are X11 apps plus some kind of dropbox synchronizer or something. If it doesn't work entirely from a browser, then it is a non-starter.
The certificate and rules sound mostly good. A private pilot's license isn't a commercial license, it's fairly easy to get, but ensures that you know the 'rules of the air' like a person with a driver's license presumably knows the rules of the road.
The rules of the air are useless 400' above the earth - they're only designed to allow pilots to operate around other aircraft, and these drones won't operate there, at least not for now. Frankly, if they ever do operate autonomously at altitude the rules of the air that exist today will be worthless anyway.
A pilot's license is also not easy to get. For starters, you generally need to not be terrified of flying in a small plane, which by itself is something probably half the US population would fail to satisfy. Mostly it is about stick and rudder skills that are useless in a drone, and which also not everybody has. It is also very expensive - at a minimum you need about 40 hours of flight time, much of it accompanied by an instructor. Planes rent for about $120/hr in most places, if you want something from the 70s.
while self-driving cars are allowed on the road for testing, unless they're on a closed track they still need a rated human driver available to take over if something goes wrong, etc...
These drones aren't being allowed "on the road" - they're being restricted to below 400' and in fairly uninhabited areas. You're legally allowed to operate a fully-autonomous vehicle in your back yard without any human inside as long as it doesn't go out onto the roads. Otherwise you'd need a license to buy a Roomba.
If Amazon wanted to test landing a drone at KSFO then certainly requiring an instrument-rated pilot at the helm would make a lot more sense - they're going to be in controlled airspace talking to ATC and essentially operating under IFR in an environment where a mistake might kill hundreds of people. This isn't that.
If you feel strongly enough about this, you can read the actual proposals and comment on them here.
Meh, they're keeping "see and avoid."
Currently, 14 CFR 91.113(b) imposes a requirement on all aircraft operations that, during flight, “vigilance shall be maintained by each person operating an aircraft so as to see and avoid other aircraft.” This see-and-avoid requirement is at the heart of the FAA's regulatory structure mitigating the risk of aircraft colliding in midair. As such, in crafting this proposed rule, the FAA sought a standard under which the small UAS operator would have the ability to see and avoid other aircraft similar to that of a manned-aircraft pilot.
This really strikes me as the wrong way to go about this. Sure, changing this principle would require completely rewriting the rules for ALL forms of air travel. However, I think that see and avoid already works poorly in practice and will become even more untenable once drone technology really takes off. There are far better ways for aircraft to avoid each other, and sooner or later we'll need to come to grips with the fact that you can build a device capable of broadcasting its position to nearby aircraft for the cost of a smartphone (you need a GPS and a radio, which virtually every phone in the US has had for over a decade due to 911 compliance rules - they had them long before they had touchscreens). There is no reason that clearances shouldn't be communicated via modems talking to computers and followed by computers to ensure that airways remain free of conflicts. Humans just mess things up. If you want to have them so that when they get disoriented they can override the computer and crash the plane to give the investigators more work, just find someplace else to fly. :)
The FAA doesn't care about the rabbit, it cares about the people. Which it's equally likely to hit if it crashes.
They're testing the drones in a low-density area. They could have just said "avoid flying over people" and left it at that.
A private pilot's license isn't that high of a bar, and it's pretty much the lowest bar the FAA has. It just ensures that the operator knows the 'rules of the air'. Sure, some of the knowledge is useless, like some of the stuff in my driver's test I'm never going to use. Same with the medical certificate, because if the drone operator croaks, it might crash before they can get another operator there. Remember, prototype. It's easier to relax restrictions than it is to crank them up.
Which of those rules of the air are relevant to flying a drone flying under 500'? About the only rule I can think of is the one that tells pilots not to fly below 500', which they're intentionally violating. Knowing which way to turn if you spot a crop-duster heading towards you isn't going to be a big help when you're not displaying navigation lights and the crop-duster has no way to know which way you're headed and won't see you anyway.
The problem is that this kind of thing stifles innovation, which means that all the R&D ends up moving overseas and the US will end up being perpetually behind in what is eventually likely to be a technology that completely replaces all aviation today (civil or military). Is that in the US interest? I'm not saying that Amazon should be flying two-ton drones over major cities without any oversight, but this is about testing concepts in the middle of nowhere.
It probably makes more sense to ensure that the people programming the drones understand the rules of flying, and then only if they're actually going to fly in conflict with other aircraft. In that situation the rules would probably need to be changed anyway, since "see and avoid" isn't going to work when you can't see the drone.
If they had different requirements ("sedentary medical certification" for example), then that would represent a heapload of additional work for them, cost for the taxpayer, and, as this is an EXPERIMENTAL PROGRAM, potentially wasted effort.
They could have different requirements like "just keep it under 400 feet and away from populated areas and do whatever you want as long as OSHA doesn't have a problem with it." I don't see how that would cost taxpayers money, or be any less safe.
equally likely? wtf. you mean to tell me there is as many bunny rabbits that are customers of amazon and living in densely populated cities of millions as there are people? try again.
The FAA didn't give Amazon permission to operate drones in cities. It gave them permission to operate drones in the middle of nowhere, where there probably are more bunnies than people walking around.
Let me get this. The FAA doesn't consider it safe for Amazon to play around with a drone outdoors in a rural area under 400' altitude unless:
1. It is in VFR conditions.
2. The specific design is pre-approved by the FAA.
3. The drone operation holds a private pilot license.
4. The drone operator holds a medical certificate.
Wouldn't want to have the operator die of a heart attack, and then have the drone go out of control and hit a rabbit. Wouldn't want the pilot sitting in a chair and looking at a monitor to have medical conditions that cause issues with balance and vertigo. Better make sure that the pilot staring at the monitor 12" from his face has good vision.
I get that they want to keep these things close to earth and away from airports. I don't get why you need to be able to glide a plane onto the runway during an engine failure in the landing pattern when you're probably flying a drone that is incapable of gliding at all and which is multi-engine besides.
Flying drones and flying planes are completely different skillsets. The FAA really needs to get away from making drone piloting an add-on to a private pilot license.
..and many companies burn through their capital on their 3rd attempt at failure. Failure isn't the goal. Forward progress is the goal. Recognizing failure or impasse quickly and cutting losses is the goal.
Isn't that basically the whole point of the article? Google tries new things, and abandons them if they don't work out.
You learn no more from failure than you learn from success. There are many ways to fail and few ways to succeed, thus it is better to learn what to do than what not to do.
Sure, but the point is that you often can't do the one without the other. Fear of failure tends to cause companies to just stick with what they already do well. That means they basically aren't learning anything at all.
Of course companies should fund the projects that they think are most likely to become profitable. They'll still fail at some of them, and willingness to embrace that increases the odds that they'll come up with something truly innovative.
A big problem with software project management is that software projects are rarely equivalent to building houses/bridges/etc. You can't just look at the square footage and guess how long it will take to build.
I agree that quality management is a value-add in any industry. However, that doesn't mean that you can manage a sports team the same way as a construction team or an orchestra or an art gallery or an ERP implementation team or a video game studio.
I do not agree in the slightest that your ability to be safe (get out of town in an emergency) should be determined solely by your ability to pay. Maybe we should also let the firefighters decide they don't really want to work today, so if you have a fire we will only work if you pay us extra.
Are you saying that individual employees should be compelled under force of law to perform a job against their will?
The $10 cabs that are already 100% saturated will still be around to not give you a cheap ride out of town even if the option to ride uber for $100 exists.
When you cap prices, it inevitably leads to shortages one way or another. The only time it makes sense is in monopoly situations. The whole point of services like Uber is to get rid of the monopolies. There is no reason there can't be 25 other similar competing services, which means that prices will generally reflect market conditions.
And in a true disaster situation, the national guard should be getting everybody out free of charge anyway. I'm referring to situations that raise prices where the national guard isn't giving everybody free rides. Clearly in those situations Uber couldn't charge a lot of money, since who would pay $200 for a ride the government would give them for free?
The user sees the license so he knows that the person has a valid taxi license. Having a valid taxi license means things like he has not been convicted of a crime.
And that is necessary for taxis, since as a passenger all you see is a random yellow-painted car pull up which might or might not be a legally-licensed taxi. Also, it is important that they haven't been convicted of a crime, since the driver could shoot you and dump you in a ditch and nobody would know you had even gotten into the car.
With a reservation system like Uber the car that shows up is pre-identified, so you don't need any kind of further validation. Likewise, your trip was recorded in advance, so it would be impossible for the driver to do something to you without being detected, which is a deterrent to crime. Sure, he could do something to you anyway, but anybody driving a car could run you over on the sidewalk too, and we can't regulate everything.
There is no area of the economy that has a 10x (let alone 50x) fluctation in price during the course of a day. Most areas would not see a 10x fluctuation in price over a decade. And when, by law, 3/4 of the cabs are on the streets at all times, there is no need for surge pricing.
I'm sure plane tickets fluctuate in value by factors this large. People can easily pay 10x more/less for the same ticket depending on when they buy it.
I don't get the problem - why wouldn't you want more people driving people around when there is more demand for drivers? If there is an emergency and I have to get out of town, I'd rather pay $100 for a ride that gets me there than try to hail one of the already-in-use cabs for a $10 ride.
How many of those regulations really offer any value in a reservation system like Uber? Why have the license in a lighted frame when you can just display it to the user on their phone or whatever? Why does the user need to see the license anyway?
Many of those regulations herald back to the day when enforcement of the law and recordkeeping was difficult, so they had to try to keep tabs on what was going on and deputize the citizenry.
I see variable fares as a benefit. Why not have reserve drivers who are willing to step in when things get busy for a higher rate? If somebody was paying $500/hr I'd be happy to get in my car and give people a lift, but I'm not going to do it for $10/hr. That is how virtually every other area in the economy works.
Rootkitted? I think that's unlikely, even the NSA isn't omnipotent. But do I think they are monitoring Tor nodes from their own nodes, probably. They'd only rootkit you if they wanted specific data from a specific person-of-interest, I think.
Well, it was already divulged that they root sysadmins to get credentials to log into boxes even when the sysadmins themselves are of no interest to them otherwise. (Ie, ISIS has a website hosted by AWS, so they find some random Amazon employee who VPNs in from home and steal their keys or such.)
I have no idea if they're rootkitting tor nodes, but it seems like a fairly obvious way to circumvent the tor network. If you have root on most of the nodes, then you can trivially follow most of the traffic.
The big thing with the NSA's hacking efforts is that it is largely automated. If they stick a list of tor IPs into a database, they'll all get hacked automatically, and then managed automatically. If somebody does a security update and only 3 of their 5 backdoors are left intact somebody will be notified to step in and open two more or whatever. It is basically what you'd get if you combined anonymous with a competent sysadmin team and a bunch of security researchers and then a bazillion SMEs to make sense of the extracted data. It probably costs as much for the NSA to hack into another PC as it costs Amazon to spool up another virtual machine.
How on earth does Apple Pay have more simplicity than a credit card? Here's how it works with a credit card:
1. Touch card or even whole wallet on reader.
2. Done!
And for more expensive transactions (over 20GBP, soon to be 30):
1. Insert card.
2. Enter PIN.
3. Done.
It doesn't get much simpler than the first one, really. I don't even have to extract my card.
The problem is that it doesn't work this way in the US. The way it works here is:
1. Usually swipe card. MAYBE use a contactless system.
2. Sign piece of paper.
3. When you leave the merchant keeps a copy of all the data required to impersonate you in future transactions.
A big part of Google Wallet and Apple Pay is getting rid of #3. In Europe chip and pin protects you against #3 for the most part - the merchant doesn't get all the credentials necessary for future transactions as some of them stay in the chip.
I like the looks of Apple Pay, and think it's a great move forward but even as an Apple fan, it seems bizarre for Apple to move forward on their own payment standard rather than the industry creating one.
Scenario 1 - Apple creates new system. If it takes off, Apple makes more money.
Scenario 2 - industry creates a new system. If it takes off, industry makes the same money they have always made. Industry is mostly run by MBAs who don't realize that if somebody else does scenario 1 they might become obsolete.
Can you see why scenario 2 not happening isn't as bizarre as you might think it should be? We're also talking about banking - hardly the most progressive industry on the planet. We still use paper checks in the US and think that taking a few days to do an electronic funds transfer is normal.
So in other words - UberX in NYC obeys insurance and safety laws, while in other states it manages to be cheaper by skipping around them?
They claim they have insurance elsewhere. They're just not licensed with a particular government agency.
I think the safety law issue is a red herring. The reason you need to license cabs is that they make trips without any kind of recording of the event. If I get in a cab and pick you up on the curb, I could kill you and leave you in a gutter someplace and nobody would know you were ever in my cab. Systems that involve reservations are inherently more secure, because there is traceability.
I think that these kinds of systems need to be regulated differently. Applying the same rules as for taxis that can be hailed just drives up costs without providing any real safety improvement.
I hate saying this, but do you "really" need to access your e-mail on more than one machine these days, with the ubiquity of phones and tablets? If you're away from your "home machine", then use the tablet/phone. Android DOES have e-mail clients that support IMAP and gnupg
Most of the IMAP clients I've seen aren't terrific about offline access - certainly not in comparison to Gmail. They also don't handle tagging well (the same email being in 10 different "folders" at the same time). I also use multiple computers at home, and some are based on ChromeOS.
I'd like to have my cake and eat it too, but there aren't a lot of threat models that thunderbird+gpg protects against that Gmail doesn't. It certainly won't stop the NSA from snooping on your email if they care to.
Sure it will stop the NSA, they can't break pgp/gpg.
They can extract the key from my PC though. The one I'm typing this on happens to run a tor relay node. How likely do you think it is that it isn't rootkitted, despite religious application of patches and generally following best practices for linux?
And you might want to replace your expired gpg pubkey on slashdot with your current one: the one with Key ID: 55EC123A Key fingerprint = 3665 3E11 22C0 8BCE A16D 1529 08C1 70DE 55EC 123A
Thanks. I'd forgotten that Slashdot even tracks such things. :)
I'd love to see a decent FOSS webmail application that supports encryption. The only options that exist right now are pretty weak compared to something like GMail.
Then use Gmail over IMAP with a proper e-mail client that does support encryption. GMail has had IMAP support for over 7 years.
I migrated away from X11 clients and IMAP for a reason. I used to GPG sign everything I sent. The problem is that when you start using multiple computers/operating-systems it becomes a real PITA. It wasn't like I was using encryption anyway, since 95% of the people I communicate with don't use it themselves.
I'd like to have my cake and eat it too, but there aren't a lot of threat models that thunderbird+gpg protects against that Gmail doesn't. It certainly won't stop the NSA from snooping on your email if they care to.
How do you store and access those keys safely and securely from a web browser? That's what most of the existing webmail plugin solutions for pgp/gpg/smime do - they just provide a local keystore and make that available to the js methods to do the work.
Oh, I get that. However, there are a few options:
1. Store it on the server. That doesn't protect you against server compromise, but it still protects you against a lot of stuff, and there is no reason the server can't be as secure as your client would otherwise be. Keep in mind that webmail does not necessarily mean 3rd party provided.
2. HTML5 storage plus javascript. This can in theory be about as secure as a local client, but my big concern here is having some way to ensure that the javascript isn't tampered with. In practice I think this is no better than #1. Google can write the best local storage system in the world and make it airtight, but if somebody compromises gmail.com they'll just send you a modified javascript file when you browse the site which will send them a copy of your keys when you unlock them.
I'd really like to see a good FOSS webmail client so that I can host my own.