Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×

Comment No, it's not a court document (Score 1) 215

Little do you realize the entire Prenda case is fiction, all written for entertainment purposes (c'mon, didn't you suspect?) and none of these things are really court documents. It's "Internet Porn" not in the sense of porn found on the Internet, but figurative porn, where the Internet is the subject (a la "food porn").

John Steele is just the name of the antagonist character (and in a recursive twist, he happens to be played by a brilliant comedic actor whose real name is "sharkmp4" and I, for one, find the writers stacked reuse of that name, to be hilarious).

And Comcast's letter was all just part of the creative work. The writers had a lot of meetings and shot-down ideas, before thought thought of having the Comcast character send that letter. And now by pirating a whole chapter, Torrentfreak is negatively impacting the resale market of the "The Prenda Case" serial, which had been licensed by Ars Technica, Slashdot, etc in order to provide ad impression fodder.

Comment Don't be a Volunteer Terroree (Score 2) 986

PJ has supplied the analogy already, to explain her mistake. I don't even need to make anything up:

I arrived in New York City, and being naive about the ways of evil doers in big cities, I rented a cheap apartment on the top floor of a six-floor walkup, in the back of the building. That of course, as all seasoned New Yorkers could have told me, meant that a burglar could climb the fire escape or get to the roof by going to the top floor via the stairs inside and then through the door to the roof and climb down to the open window of my apartment.

That is exactly what happened.

In other words, lots of people knew about the risk all along, but someone didn't and got burned as a result. (That sucks. I sympathize.) When that person finally realized how hostile the environment was, and had always been, and what countermeasures that person should have always been using, what happened?

PJ, did you take the burglary as a lesson to move away from NYC and also infer that all rational people should also move away from NYC?

Or did you learn to secure your windows, see fire escapes as possible attack surfaces, etc?

What the government is doing is sucks, and every person in Congress and the White House ought to be issuing press releases that they are going to make the gag orders illegal (so that violating the illegal gags has no negative legal consequence to the speaker). They should say they're going to put an end to the US government working against the interest of US citizens. And every American who would even consider voting for someone who isn't issuing such a press release, ought to be ashamed. And yet, ALL THIS IS A SEPERATE ISSUE. When it comes to privacy, your government is always a potential problem, but it's never your only problem. So no matter what happens to this country, your problem remains and even if you had a big enough gun, you can't solve the privacy problem by pointing your gun at your government's face.

Just like how moving away from NYC, doesn't mean you never have to worry about another burglary.

In 1991, PRZ released the first version of PGP. Why do you think he did that? We have known literally for decades that lots of people are able to read our unencrypted email. I just don't understand how this basic and obvious fact is still wished away. Nothing Bushbama did, changed that. Nothing the NSA has done, changed that. The PATRIOT Act didn't change it, CALEA didn't change it, and future CALEA expansion isn't going to change it. On September 10th 2001, you could have just as easily and accurately written about unencrypted email,

I can't continue. There is now no shield from forced exposure.

and it would be no less relevant or true than it is today. And similarly, when I see

You don't expect a stranger to read your private communications to a friend.

I have to call bullshit. You "don't expect" it, in the sense that you think it's undesireable that it happens, and you wish your plaintext communication could be private. But seriously, for decades you certainly have "expected" it in the sense that you predict it and have had reason to think it could happen, committed by any of many parties including the government, and that it can be done passively and inexpensively, without anyone ever detecting it and being held accountable. Your love letters were always on the wind.

The environment didn't change, PJ. You did. The world is no darker now than it was when you launched groklaw. If anything, things are better. You went from not giving a fuck if people read your email, or maybe from living in denial of what every single Internet tech-minded person eventually realizes, to understanding how vulnerable unsecured communications are, and caring about it.

That's good! It's progress, a first step. But after that, you've just now stumbled horribly, staggering into depressed crazyland, throwing up your hands instead of doing something about it (like, say, making sure everyone you meet has your key's fingerprint). And that (giving up) doesn't sound to me, like someone who really cares. I hope I'm wrong; I hope the real problem is that you simply don't know that the ball is in your court. That seems weird to me, but I guess lots of people still don't know about PGP. I hope history calls 2013 not the year of spying, but the year spying got a lot harder because the public finally did what they had long needed to do.

If you have to stay on the Internet, my research indicates that the short term safety from surveillance, to the degree that is even possible, is to use a service like Kolab for email, which is located in Switzerland, and hence is under different laws than the US, laws which attempt to afford more privacy to citizens.

*sigh* To me, this is so 2002 thinking. Jurisdiction shopping was one of my first gut reactions too, but as a stratagy it did not survive exposure to real life facts. Go ahead and continue your "research" and I think you'll be apologetically taking back your Swiss recommendation soon enough.

Jurisdiction shopping doesn't make users secure their communications. Giving a fuck is what makes users secure their communications. And users (not governments, not services) are the only hope, ever; people who care can have security, and people who don't care, cannot. Your unlocked-door-on-the-Internet in Switzerland is just as open and waiting for whoever wants what is within, as your unlocked-door-on-the-Internet in USA. Why don't you try locking the door instead? Then you can have it anywhere.

You're so close. Now back up just one step, reverting the depressed crazy talk, and start acting like someone who gives a fuck. Start building short links on the WoT, spread the word. For decades we have seen the problem, but also, for decades We have had the power to do something about it. Join up.

Comment Re:Party!!! (Score 1) 399

Crud, I wish I had seen this earlier... I really hope no PGP newbies read what you said without thinking and then believed it.

Having a signed key is neither necessary of significant when sending encrypted mail using a key provided to you
by your recipient. He could be a completely bogus person trying to 419 you into sending a credit card.
If so, his key will be signed. And the signatures will look impressive. But they too will be bogus.
And no part of attending signing parties will prevent this.

Exactly wrong. The way that keysigning parties prevent the problem, is that they cause signatures to become meaningful! The one and only way that signatures ever gain the capacity to "look impressive" is when you have a trust path through them. A 419 scammer's fake id isn't ever going to be signed by people that you know, where as a real person might, thanks to keysigning parties.

Comment How this really does help (a little) (Score 1) 217

If you strain to look at things in the best possible light, you will figure out there are some scenarios where this helps. And if you take a pessimistic view, I think the conclusion is that this is completely harmless. Unfortunately, it's also very dishonest, so Google earns a demerit anyway, but that's another topic that plenty of people are already going on about. ;-)

Obviously this doesn't protect the data if Google is coerced into giving up the key, or if Marketing decides there might be profiling advantages to be gained in examining the plaintext.

But it does help against certain types of inadvertent leaks or subterfuge. For example, my server's disks are encrypted with a key that is easily available on their boot SSD. If you steal my server you have my data. If you sneak into my server, you have my data. The encryption is starting to sound useless, isn't it?

But if you take a disk and leave the rest of the machine behind, you don't have anything. If I get too many Offline Uncorrectable SMART errors and send you my disk for a warranty replacement, you (nor anyone who intercepts the delivery) don't have my data. The encryption isn't useless; it's just mostly useless. Except that it's useful in what just happens to be the most common scenario, something that happens 3-4 times per year as various disks rotate through UPS' fine delivery service.

If Google is doing something like that, cool. And if they're using iSCSI or something where disk blocks are moving through their own internal network where the attack surface is even larger, and now a sneaky tap on their storage network will start seeing ciphertext instead of plaintext, I say: good!.

It's a bit slimy that Google is announcing this common-sense minor edge-case precaution right now, when the public is thinking about totally different threat models thanks to this years' news stories. And the announcement itself is completely full of bullshit. But nevertheless, look carefully and you really will see something with just a little bit of positive value.

I'll say what I've said before: It's good to fear and act against Big Brother, but the thousand Little Brothers out there are attacking you much more often and overall probably causing us all more long-term average loss. Deal with them, and you'll also be incrementally dealing with Big Brother too, by changing how we think about info security.

Comment Re:Party!!! (Score 1) 399

[keysigning party] Not helpful in obtaining a key with which to send email.

That's because you looked at the answer from the sender's PoV and/or in the short-term. First rule of interviews: don't answer the quest you're asked; answer the question you wish you were asked. ;-)

Flip the submitter's question: "I need a client to send me personal info, but he doesn't know my key. Hey Slashdot, how can he get my key?" If you're an organization that is for whatever reason making a habit of requesting personal information from various people, then your org's people ought to be going to keysigning parties.

Or distort the timeliness: "Ten years from now, I'll need to send personal information to someone. How do I bring about an environment where I'm likely to be able to easily get their key and believe it?"

Comment Re: This is why encryption isn't popular (Score 1) 399

You wouldn't use it for any private communications. The best use for something like this would be to use your government-issued key to sign your "real" personal key. Such a system would allow anyone to trust your key's identity by as much as they trust the government. As a lower bound.

Think about what those of us without these government-issued crypto systems are doing. We meet strangers and check their government-issued ids (drivers' license, passport) and then say "ok, I think you're really Joe Schmoe" and sign Joe's key. There's an implicit government-trust hop in each one of these, except that a third party observer can't tell whether or not it's there. If I see you signed Joe's key, I don't know if that's because you actually know Joe, or if it's because y'all set up a keysigning meeting and then signed each other based on trusting each other's government ids.

With cryptographic government issued ids, we could stop having these kinds of keysigning meetings, and raise the standard for keysigning. I would no longer sign you simply because you have a government document saying you are who you are, because you would be able to do that yourself. You would have to verify your identity to me, some other way. This would raise the overall reliability of signatures.

I totally want the government to get in on keysigning. Just don't fucking use single-signer systems. such as X.509. The government's attestations should be additive not exclusive. We ruined HTTPS; let's learn from that mistake and do "email2" right.

Comment Use the key which signed the request (Score 1) 399

An organization wants me to send them my personal data by email

Whatever key was used to sign that request, is the key to use. Since you've already verified the request, you must already have the key and have verified its identity.

*pause*

Whaddya mean, "the request wasn't signed?" Hmmm.. Are you sure you know who is asking?

Comment Crypto is the answer, jurisdiction-shopping isn't (Score 2) 470

What the government is doing is repugnant, but only because most people are stupid and take the wrong lessons from it. If people had their shit together, then it would actually cause a positive effect, and we'd be talking about how US government's thuggery inadvertently did everyone a favor.

I never even heard of these encrypted email services until yesterday (except for hushmail about a decade ago but that was an even dumber beast) and the more I look into them, the more apparent it is that they sell .. well .. "snakeoil" is maybe too harsh, but I guess I'd have to say they sell the service of closing barn doors after horses escape. If I had to put it really nicely, to the point of sickening insincere sweetness, I suppose I could say they help you deploy "defense in depth" and I might be able to avoid making any gagging sounds as I did it.

Either the sender encrypts your email with your key, or they don't.

If they do it (i.e. if people do things right), then you don't need any service's special help with anything. All you want from your service are reliability, performance, and low prices -- a commodity, just like ISP's service of packet-passing.

If the sender doesn't encrypt the email with your key, then you're fucked. This is the common scenario, and the fact that people are basically fucked but still want to somehow mitigate it, is how this market emerged. Fair enough, I get it: when life hands you lemons, you make lemonaide. But you're taking it way too seriously, expecting far too much from a lossy premise. Your lemonaide is never going to be Dogfish Head 90 Minute IPA, ever, period. You should lament that, that people don't encrypt. You don't know who all read your PLAINTEXT before it got to Silent Circle or Lavabit and then they encrypted the storage of it.

(Worse, from what people are hinting about how lavabit worked, it sounds like they did the storage wrong, and that everyone always knew they would be able to decrypt things under certain circumstances, if forced.)

Users and their endpoint software must provide security. Other people's media and services running on other people's computers, can't really help you. Everything in between the endpoints is untrusted. Gag orders, CALEA-like laws, etc will make even the best-meaning services untrustworthy.

So. If it makes users feel better to move their hosting to other jurisdictions, fine. But for fuck's sake, go beyond just trying to make yourself feel better, and actually do something to make things really better: have a keysigning party. Help webmail users find and upgrade to decent (i.e. openpgp-compatible) mailreaders. And so on. Every time you see an unencrypted email come in, think about WTF went wrong and how that could have been prevented. And if you really do this, then you'll find that you can still host in America.

BTW, we've been through all this before. It's not like anything truly new is happening. All the same issues were coming up ten years ago, and ten years before that. (And probably ten years before that but I missed out on that round.) It always comes down to jurisdiction-shopping being a waste of time. You have the ultimate weapon which makes it all obsolete: 1970s PK tech. The only time you need jurisdiction-shopping is if your government outlaws the tech (France still? Not sure.).

Comment Re:What _is_ an "encrypted email provider"? (Score 1) 771

Thank you for your contribution, AC. I have known you for many years and always felt your reputation was unfairly earned and the people who accused you of crapflooding, over-specificity, and blabbering endlessly, were just projecting. I know I was, when I flamed you for it. Or maybe I am looking at history through rose-tinted lenses, and you were actually guilty back then, but I would hate to have the indiscretions of my youth thrown in my face. My, how the years have changed us, AC. So, it's this. This is what it's like to be old. I have changed, and so have you.

Over-specificity. As if. If only they could see you now!

Comment Re:Excellent Idea (Score 1) 321

That provides end to end and is vulnerable to MITM attacks

Being vulnerable to MitM attacks is ok, provided you do two things:

1) You tell the user. (Probably not with some kind of scary modal thing, but the population does need to somehow get educated about what it means.) This is something that current web browsers do wrong, when they treat untrusted certs as being worse than plaintext.

2) You have the option, for users who are willing to go to some extra trouble for key exchange, to take countermeasures against MitM. This is what makes Diffie-Helman alone be insufficient. There should always be a MitM-proofable wrapper around it, even if by default for novice users, it isn't MitM-proof. Then whenever hubby reads one too many news stories and decides he cares a little bit, he can say, "ok, I'll have my wife read her key signature to me" and then by magic he's actually done something useful and the situation really gets better, and the novice has easily and incrementally become a beginner.

Slashdot Top Deals

It is easier to write an incorrect program than understand a correct one.

Working...