I speculate that I'd let the TLS ports through by default:
SSL/TLS encrypted IMAP uses port 993
encrypted POP uses port 995
encrypted SMTP uses port 465
and open the ordinary ports if someone said the equivalent of "I'm using STARTTLS on port 25, could I have it open?"
This gives a starting point for people who wish some degree of security: it's only over the 'net, and it can be subverted by either NSA-forged certificates (:-)) or server attacks, but it allows for a migration to stronger key/certificate scenarios.
The German security service tested an older but still good version from Siemens (my employer) years ago, and stopped as soon as they discovered that the "birthday paradox" made it totally unsuitable for large-scale use.
If you scan for one particular person out of thousands in an airline terminal, you get a certain small number of false positives, so it sorta works for that case. If, however, you search for the entire Baader-Meinhoff gang and all the other terrorists of the day in the same terminal, you get an insane number of false positives, because you're doing (N*M)! comparisons, each with a small chance of a false positive.
This is the same thing that causes the "birthday paradox", where you get a 50% probability of two people at a party having the same birthday when you have only 23 people present. One would normally expect it to take 367 people, but you're actually comparing (23 * 22) people, not (1 * 23)...
There has been some good work done with the technology, and the Ontario Privacy Commissioner has successfully used it to identify small numbers of self-selected problem gamblers at casinos, but until the technology literally becomes perfect, it will fail by creating false positives for any N * M problem where both N and M are large.
The German BND took one look at the in-the-large problem and said "No thanks, that will have us arresting my grandma as a Baader-Meinhoff member, and she'd never forgive me".
To be fair, I'd love if people started doing lots more TLS for email, and more end-to-end, but in that case I was addressing what %$&^!!! idiots ISPs do now (:-))
Even the lower houses of parliament have such a high cost to get nominated that they select only for the seriously well-to-do. Such folks tend toward static and statist systems, to preserve their wealth.
It's probably fair to call the U.S. Congress the "Houe of Lords", and take their statements with just a tiny grain of salt.
A good packet-inspection gateway between an ISP' and the internet can catch both incoming and outgoing spam in unencrypted mail, and in the latter case warn the sender that they've be turned into a 'bot.
Customers don't like being told they're sending spam, but are mildly supportive if you tell them they've been attacked by a virus, should run XXX to fix the problem, and won't be bothered again about sending spam for a week, unless of course they get a new and different virus (;-))
As another commentator pointed out, she may have been found to be part of a chain of connections to someone the NSA is actively interested in, such as the spouse of a reporter, perhaps. That makes her, will-nilly, a subject of investigation.
The experience of being dragged into even a public, legal dispute is not a pleasant one. For example, being a witness in court against an RSM* is both stressful and a career-ending move. I could well imagine she would not wish her friends and correspondents dragged into a secret court...
As one of the early adopters (on mainframes, no less) I find SMTP more subverted than insecure.
We used to look up the A (now MX) record of the recipient and make a direct connection to their mail-server, which allows end-to-end encryption and hides all the "envelope" information.
These days my ISP prohibits that, and inserts themselves as a man-in-the-middle, and do store-and-forward on a protocol that was written for end-to-end. They says it's for spam, but since they only filter on ingress, their justification of egress controls ring false.
It doesn't have to be that way: our first major ISP in Canada (uuunet.ca) assumed everyone depended on normal SMTP behaviour, and refrained from filtering it out!
She's a lovely source of starting-points, from which to do one's connection matrix. As I noted elsewhere in this discussion, a copyright maximalist or a NSA fanboy would see her as an attractive target.
If someone is reading all our (insecure) emails to and from a known "person of interest", such as, for example, a well-known web site, then they can build the kind of interconnection matrix that will lead them to the supporters and fellow-travellers of that website.
Were I a copyright maximalist, I would regard groklaw as a criminal conspiracy, and the centre of a matrix of criminals and fellow-travellers. Based on that, I'd then petition the communications security establishment for a (secret) order allowing me to identify the conspirators and their fellow-travellers for (equally secret) investigation, leading to either prosecution or private revenge...
No, just tunnel your email out of your ISP to where the MX record of your recipient says and use PGP.
Methinks Silent Circle will eventually offer that (note who is a founder, and what his relationship to PGP was)
Phil Zimmerman, one of the Silent Circle founders, has a history of fighting back publicly. See also https://en.wikipedia.org/wiki/Phil_Zimmermann
I therefor suspect he is telling the truth.
In addition, orders to not disclose the existence of orders have been found unenforcable in common-law jurisdictions.
--dave
Re:They keep trying and failing
on
How DRM Won
·
· Score: 1
The vendors sell them as the same thing: protection against a customer taking a copy of the thing they bought.
Re:They keep trying and failing
on
How DRM Won
·
· Score: 1
The price of a DRM package for a 286 DOS application was approximately the size of our profit margin, and would have pushed the total cost into a whole different bracket (:-))
They keep trying and failing
on
How DRM Won
·
· Score: 4, Interesting
The first DRM I saw was funny formats on Apple ][ floppies, followed by DOS format misfeatures, followed by dongles, followed by own-code in apps, followed by... ite ad infinitum.
Note that you don't see these forms of DRM any more. What you do see is that, each time a new format of anything comes out, some DRM vendor talks the publishers into "protecting" their work[1].
As long as new publishers are suckers, the DRM vendors will suck them in, and make lots of money off a technology that motivates people to not buy the publications.
The publishers lose two ways!
--dave
[1. One of my former employers almost got taken in by this scam, but the techies caught it. ]
Not too long ago The Economist noted the lack of new graduates in India to take up the development jobs the outsourcing companies had on offer. Comments from an individual outsourcer seemed to support that...
I'd take this one with a mine of salt, and speculate that by "developer" they mean "someone who wants to be a developer", without consideration of whether they have experience or training.
Yes: xml can be edited, read by humans and repaired when wrong. It is typically parsed via commonly available packages, reducing human effort writing and computer effort running ad-hoc parsers for home-made formats. In principle it can be parsed quickly, although in practice that does vary a lot.
It's a win in spreadsheets and anything with sparse matrices.
It's a wash in text, as you'd imagine,
It's a lose in databases, which are almost always data-intensive,
It's horrid if you are encoding bit patterns in something like base-50.
We did a number of sanity-tests before using it in Ability, and later did a check using customer data that had been given to us, from which comes the characterization above.
Slashdot Mirror
He's in Quebec, SLAPP suits are actionable there.
Cool!
I speculate that I'd let the TLS ports through by default:
and open the ordinary ports if someone said the equivalent of "I'm using STARTTLS on port 25, could I have it open?"
This gives a starting point for people who wish some degree of security: it's only over the 'net, and it can be subverted by either NSA-forged certificates (:-)) or server attacks, but it allows for a migration to stronger key/certificate scenarios.
--dave
The German security service tested an older but still good version from Siemens (my employer) years ago, and stopped as soon as they discovered that the "birthday paradox" made it totally unsuitable for large-scale use.
If you scan for one particular person out of thousands in an airline terminal, you get a certain small number of false positives, so it sorta works for that case. If, however, you search for the entire Baader-Meinhoff gang and all the other terrorists of the day in the same terminal, you get an insane number of false positives, because you're doing (N*M)! comparisons, each with a small chance of a false positive.
This is the same thing that causes the "birthday paradox", where you get a 50% probability of two people at a party having the same birthday when you have only 23 people present. One would normally expect it to take 367 people, but you're actually comparing (23 * 22) people, not (1 * 23)...
There has been some good work done with the technology, and the Ontario Privacy Commissioner has successfully used it to identify small numbers of self-selected problem gamblers at casinos, but until the technology literally becomes perfect, it will fail by creating false positives for any N * M problem where both N and M are large.
The German BND took one look at the in-the-large problem and said "No thanks, that will have us arresting my grandma as a Baader-Meinhoff member, and she'd never forgive me".
--dave
To be fair, I'd love if people started doing lots more TLS for email, and more end-to-end, but in that case I was addressing what %$&^!!! idiots ISPs do now (:-))
This is what Plato wrote about, and the Romans invented a cool variant on: "bread and circuses". The latter led to / supported tyranny, the Empire.
Plato observed a cycle in Greek city-states: aristocracy to timocracy, to an oligarchy, to a democracy, to tyranny and thence back to aristocracy.
Check out "responsibility to protect", which is the hook less-insane countries use to hang a duty to not kill your own citizens on (;-))
Even the lower houses of parliament have such a high cost to get nominated that they select only for the seriously well-to-do. Such folks tend toward static and statist systems, to preserve their wealth.
It's probably fair to call the U.S. Congress the "Houe of Lords", and take their statements with just a tiny grain of salt.
--dave
A good packet-inspection gateway between an ISP' and the internet can catch both incoming and outgoing spam in unencrypted mail, and in the latter case warn the sender that they've be turned into a 'bot.
Customers don't like being told they're sending spam, but are mildly supportive if you tell them they've been attacked by a virus, should run XXX to fix the problem, and won't be bothered again about sending spam for a week, unless of course they get a new and different virus (;-))
"Solved problem in computer science"
--dave
Citation, please?
As another commentator pointed out, she may have been found to be part of a chain of connections to someone the NSA is actively interested in, such as the spouse of a reporter, perhaps. That makes her, will-nilly, a subject of investigation.
The experience of being dragged into even a public, legal dispute is not a pleasant one. For example, being a witness in court against an RSM* is both stressful and a career-ending move. I could well imagine she would not wish her friends and correspondents dragged into a secret court...
--dave
* Regimental Sergeant-Major
People are amazingly bad at understanding risks: In the U.S., for example... http://www.cato.org/blog/youre-eight-times-more-likely-be-killed-police-officer-terrorist
As one of the early adopters (on mainframes, no less) I find SMTP more subverted than insecure.
We used to look up the A (now MX) record of the recipient and make a direct connection to their mail-server, which allows end-to-end encryption and hides all the "envelope" information.
These days my ISP prohibits that, and inserts themselves as a man-in-the-middle, and do store-and-forward on a protocol that was written for end-to-end. They says it's for spam, but since they only filter on ingress, their justification of egress controls ring false.
It doesn't have to be that way: our first major ISP in Canada (uuunet.ca) assumed everyone depended on normal SMTP behaviour, and refrained from filtering it out!
She's a lovely source of starting-points, from which to do one's connection matrix. As I noted elsewhere in this discussion, a copyright maximalist or a NSA fanboy would see her as an attractive target.
J. Edgar Hoover, perhaps?
The hard part of finding people to prosecute is *finding* candidates. Once you know who one person is, you can do traffic analysis and find all their friends. See, for example http://kieranhealy.org/blog/archives/2013/06/09/using-metadata-to-find-paul-revere/
If someone is reading all our (insecure) emails to and from a known "person of interest", such as, for example, a well-known web site, then they can build the kind of interconnection matrix that will lead them to the supporters and fellow-travellers of that website.
Were I a copyright maximalist, I would regard groklaw as a criminal conspiracy, and the centre of a matrix of criminals and fellow-travellers. Based on that, I'd then petition the communications security establishment for a (secret) order allowing me to identify the conspirators and their fellow-travellers for (equally secret) investigation, leading to either prosecution or private revenge...
--dave
Indeed: it rather reminds me of usenet, where the community created a computer network over the objections of a telephone polyopoly.
I speculate we will see two out of perhaps three or four initiatives suceed in the next little while:
--dave
No, just tunnel your email out of your ISP to where the MX record of your recipient says and use PGP. Methinks Silent Circle will eventually offer that (note who is a founder, and what his relationship to PGP was)
They did, in China. Alas, home was the U.S.
Phil Zimmerman, one of the Silent Circle founders, has a history of fighting back publicly. See also https://en.wikipedia.org/wiki/Phil_Zimmermann I therefor suspect he is telling the truth.
In addition, orders to not disclose the existence of orders have been found unenforcable in common-law jurisdictions.
--dave
The vendors sell them as the same thing: protection against a customer taking a copy of the thing they bought.
The price of a DRM package for a 286 DOS application was approximately the size of our profit margin, and would have pushed the total cost into a whole different bracket (:-))
The first DRM I saw was funny formats on Apple ][ floppies, followed by DOS format misfeatures, followed by dongles, followed by own-code in apps, followed by ... ite ad infinitum.
Note that you don't see these forms of DRM any more. What you do see is that, each time a new format of anything comes out, some DRM vendor talks the publishers into "protecting" their work[1].
As long as new publishers are suckers, the DRM vendors will suck them in, and make lots of money off a technology that motivates people to not buy the publications.
The publishers lose two ways!
--dave
[1. One of my former employers almost got taken in by this scam, but the techies caught it. ]
Not too long ago The Economist noted the lack of new graduates in India to take up the development jobs the outsourcing companies had on offer. Comments from an individual outsourcer seemed to support that...
I'd take this one with a mine of salt, and speculate that by "developer" they mean "someone who wants to be a developer", without consideration of whether they have experience or training.
--dave
Yes: xml can be edited, read by humans and repaired when wrong. It is typically parsed via commonly available packages, reducing human effort writing and computer effort running ad-hoc parsers for home-made formats. In principle it can be parsed quickly, although in practice that does vary a lot.
--dave
It's a win in spreadsheets and anything with sparse matrices.
It's a wash in text, as you'd imagine,
It's a lose in databases, which are almost always data-intensive,
It's horrid if you are encoding bit patterns in something like base-50.
We did a number of sanity-tests before using it in Ability, and later did a check using customer data that had been given to us, from which comes the characterization above.
--dave