I would disagree - intercept technology has advanced to the point that the law no longer covers its use - but that doesn't always make it right to use.
If technology found a way to make each TV set in america a camera - so that they could monitor a room with a TV set at will - would you say the law needed to authorise them to randomly look out of any TV set just so the law can "keep up"?
They can't use it in court - but half the time you don't *need* to use intel in court, just turning up when the arms shipment is coming in is more than enough:)
Actually, generating the numbers is quite easy - use a decent low grade entropy source (I find good results from a ordinary radio tuned to a dead station and samped via a sound card) then hash in a suitable ratio to "concentrate" the entropy into a smaller area (so if you hash 1K blocks down to 128 bits, you have a pretty high-grade random block of 16 bytes - and a 1K/sec sample is quite easy to get. obviously, 650mb will take a fair while at that rate (it works out a little over a K a minute, or a 1.3MB/day - but if you aren't in any real hurry to generate bits, and you can do it continuously for several weeks as a background task.
This is a serious question - has anyone gotten the server to compile under the current Cygwin?
I would *love* to try silc, but the client is a console-only cygwin app after compile, and is only compatable with the silc server (which I can't compile under Cygwin, and am unwilling to trust on one of my solaris boxen)
It is much simpler just to xor a random OTP pool to the plaintext - and doesn't restrict what you can say.
The difficult part of OTP is not the crypto (you can do that on a *watch* these days) but getting the random pad data safely to the recipient before sending the message, and keeping it secure until it needs to be used (after which it should be destroyed of course)
What you are describing is a codebook - and codebooks CAN be broken given enough data.
Q: Why would 99.9999% of internet users ever bother with crypting their emails anyway? The canonical answer is - for the same reason you put a letter in an envelope, not just write on a postcard every time. for a lot of things (particularly love letters and business negotiations) you don't want anyone but the intended recipient to read it.
Q: Why would a Terrorist use software that has a US/UK/UN backdoor, surely they'd write it themselves (hard) or download it from the net (easy)?
They wouldn't use broken software, and it is impossible to force them to without a 100% scan of all email.
Q: Assuming most T's are small organisations surely they'd use replacement words, which unless you've infiltrated the group, you'll never understand.
Or use steganography, yes.
Q: The UK government have been talking about bringing in ID cards in the face of the WTC horrors. Doesn't the US have ID cards already? Every time I wanted a drink in Las Vegas I got 'carded' and I'm 30, so it's not like they don't get checked.
It is sad, but all sorts of control freaks have come out of the woodwork, waving laws that got voted down last time they tried it with "terrorist" scribbled in at a few places to make them look a bit different. ID cards would do nothing against terrorism - it is likely we will never know the real names of the terrorists, given how many seem to have popped up and said "no, I am still alive here" when named.
I know a group of guys who were literally taking all of the porn off of the alt.binaries newsgroups to look for hidden messages, but gave up do to the volune, the low chance of actually getting a hit, etc.
"No honey, I am *not* looking at dirty pictures, I am helping the Fight against Terrorism. see, it says so right here....."
For example, one of the new powers that Ashcroft has called for is that when a surveillance warrant is granted, it be tied to the individual rather than a specific phone, which seems totally reasonable to me.
It *sounds* reasonable, until you try to impliment it - and realise there is no way to wiretap a person, you have to wiretap any device he might *possibly* use.
Taken to extremes, it would justify tapping every phone line at a hotel because he stopped off for a meal there....
I have three (well, a base of three) crypto-capable packages installed right now.
PGP - obvious, the de-facto standard for email encryption, but unless you can handle GPG is expensive closed source payware.
Scramdisk - powerful, OTF encryption with steganographic capabilities, but requires that the host file be created and formatted before use - pretty useless for email, but very good indeed for local storage
S/Mime - built into Netscape, Outlook and Outlook Express for free; lusers can get a free key from www.thawte.com for the effort of going there, and the system is transparent. I generate my own keys using OpenSSL, but the big name packages mentioned above don't like that - it isn't in their hierachical trust structure...
Best application for StegCrypto I know of is Scramdisk - it only supports 16 bit WAV files (for now) but for ease of use it is unbeatable. the lower four bits of each sample are "formatted" to form a virtual disk drive (a bit like a floppy disk).
To open this virtual disk, you drag and drop the wav file on top of the scramdisk app (there are other ways, but that is the simplest) and type in your password. unless you know the password, the volume won't open, and if you examine the file you can't even prove the scramdisk is there (yes, the file's lower four bits will be statistically at random, but this is true of anything but a pure CD rip anyhow - sound cards just can't sample accurately enough to get a clean lower four bits)
Scramdisk is free (with source) from www.scramdisk.clara.net
This has a number of sides
It is pretty awkward to sue a company that has just gone down the pan, owing you a month or more back wages.... and if you wait until the case comes to court, you will almost certainly win - against a company that has already sold all assets and handed over the money to the secured creditors (who aren't going to hand any back)
however, in some cases the assets are never owned by the company that has just gone under - a number of "asset loans" are just that - loans of the assets (for a nominal $1 rent a year and a stack of shares) in which case they aren't stealing from the dot-comm, but from the rental company (and of course, even if the dot com got real money, it may have rented, rather than bought, major assets such as servers.
The odds of it costing only $3 for the entire production run materials doesn't make sense - so neither does your comment.
Only if you consider it to be $3+$10 per copy sold does the comment make sense - and at that I would consider the material cost fairly high. I have seen prices of a few cents per copy for a large "pressed" CD run, and I doubt the box and little booklet would be that expensive.
I've heard at least one report that they did in fact use Yahoo.
I have seen TWO - both of which said they "must have" been using Yahoo or why would the FBI be installing carnivore there.....
You must be joking. Many of the conspirators were right here in the US, where they enjoyed every one of the freedoms that have been debated on this site.
Indeed true - but the communications were almost certainly planned and structured before they were sent to the US. You *don't* send an agent into the field with orders to "see if you can find some way to communicate.
As for the ones outside our borders...they have satphones, laptops, etc I think you will find they don't. Comms and Computers are heavily restricted in that country, in case the people should hear a different message than their religous leaders would wish. While the government has most (if not all) of those, I doubt they would be pressed into service for spy communications.
Feel free to post a few links that show otherwise of course - I don't claim to be infallable.
I would agree with this, but for one point - There is no reason to assume that increasing LEA and "spook" rights to intercept communications, decode private information and break into your machine remotely will in any way increase their ability to locate terrorists.
Think about this - the FBI rushed Carnivore into service at the "freemail" providers like yahoo, when there was no evidence that the terrorists even knew freemail existed - why would they? the internet is banned for the afgan people; the phone service barely exists there, never mind ISPs. If any communications took place outside of the original mission briefing, they were almost certainly by way of "innocent sounding" telephone conversations and/or letters with hidden text. consider the following conversation:
Hi John! have you booked your tickets yet?
Yes, I am flying out of boston at 8am; Hoping to meet up with Clive at the WTC around 9
I am sure you two will make an impression there; I would come too, but I have to attend a meeting at the government place about that time..
Ok, a little contrived - but you see my point. there is *no* way someone, even suspious of one or more of the parties involved, could have guessed at their real plans from that conversation - and they would have to monitor *every* phone call in america, no matter how innocent, to pick it up at all.
Similar statements could be made about almost any of the measures proposed - for each one you should be asking yourself "what will this achieve? will the cost of giving this up be matched by a equal gain in the protection I will get from my government? In this case, the answer is no. it is an attempt to exploit the grief and suffering of the american people to push though "reforms" that the american courts and people have been rejecting for years now. Would you really want the US to be the only country in the world where online banking is insecure, because you have to make sure the police can decode it, and almost any private eye can bribe his way into a couple of juicy keys?
I hate to tell you this - but there is a long history in the UK of judges saying "this evidence was illegally gathered; I expect disiplinary action against the officers concerned, but as your case relies on it I won't throw it out...."
I agree - for home users, even Windows is hard to use at first.
The advantage of a corporate site like this one is that the same helpdesk team that would be supporting the users with MS Office, outlook and solitare will be trained to help the users with StarOffice, sendmail, and XBill:)
Hush 2.0 - first impressions
on
Hushmail 2.0
·
· Score: 3
Ok, I thought I would give this the once over - noting in passing that the uk "cyber-rights.org" version is not yet on the conversion path.
Upgrade
Actual upgrade went smoothly; I was presented with a upgrade applet, which (once given some mouse movements to generate a few numbers) calculated my key for me and then threw me into a conventional-looking login screen. One minor glitch - one of the intermediate screens has a different server key to the main site, so a anti-spoof web browser throws an error.
First Impressions
At first glance, everything looks pretty and shiny; in a similar display to Outlook and many mail clients, Hush now has a folder list down the left edge (with number of unread displayed) and a toolbar along the top with check mail, compose, contacts (address book) and quit.
that the app does not take you to the inbox on start or on "check mail" is a little surprising, but I can live with that. compose works smoothly, and there is a little "options" tab at the top right that will take you to the now integrated "pager" and "change password" functions.
Testing - OpenPGP
well, obviously the big test of the system (and the major change) is pgp. For testing purposes, I did the following:
sent two messages from a conventional mail client, signed with keys long established on the keyservers (youngest is two years old) - one each of RSA and DSS.
results were disappointing - not only did the program fail to verify either key, but no options are available to display signature info (time/date and so forth) to manually specify a key, or to specify a keyserver to check for keys.
Ok, I thought - possibly you have to add the sender to your contacts list. yes? no. a careful search found no way to add the sender of the current message to the contacts. ok, so lets do it manually. copied the address data into the clipboard, then contacts->add. hmm. I can add a contact easily enough (nickname, full name, email, much as always) but no sign or concept of key or default encryption. very odd.
sent mail from Hush to the named recipients Manually composed messages to the recipients, and hit encrypt. no dice... "no key". oh. ok, I will send a unencrypted, but signed message to one of them, and see what it looks like on receipt. oops, hangs indefinitely. this is not good......
send mail hush-->hush a bit shaken, I decided I had better test the functionality that was working before - encrypted mail from a hush user to another hush user. results were the same as to a non-hush encrypted user - encrypt fails with a no-key error, sign hangs the system
I am a bit disappointed in this initial rollout, but am assuming it is still in the teething stage and a few bugs are to be expected (I also have none of my previous mail, but they apologised for that in advance and said it will be moved to the new server soon) however, I think it is unworkable for the short term, and will be waiting to see what happens in the next few weeks. --
Re:Does it work? Well, let's see
on
Hushmail 2.0
·
· Score: 2
yes, I got that error too - it is related to the downloaded (cached) copy of the client. to fix this, go into c:\windows\downloaded program files and delete the hushmail app you will find there.
when you try again, you will get the online (non installed) version which does work. --
NSA has two halves. One half has the purpose of recommending security systems (e.g. DES many moons back).
Ah yes - DES which was deliberately weakened from 128 bits (which was the original recommendation) to 40 (which the NSA could break but hoped nobody else could)
and this supports your argument how? --
I think it would be more appropriate to say they took an OSS product, and modified it to suit what they wanted it to look like - as doing so is one of the strengths of Open Source.
I doubt they actually WANT secure versions of windows out there - several governments seem to be viewing windows with mounting suspicion for official use.... --
Slashdot Mirror
I would disagree - intercept technology has advanced to the point that the law no longer covers its use - but that doesn't always make it right to use.
If technology found a way to make each TV set in america a camera - so that they could monitor a room with a TV set at will - would you say the law needed to authorise them to randomly look out of any TV set just so the law can "keep up"?
They can't use it in court - but half the time you don't *need* to use intel in court, just turning up when the arms shipment is coming in is more than enough :)
Actually, generating the numbers is quite easy - use a decent low grade entropy source (I find good results from a ordinary radio tuned to a dead station and samped via a sound card) then hash in a suitable ratio to "concentrate" the entropy into a smaller area (so if you hash 1K blocks down to 128 bits, you have a pretty high-grade random block of 16 bytes - and a 1K/sec sample is quite easy to get. obviously, 650mb will take a fair while at that rate (it works out a little over a K a minute, or a 1.3MB/day - but if you aren't in any real hurry to generate bits, and you can do it continuously for several weeks as a background task.
This is a serious question - has anyone gotten the server to compile under the current Cygwin?
I would *love* to try silc, but the client is a console-only cygwin app after compile, and is only compatable with the silc server (which I can't compile under Cygwin, and am unwilling to trust on one of my solaris boxen)
well yes, of course.
What do you think the RSA patent (which just expired) was about if it wasn't math?
It is much simpler just to xor a random OTP pool to the plaintext - and doesn't restrict what you can say.
The difficult part of OTP is not the crypto (you can do that on a *watch* these days) but getting the random pad data safely to the recipient before sending the message, and keeping it secure until it needs to be used (after which it should be destroyed of course)
What you are describing is a codebook - and codebooks CAN be broken given enough data.
Ok I give up - how is this offtopic?
The canonical answer is - for the same reason you put a letter in an envelope, not just write on a postcard every time. for a lot of things (particularly love letters and business negotiations) you don't want anyone but the intended recipient to read it.
Q: Why would a Terrorist use software that has a US/UK/UN backdoor, surely they'd write it themselves (hard) or download it from the net (easy)?
They wouldn't use broken software, and it is impossible to force them to without a 100% scan of all email.
Q: Assuming most T's are small organisations surely they'd use replacement words, which unless you've infiltrated the group, you'll never understand.
Or use steganography, yes.
Q: The UK government have been talking about bringing in ID cards in the face of the WTC horrors. Doesn't the US have ID cards already? Every time I wanted a drink in Las Vegas I got 'carded' and I'm 30, so it's not like they don't get checked.
It is sad, but all sorts of control freaks have come out of the woodwork, waving laws that got voted down last time they tried it with "terrorist" scribbled in at a few places to make them look a bit different. ID cards would do nothing against terrorism - it is likely we will never know the real names of the terrorists, given how many seem to have popped up and said "no, I am still alive here" when named.
I know a group of guys who were literally taking all of the porn off of the alt.binaries newsgroups to look for hidden messages, but gave up do to the volune, the low chance of actually getting a hit, etc.
"No honey, I am *not* looking at dirty pictures, I am helping the Fight against Terrorism. see, it says so right here....."
It *sounds* reasonable, until you try to impliment it - and realise there is no way to wiretap a person, you have to wiretap any device he might *possibly* use.
Taken to extremes, it would justify tapping every phone line at a hotel because he stopped off for a meal there....
- PGP - obvious, the de-facto standard for email encryption, but unless you can handle GPG is expensive closed source payware.
- Scramdisk - powerful, OTF encryption with steganographic capabilities, but requires that the host file be created and formatted before use - pretty useless for email, but very good indeed for local storage
- S/Mime - built into Netscape, Outlook and Outlook Express for free; lusers can get a free key from www.thawte.com for the effort of going there, and the system is transparent. I generate my own keys using OpenSSL, but the big name packages mentioned above don't like that - it isn't in their hierachical trust structure...
What do other people here use?I would - and gladly - but I can't get the server to compile under cygwin *sigh*
Come to think of it - if you can do that, just force THEM to use it and leave us alone :)
Not that rare - I have seen it take better than six hours before a submission is even looked at....
Best application for StegCrypto I know of is Scramdisk - it only supports 16 bit WAV files (for now) but for ease of use it is unbeatable. the lower four bits of each sample are "formatted" to form a virtual disk drive (a bit like a floppy disk).
To open this virtual disk, you drag and drop the wav file on top of the scramdisk app (there are other ways, but that is the simplest) and type in your password. unless you know the password, the volume won't open, and if you examine the file you can't even prove the scramdisk is there (yes, the file's lower four bits will be statistically at random, but this is true of anything but a pure CD rip anyhow - sound cards just can't sample accurately enough to get a clean lower four bits) Scramdisk is free (with source) from www.scramdisk.clara.net
This has a number of sides
It is pretty awkward to sue a company that has just gone down the pan, owing you a month or more back wages.... and if you wait until the case comes to court, you will almost certainly win - against a company that has already sold all assets and handed over the money to the secured creditors (who aren't going to hand any back)
however, in some cases the assets are never owned by the company that has just gone under - a number of "asset loans" are just that - loans of the assets (for a nominal $1 rent a year and a stack of shares) in which case they aren't stealing from the dot-comm, but from the rental company (and of course, even if the dot com got real money, it may have rented, rather than bought, major assets such as servers.
The odds of it costing only $3 for the entire production run materials doesn't make sense - so neither does your comment.
Only if you consider it to be $3+$10 per copy sold does the comment make sense - and at that I would consider the material cost fairly high. I have seen prices of a few cents per copy for a large "pressed" CD run, and I doubt the box and little booklet would be that expensive.
I have seen TWO - both of which said they "must have" been using Yahoo or why would the FBI be installing carnivore there.....
You must be joking. Many of the conspirators were right here in the US, where they enjoyed every one of the freedoms that have been debated on this site.
Indeed true - but the communications were almost certainly planned and structured before they were sent to the US. You *don't* send an agent into the field with orders to "see if you can find some way to communicate.
As for the ones outside our borders...they have satphones, laptops, etc
I think you will find they don't. Comms and Computers are heavily restricted in that country, in case the people should hear a different message than their religous leaders would wish. While the government has most (if not all) of those, I doubt they would be pressed into service for spy communications.
Feel free to post a few links that show otherwise of course - I don't claim to be infallable.
Think about this - the FBI rushed Carnivore into service at the "freemail" providers like yahoo, when there was no evidence that the terrorists even knew freemail existed - why would they? the internet is banned for the afgan people; the phone service barely exists there, never mind ISPs. If any communications took place outside of the original mission briefing, they were almost certainly by way of "innocent sounding" telephone conversations and/or letters with hidden text. consider the following conversation:
- Hi John! have you booked your tickets yet?
- Yes, I am flying out of boston at 8am; Hoping to meet up with Clive at the WTC around 9
- I am sure you two will make an impression there; I would come too, but I have to attend a meeting at the government place about that time..
Ok, a little contrived - but you see my point. there is *no* way someone, even suspious of one or more of the parties involved, could have guessed at their real plans from that conversation - and they would have to monitor *every* phone call in america, no matter how innocent, to pick it up at all.Similar statements could be made about almost any of the measures proposed - for each one you should be asking yourself "what will this achieve? will the cost of giving this up be matched by a equal gain in the protection I will get from my government? In this case, the answer is no. it is an attempt to exploit the grief and suffering of the american people to push though "reforms" that the american courts and people have been rejecting for years now. Would you really want the US to be the only country in the world where online banking is insecure, because you have to make sure the police can decode it, and almost any private eye can bribe his way into a couple of juicy keys?
I hate to tell you this - but there is a long history in the UK of judges saying "this evidence was illegally gathered; I expect disiplinary action against the officers concerned, but as your case relies on it I won't throw it out...."
The advantage of a corporate site like this one is that the same helpdesk team that would be supporting the users with MS Office, outlook and solitare will be trained to help the users with StarOffice, sendmail, and XBill :)
- Upgrade
- First Impressions
- Testing - OpenPGP
- sent two messages from a conventional mail client, signed with keys long established on the keyservers (youngest is two years old) - one each of RSA and DSS.
- sent mail from Hush to the named recipients
- send mail hush-->hush
I am a bit disappointed in this initial rollout, but am assuming it is still in the teething stage and a few bugs are to be expected (I also have none of my previous mail, but they apologised for that in advance and said it will be moved to the new server soon) however, I think it is unworkable for the short term, and will be waiting to see what happens in the next few weeks.Actual upgrade went smoothly; I was presented with a upgrade applet, which (once given some mouse movements to generate a few numbers) calculated my key for me and then threw me into a conventional-looking login screen. One minor glitch - one of the intermediate screens has a different server key to the main site, so a anti-spoof web browser throws an error.
At first glance, everything looks pretty and shiny; in a similar display to Outlook and many mail clients, Hush now has a folder list down the left edge (with number of unread displayed) and a toolbar along the top with check mail, compose, contacts (address book) and quit.
that the app does not take you to the inbox on start or on "check mail" is a little surprising, but I can live with that. compose works smoothly, and there is a little "options" tab at the top right that will take you to the now integrated "pager" and "change password" functions.
well, obviously the big test of the system (and the major change) is pgp. For testing purposes, I did the following:
results were disappointing - not only did the program fail to verify either key, but no options are available to display signature info (time/date and so forth) to manually specify a key, or to specify a keyserver to check for keys.
Ok, I thought - possibly you have to add the sender to your contacts list. yes? no. a careful search found no way to add the sender of the current message to the contacts. ok, so lets do it manually. copied the address data into the clipboard, then contacts->add. hmm. I can add a contact easily enough (nickname, full name, email, much as always) but no sign or concept of key or default encryption. very odd.
Manually composed messages to the recipients, and hit encrypt. no dice... "no key". oh. ok, I will send a unencrypted, but signed message to one of them, and see what it looks like on receipt. oops, hangs indefinitely. this is not good......
a bit shaken, I decided I had better test the functionality that was working before - encrypted mail from a hush user to another hush user. results were the same as to a non-hush encrypted user - encrypt fails with a no-key error, sign hangs the system
--
yes, I got that error too - it is related to the downloaded (cached) copy of the client. to fix this, go into c:\windows\downloaded program files and delete the hushmail app you will find there. when you try again, you will get the online (non installed) version which does work.
--
NSA has two halves. One half has the purpose of recommending security systems (e.g. DES many moons back).
Ah yes - DES which was deliberately weakened from 128 bits (which was the original recommendation) to 40 (which the NSA could break but hoped nobody else could)
and this supports your argument how?
--
I think it would be more appropriate to say they took an OSS product, and modified it to suit what they wanted it to look like - as doing so is one of the strengths of Open Source. I doubt they actually WANT secure versions of windows out there - several governments seem to be viewing windows with mounting suspicion for official use....
--