A codebook in classical crypto is a mapping from a chosen number of
normal or nonsense words to the corresponding intended words or even
phrases and sentences. Wouldn't a generalization of that idea, namely a mapping from certain types of sentences to the corresponding intended
phrases or sentences be much more versatile and useful in practice for purposes of steganographical transmission of secret messages? I mean
e.g. a rather mundane sentence expressing the fact that some friends
came to visit us and we have together done something and that they'll
further travel to somewhere could be written in a quite large number of different ways textually via varying the names of the visitors and what
we have jointly done etc. as well as via varying the grammatical constructions of the sentence.
M. K. Shen
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of
normal or nonsense words to the corresponding intended words or even
phrases and sentences. Wouldn't a generalization of that idea, namely a
mapping from certain types of sentences to the corresponding intended
phrases or sentences be much more versatile and useful in practice for
purposes of steganographical transmission of secret messages? I mean
e.g. a rather mundane sentence expressing the fact that some friends
came to visit us and we have together done something and that they'll
further travel to somewhere could be written in a quite large number of
different ways textually via varying the names of the visitors and what
we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt
10^9 bytes, it is hopeless. Languages are pretty well structured, and
meaning is also well structured. It is really hard to tell a computer
how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
Am 08.04.2016 um 03:30 schrieb William Unruh:
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of
normal or nonsense words to the corresponding intended words or even
phrases and sentences. Wouldn't a generalization of that idea, namely a
mapping from certain types of sentences to the corresponding intended
phrases or sentences be much more versatile and useful in practice for
purposes of steganographical transmission of secret messages? I mean
e.g. a rather mundane sentence expressing the fact that some friends
came to visit us and we have together done something and that they'll
further travel to somewhere could be written in a quite large number of
different ways textually via varying the names of the visitors and what
we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt
10^9 bytes, it is hopeless. Languages are pretty well structured, and
meaning is also well structured. It is really hard to tell a computer
how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
I must say that I personally never have dreamed of ever having a need
to send a message of 10^9 bytes. What I meant is that one could employ
the idea I sketched to sufficiently well send some short pieces of
secret messages steganographically. In general it is obvious that
really high secret messages are invariably short.
Your last but one sentence is a misunderstanding of my post. I didn't
mean that an intelligent piece of software should do or help to do
the steganographical work. The communication partners have to construct
the rules/conventions to write the sentences to convey the informations
that are hidden in them.
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 08.04.2016 um 03:30 schrieb William Unruh:
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of
normal or nonsense words to the corresponding intended words or even
phrases and sentences. Wouldn't a generalization of that idea, namely a >>>> mapping from certain types of sentences to the corresponding intended
phrases or sentences be much more versatile and useful in practice for >>>> purposes of steganographical transmission of secret messages? I mean
e.g. a rather mundane sentence expressing the fact that some friends
came to visit us and we have together done something and that they'll
further travel to somewhere could be written in a quite large number of >>>> different ways textually via varying the names of the visitors and what >>>> we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt
10^9 bytes, it is hopeless. Languages are pretty well structured, and
meaning is also well structured. It is really hard to tell a computer
how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
I must say that I personally never have dreamed of ever having a need
to send a message of 10^9 bytes. What I meant is that one could employ
the idea I sketched to sufficiently well send some short pieces of
secret messages steganographically. In general it is obvious that
really high secret messages are invariably short.
Not obvious at all. The Panama papers were almost certainly sent
encryptoed, and were hardly short. The Snowdon papers the same.
If you want to do it, go ahead, but I thought that you were asking a
general question not one which just referred to you.
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 08.04.2016 um 03:30 schrieb William Unruh:
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of >>>>> normal or nonsense words to the corresponding intended words or even >>>>> phrases and sentences. Wouldn't a generalization of that idea, namely a >>>>> mapping from certain types of sentences to the corresponding intended >>>>> phrases or sentences be much more versatile and useful in practice for >>>>> purposes of steganographical transmission of secret messages? I mean >>>>> e.g. a rather mundane sentence expressing the fact that some friends >>>>> came to visit us and we have together done something and that they'll >>>>> further travel to somewhere could be written in a quite large number of >>>>> different ways textually via varying the names of the visitors and what >>>>> we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt >>>> 10^9 bytes, it is hopeless. Languages are pretty well structured, and >>>> meaning is also well structured. It is really hard to tell a computer >>>> how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
I must say that I personally never have dreamed of ever having a need
to send a message of 10^9 bytes. What I meant is that one could employ >>> the idea I sketched to sufficiently well send some short pieces of
secret messages steganographically. In general it is obvious that
really high secret messages are invariably short.
Not obvious at all. The Panama papers were almost certainly sent
encryptoed, and were hardly short. The Snowdon papers the same.
If you want to do it, go ahead, but I thought that you were asking a
general question not one which just referred to you.
Well, one could certainly find exceptions or at least reasonalble exceptions. I tend to think of situations like the case where a representative of a firm in a negotiation with a customer (under competition of rival firms) needs to question his chief and get instructions from him. I suppose the diverse possible essential
questions and answers could be somehow foreseen and so managed to
be hidden in innocent messages to prevent the potential risks of
espionage.
It is called pgp. You just encrypt your message. No need to anticipate
the questions beforehand. It can ge used for any possible message you
would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you
are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really
creative in the way you manufacture the message. It is also horrendously "expensive" (transmitted text much much much larger than the message to
be sent).
Note that this type of crypto is ancient. It is also pretty limited in
its usefulness. Modern crypto means that you do not have to worry about
your message being readable. No Stego needed.
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 08.04.2016 um 03:30 schrieb William Unruh:
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of
normal or nonsense words to the corresponding intended words or even >>>>> phrases and sentences. Wouldn't a generalization of that idea, namely a >>>>> mapping from certain types of sentences to the corresponding intended >>>>> phrases or sentences be much more versatile and useful in practice for >>>>> purposes of steganographical transmission of secret messages? I mean >>>>> e.g. a rather mundane sentence expressing the fact that some friends >>>>> came to visit us and we have together done something and that they'll >>>>> further travel to somewhere could be written in a quite large number of >>>>> different ways textually via varying the names of the visitors and what >>>>> we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt
10^9 bytes, it is hopeless. Languages are pretty well structured, and
meaning is also well structured. It is really hard to tell a computer
how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
I must say that I personally never have dreamed of ever having a need
to send a message of 10^9 bytes. What I meant is that one could employ
the idea I sketched to sufficiently well send some short pieces of
secret messages steganographically. In general it is obvious that
really high secret messages are invariably short.
Not obvious at all. The Panama papers were almost certainly sent
encryptoed, and were hardly short. The Snowdon papers the same.
If you want to do it, go ahead, but I thought that you were asking a
general question not one which just referred to you.
Well, one could certainly find exceptions or at least reasonalble
exceptions. I tend to think of situations like the case where a representative of a firm in a negotiation with a customer (under
competition of rival firms) needs to question his chief and get
instructions from him. I suppose the diverse possible essential
questions and answers could be somehow foreseen and so managed to
be hidden in innocent messages to prevent the potential risks of
espionage.
M. K. Shen
Am 09.04.2016 um 22:21 schrieb William Unruh:...
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
It is called pgp. You just encrypt your message. No need to anticipate
the questions beforehand. It can ge used for any possible message you
would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you
are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really
creative in the way you manufacture the message. It is also horrendously
"expensive" (transmitted text much much much larger than the message to
be sent).
Note that this type of crypto is ancient. It is also pretty limited in
its usefulness. Modern crypto means that you do not have to worry about
your message being readable. No Stego needed.
I don't understand why you want to unconditionally neglect certain
feasible alternatives which under suitable conditions could be useful. Analogy: If you have a motorcycle, you can from your home extremely
fast reach any arbitrary point of your city. But in some cases riding
a bicycle may be just as fine and perhaps also more enjoyable, though
the radius of reachability is for practical reasons comparatively
limited and it takes more time to get to your destinations.
M. K. Shen
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 08.04.2016 um 03:30 schrieb William Unruh:
On 2016-04-07, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
A codebook in classical crypto is a mapping from a chosen number of >>>>>> normal or nonsense words to the corresponding intended words or even >>>>>> phrases and sentences. Wouldn't a generalization of that idea, namely a >>>>>> mapping from certain types of sentences to the corresponding intended >>>>>> phrases or sentences be much more versatile and useful in practice for >>>>>> purposes of steganographical transmission of secret messages? I mean >>>>>> e.g. a rather mundane sentence expressing the fact that some friends >>>>>> came to visit us and we have together done something and that they'll >>>>>> further travel to somewhere could be written in a quite large number of >>>>>> different ways textually via varying the names of the visitors and what >>>>>> we have jointly done etc. as well as via varying the grammatical
constructions of the sentence.
If you have 5 words you want to transmit, OK. If you want to encrypt >>>>> 10^9 bytes, it is hopeless. Languages are pretty well structured, and >>>>> meaning is also well structured. It is really hard to tell a computer >>>>> how to manufacture plausible and meaningful text.
So no, it would not be more versatile and useful.
I must say that I personally never have dreamed of ever having a need
to send a message of 10^9 bytes. What I meant is that one could employ >>>> the idea I sketched to sufficiently well send some short pieces of
secret messages steganographically. In general it is obvious that
really high secret messages are invariably short.
Not obvious at all. The Panama papers were almost certainly sent
encryptoed, and were hardly short. The Snowdon papers the same.
If you want to do it, go ahead, but I thought that you were asking a
general question not one which just referred to you.
Well, one could certainly find exceptions or at least reasonalble
exceptions. I tend to think of situations like the case where a
representative of a firm in a negotiation with a customer (under
competition of rival firms) needs to question his chief and get
instructions from him. I suppose the diverse possible essential
questions and answers could be somehow foreseen and so managed to
be hidden in innocent messages to prevent the potential risks of
espionage.
It is called pgp. You just encrypt your message. No need to anticipate
the questions beforehand. It can ge used for any possible message you
would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you
are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really
creative in the way you manufacture the message. It is also horrendously "expensive" (transmitted text much much much larger than the message to
be sent).
Note that this type of crypto is ancient. It is also pretty limited in
its usefulness. Modern crypto means that you do not have to worry about
your message being readable. No Stego needed.
On 2016-04-10, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 22:21 schrieb William Unruh:...
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
It is called pgp. You just encrypt your message. No need to anticipate
the questions beforehand. It can ge used for any possible message you
would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you
are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really
creative in the way you manufacture the message. It is also horrendously >>> "expensive" (transmitted text much much much larger than the message to
be sent).
Note that this type of crypto is ancient. It is also pretty limited in
its usefulness. Modern crypto means that you do not have to worry about
your message being readable. No Stego needed.
I don't understand why you want to unconditionally neglect certain
feasible alternatives which under suitable conditions could be useful.
Analogy: If you have a motorcycle, you can from your home extremely
fast reach any arbitrary point of your city. But in some cases riding
a bicycle may be just as fine and perhaps also more enjoyable, though
the radius of reachability is for practical reasons comparatively
limited and it takes more time to get to your destinations.
Enjoyable? You can play whatever games you want, by yourself. Your
attitude presents a really silly approach to crypto. It is a deadly
serious business, and advocating bad crypto to people whose lives may
depend on it is not very good. It is extremely labourious, and, like a
one time pad, is really bad for repeated use. It is extremely limited in
the messages it can send. Sure it could be useful in extremely special
and limited cases, and has been used that way for decades and on every
TV show. (Just last week I watched on in which a person was wired up an
in a nest of criminals, and if he mentioned chickens his collegues
should rush in and rescue him.)
Am 10.04.2016 um 23:28 schrieb William Unruh:
On 2016-04-10, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 22:21 schrieb William Unruh:...
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
It is called pgp. You just encrypt your message. No need to anticipate >>>> the questions beforehand. It can ge used for any possible message you
would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you
are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really
creative in the way you manufacture the message. It is also horrendously >>>> "expensive" (transmitted text much much much larger than the message to >>>> be sent).
Note that this type of crypto is ancient. It is also pretty limited in >>>> its usefulness. Modern crypto means that you do not have to worry about >>>> your message being readable. No Stego needed.
I don't understand why you want to unconditionally neglect certain
feasible alternatives which under suitable conditions could be useful.
Analogy: If you have a motorcycle, you can from your home extremely
fast reach any arbitrary point of your city. But in some cases riding
a bicycle may be just as fine and perhaps also more enjoyable, though
the radius of reachability is for practical reasons comparatively
limited and it takes more time to get to your destinations.
Enjoyable? You can play whatever games you want, by yourself. Your
attitude presents a really silly approach to crypto. It is a deadly
serious business, and advocating bad crypto to people whose lives may
depend on it is not very good. It is extremely labourious, and, like a
one time pad, is really bad for repeated use. It is extremely limited in
the messages it can send. Sure it could be useful in extremely special
and limited cases, and has been used that way for decades and on every
TV show. (Just last week I watched on in which a person was wired up an
in a nest of criminals, and if he mentioned chickens his collegues
should rush in and rescue him.)
IMHO your example actually supports the usefulness/efficiency of steganography.
Anyway, I don't agree with your one-size-fits-all strategy. Even for
hammers, one commonly would need different sizes of them, depending
on the nails.
As to the work being laborious, everything of utility in life has
its corresponding cost, cf. the Principle of No Free Lunch. Consider
my earlier example: Wouldn't the upper bound of the cost of devising
a stego scheme be substantially high, if the competition were something
like, say, between Boeing and Airbus for the sale of 100 machines?
(Actually longtime ago there were rumours that in one case Airbus
failed to get the contract as consequence of espionage.)
Note that in my earlier example the communications between the
representative and his manager could under circumstances contain
stuffs that would be disadvantageous to the firm -- not only for the
present but also in the future -- if these became known to the
customer. Now, if encryption is used, then in countries like the
UK there is the risk of law enforcement's demanding delivery of the
key. In that case the content of the message would not only be
releaved but also proved to be genuinely stemming from the firm.
In the stego alternative, the content wouldn't be known, since the
cover message is innocent, not to say could be proved to be genuine.
On 2016-04-11, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 10.04.2016 um 23:28 schrieb William Unruh:
On 2016-04-10, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 22:21 schrieb William Unruh:...
On 2016-04-09, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
Am 09.04.2016 um 00:44 schrieb William Unruh:
On 2016-04-08, Mok-Kong Shen <mok-kong.shen@t-online.de> wrote:
It is called pgp. You just encrypt your message. No need to anticipate >>>>> the questions beforehand. It can ge used for any possible message you >>>>> would want to send, even unforseen ones. Why the need for
steganography. Sure you could agree with your boss that if dogs are
mentioned int he message you mean yes, and if cats, no.
But steganography is limited to the rare cases when the fact that you >>>>> are using hiding should not be obvious to anyone, requires lots of
advance preparation, and is completely obvious unless you are really >>>>> creative in the way you manufacture the message. It is also horrendously >>>>> "expensive" (transmitted text much much much larger than the message to >>>>> be sent).
Note that this type of crypto is ancient. It is also pretty limited in >>>>> its usefulness. Modern crypto means that you do not have to worry about >>>>> your message being readable. No Stego needed.
I don't understand why you want to unconditionally neglect certain
feasible alternatives which under suitable conditions could be useful. >>>> Analogy: If you have a motorcycle, you can from your home extremely
fast reach any arbitrary point of your city. But in some cases riding
a bicycle may be just as fine and perhaps also more enjoyable, though
the radius of reachability is for practical reasons comparatively
limited and it takes more time to get to your destinations.
Enjoyable? You can play whatever games you want, by yourself. Your
attitude presents a really silly approach to crypto. It is a deadly
serious business, and advocating bad crypto to people whose lives may
depend on it is not very good. It is extremely labourious, and, like a
one time pad, is really bad for repeated use. It is extremely limited in >>> the messages it can send. Sure it could be useful in extremely special
and limited cases, and has been used that way for decades and on every
TV show. (Just last week I watched on in which a person was wired up an
in a nest of criminals, and if he mentioned chickens his collegues
should rush in and rescue him.)
IMHO your example actually supports the usefulness/efficiency of
steganography.
Anyway, I don't agree with your one-size-fits-all strategy. Even for
hammers, one commonly would need different sizes of them, depending
on the nails.
As to the work being laborious, everything of utility in life has
its corresponding cost, cf. the Principle of No Free Lunch. Consider
my earlier example: Wouldn't the upper bound of the cost of devising
a stego scheme be substantially high, if the competition were something
like, say, between Boeing and Airbus for the sale of 100 machines?
(Actually longtime ago there were rumours that in one case Airbus
failed to get the contract as consequence of espionage.)
And if anyone in either company suggested using your scheme I would fire
them immediately. Well established cryptography would be far far safer.
Note that in my earlier example the communications between the
representative and his manager could under circumstances contain
stuffs that would be disadvantageous to the firm -- not only for the
present but also in the future -- if these became known to the
customer. Now, if encryption is used, then in countries like the
UK there is the risk of law enforcement's demanding delivery of the
key. In that case the content of the message would not only be
releaved but also proved to be genuinely stemming from the firm.
In the stego alternative, the content wouldn't be known, since the
cover message is innocent, not to say could be proved to be genuine.
Yes, sure. A salesman suddenly sending emails to head office talking
about kittens would not be suspicious at all.
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 296 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 86:11:28 |
| Calls: | 6,658 |
| Files: | 12,203 |
| Messages: | 5,333,786 |