Pages:
Author

Topic: [ANNOUNCE] Bitmessage - P2P Messaging system based partially on Bitcoin - page 21. (Read 89884 times)

sr. member
Activity: 249
Merit: 251
Yes, it would be possible to separate the address book from the messages.dat file. I'll add it to the list of requests although there are other priorities.

There is currently no way to command the deletion of content from the network; all content simply expires. Adding such a feature would be a drastic change. You should already be receiving messages on both clients if they share the same keys. As long as they each go online once every two days, they'll both receive all messages. If one stays online all the time and another goes online only once every four days then that second client will only receive half of the messages (from the last two days).
staff
Activity: 4270
Merit: 1209
I support freedom of choice
Two feature request (if they are possible):

- Can you divide the address book and messages? So that we can have: addressbook.dat and messages.dat

- Right now if I receive a message, it disappears from the network.
I mean that I'm not able to receive it from another client.
Is it possible to set a time expiration before deleting it from the network?
Something as it is now with the mail settings, where you can set to leave message on the server for X days.

So I can also have two different installation of Bitmessage (ex: in two different computer)
With installation A, I'll set to receive and maintain messages on the network for 10 days.
With the installation B, I'll set to receive the messages and delete them from the network.
If I'll open the installation B before the first one, then I won't be able to receive again messages on the installation A.

I know that messages that are older than X weeks will be deleted anyway from the network, I'm not asking to change this rule.
sr. member
Activity: 249
Merit: 251
The code should not be distributing and connecting to peer addresses for internal / site-specific / unrouted addresses.

You're right; it shouldn't. I will patch it appropriately tomorrow.

Thank you
legendary
Activity: 1596
Merit: 1100
Peer address issue...

Code:
Trying an outgoing connection to 10.10.10.1 : 8444
[...]
Could NOT connect to 10.10.10.1 during outgoing attempt. timed out

The code should not be distributing and connecting to peer addresses for internal / site-specific / unrouted addresses.

hero member
Activity: 784
Merit: 1000
0xFB0D8D1534241423
I doubt that because I can use your public key from the blockchain to encrypt messages against your will.

But then you can see whether or not I could successfully decrypt it and in some cases you can also see the plaintext (like if Bitmessage is being used on a server for some sort of communication). This additional information could be utilized. Cryptographers will tell you that it is bad practice to use the same key for encryption and signing. And if there are people on the forum telling everyone else that it is unsafe to share keys between Bitcoin and Bitmessage then no one is going to do it (not that I would want them to).
I agree things are simpler and cleaner, thus potentially safer with separate keys. But it might be highly beneficial to be able to send encrypted messages directly to a BTC address, so this concept deserves some thought.
For example, to tell someone why you're donating to their signature without bloating the blockchain.
sr. member
Activity: 249
Merit: 251
That's a known issue that happens when you rename the EXE. It will happen with this version and possibly versions going forward unless I can find a work-around or unless pyinstaller fixes this rather common problem.

http://www.pyinstaller.org/ticket/522


EDIT: I was able to find a workaround. The currently published EXE should function like normal.
legendary
Activity: 916
Merit: 1003
When I start the new version 0.2.7 it displays this error pop-up:



If I click OK the GUI appears and eventually connects to the network and appears to work normally.
staff
Activity: 4270
Merit: 1209
I support freedom of choice
0.2.7
- Added API. See API Reference.
- Added error handling for the case where the client tries to send a message from an address for which the human has deleted the keys.
- Improved GUI messages when doing work (or pending work) for broadcast messages.
- Added error handling for the case where the proof of work takes practically no time.
legendary
Activity: 1708
Merit: 1020
I doubt that because I can use your public key from the blockchain to encrypt messages against your will.

But then you can see whether or not I could successfully decrypt it and in some cases you can also see the plaintext (like if Bitmessage is being used on a server for some sort of communication). This additional information could be utilized. Cryptographers will tell you that it is bad practice to use the same key for encryption and signing. And if there are people on the forum telling everyone else that it is unsafe to share keys between Bitcoin and Bitmessage then no one is going to do it (not that I would want them to).
I agree things are simpler and cleaner, thus potentially safer with separate keys. But it might be highly beneficial to be able to send encrypted messages directly to a BTC address, so this concept deserves some thought.


[...]
Actually I suppose that a bitcoin private key could be used to sign a message saying that they assert that new_pub_key_1 is to be used for signing and new_pub_key_2 is to be used for encryption. That would work. Then users could use their Bitcoin addresses in Bitmessage. This would require that users import their wallet keys into Bitmessage (or vise-versa).
[...]
Sounds like a plan  Grin
sr. member
Activity: 249
Merit: 251
But to sign a message with a bitcoin address we use not the public key but the private key, right?

Right

is it possible to generate a different type of address (the label I threw out before was "Dual Address") which, unlike usual BM addresses, is the hash of the unique public key associated with a given bitcoin ECDSA private key (and possibly starts with "BTC-" instead of "BM-")?

Yes. We could also just use Bitcoin addresses at that point which might make users happier.

Is it actually impossible to get that idea to work with the BitMessage system?

Actually I suppose that a bitcoin private key could be used to sign a message saying that they assert that new_pub_key_1 is to be used for signing and new_pub_key_2 is to be used for encryption. That would work. Then users could use their Bitcoin addresses in Bitmessage. This would require that users import their wallet keys into Bitmessage (or vise-versa).

I believe that this is futile in any case: Bitcoin will (or at least should) someday switch to Pay-to-script-hash because it will help keep the long-term size of the blockchain down.
member
Activity: 115
Merit: 10
I doubt that because I can use your public key from the blockchain to encrypt messages against your will.

But then you can see whether or not I could successfully decrypt it and in some cases you can also see the plaintext (like if Bitmessage is being used on a server for some sort of communication). This additional information could be utilized. Cryptographers will tell you that it is bad practice to use the same key for encryption and signing. And if there are people on the forum telling everyone else that it is unsafe to share keys between Bitcoin and Bitmessage then no one is going to do it (not that I would want them to).

Thanks for the fast response!

But to sign a message with a bitcoin address we use not the public key but the private key, right?

Again I'm not suggesting the entire messaging system be reworked, just that a neat feature be added...is it possible to generate a different type of address (the label I threw out before was "Dual Address") which, unlike usual BM addresses, is the hash of the unique public key associated with a given bitcoin ECDSA private key (and possibly starts with "BTC-" instead of "BM-")?

Is it actually impossible to get that idea to work with the BitMessage system?
sr. member
Activity: 249
Merit: 251
I doubt that because I can use your public key from the blockchain to encrypt messages against your will.

But then you can see whether or not I could successfully decrypt it and in some cases you can also see the plaintext (like if Bitmessage is being used on a server for some sort of communication). This additional information could be utilized. Cryptographers will tell you that it is bad practice to use the same key for encryption and signing. And if there are people on the forum telling everyone else that it is unsafe to share keys between Bitcoin and Bitmessage then no one is going to do it (not that I would want them to).
legendary
Activity: 1708
Merit: 1020
Random Thought:

Why not import an actual bitcoin private key as the address-generator passphrase, and force Bitmessage to generate "the" associated bitcoin address (http://brainwallet.org/), and then concatenate "BM-" to that as the BitMessage Address? This could be have a special designation like "Dual Address" (possibly you actually concatenate "BTC-" to it instead).

Then BM could work WITH Bitcoin and every Bitcoin user would have the ability to receive BM messages (albeit through the BM chain). Theoretically I could send a message to 1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T (at BM-1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T), and anyone with the private key (5KJvsngHeMpm884wtkJNzQGaCErckhHJBGFsvd3VyK5qMZXj3hS) in their Bitcoin wallet.dat could, assuming they haven't already, install BitMessage and retrieve all of the messages sent to them. Easier to manage one online identity than two.

My understanding is that people are complaining about Satoshi dice using BTC to communicate or something, this would help solve that problem, wouldn't it? Did I misunderstand something?

People would of course retain the ability to make NEW identities/use all of the other great features of BM, but if each Bitcoin address has a provably associated BM address I think that would be pretty convenient and foster adoption!

Thoughts?

A Bitmessage address can be turned into a Bitcoin address but a Bitcoin address cannot be turned into a Bitmessage address. This is because a Bitcoin address is a hash of one public key while a Bitmessage address is a hash of two public keys- one for encryption and one for signing. If we were to retool Bitmessage and use a Bitcoin address as a Bitmessage address, the first problem that comes to mind is that we would essentially be using an cryptographic key previously used for only signing now also for encryption which could leak information about the private key.

I don't think that Bitcoin and Bitmessage will ever share cryptographic keys although it is certainly possible, if Bitcoin and something like Namecoin are ever more integrated, to be able to derive one type of address from the other in a very user friendly way. This would have other benefits- anyone could send a bitcoin or a bitmessage to 'AsymmetricInformation' without having to deal with either type of address at all.

I doubt that because I can use your public key from the blockchain to encrypt messages against your will.
sr. member
Activity: 249
Merit: 251
Random Thought:

Why not import an actual bitcoin private key as the address-generator passphrase, and force Bitmessage to generate "the" associated bitcoin address (http://brainwallet.org/), and then concatenate "BM-" to that as the BitMessage Address? This could be have a special designation like "Dual Address" (possibly you actually concatenate "BTC-" to it instead).

Then BM could work WITH Bitcoin and every Bitcoin user would have the ability to receive BM messages (albeit through the BM chain). Theoretically I could send a message to 1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T (at BM-1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T), and anyone with the private key (5KJvsngHeMpm884wtkJNzQGaCErckhHJBGFsvd3VyK5qMZXj3hS) in their Bitcoin wallet.dat could, assuming they haven't already, install BitMessage and retrieve all of the messages sent to them. Easier to manage one online identity than two.

My understanding is that people are complaining about Satoshi dice using BTC to communicate or something, this would help solve that problem, wouldn't it? Did I misunderstand something?

People would of course retain the ability to make NEW identities/use all of the other great features of BM, but if each Bitcoin address has a provably associated BM address I think that would be pretty convenient and foster adoption!

Thoughts?

A Bitmessage address can be turned into a Bitcoin address but a Bitcoin address cannot be turned into a Bitmessage address. This is because a Bitcoin address is a hash of one public key while a Bitmessage address is a hash of two public keys- one for encryption and one for signing. If we were to retool Bitmessage and use a Bitcoin address as a Bitmessage address, the first problem that comes to mind is that we would essentially be using an cryptographic key previously used for only signing now also for encryption which could leak information about the private key.

I don't think that Bitcoin and Bitmessage will ever share cryptographic keys although it is certainly possible, if Bitcoin and something like Namecoin are ever more integrated, to be able to derive one type of address from the other in a very user friendly way. This would have other benefits- anyone could send a bitcoin or a bitmessage to 'AsymmetricInformation' without having to deal with either type of address at all.
legendary
Activity: 1708
Merit: 1020
Random Thought:

Why not import an actual bitcoin private key as the address-generator passphrase, and force Bitmessage to generate "the" associated bitcoin address (http://brainwallet.org/), and then concatenate "BM-" to that as the BitMessage Address? This could be have a special designation like "Dual Address" (possibly you actually concatenate "BTC-" to it instead).

Then BM could work WITH Bitcoin and every Bitcoin user would have the ability to receive BM messages (albeit through the BM chain). Theoretically I could send a message to 1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T (at BM-1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T), and anyone with the private key (5KJvsngHeMpm884wtkJNzQGaCErckhHJBGFsvd3VyK5qMZXj3hS) in their Bitcoin wallet.dat could, assuming they haven't already, install BitMessage and retrieve all of the messages sent to them. Easier to manage one online identity than two.

My understanding is that people are complaining about Satoshi dice using BTC to communicate or something, this would help solve that problem, wouldn't it? Did I misunderstand something?

People would of course retain the ability to make NEW identities/use all of the other great features of BM, but if each Bitcoin address has a provably associated BM address I think that would be pretty convenient and foster adoption!

Thoughts?
+1
member
Activity: 115
Merit: 10
Random Thought:

Why not import an actual bitcoin private key as the address-generator passphrase, and force Bitmessage to generate "the" associated bitcoin address (http://brainwallet.org/), and then concatenate "BM-" to that as the BitMessage Address? This could be have a special designation like "Dual Address" (possibly you actually concatenate "BTC-" to it instead).

Then BM could work WITH Bitcoin and every Bitcoin user would have the ability to receive BM messages (albeit through the BM chain). Theoretically I could send a message to 1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T (at BM-1JwSSubhmg6iPtRjtyqhUYYH7bZg3Lfy1T), and anyone with the private key (5KJvsngHeMpm884wtkJNzQGaCErckhHJBGFsvd3VyK5qMZXj3hS) in their Bitcoin wallet.dat could, assuming they haven't already, install BitMessage and retrieve all of the messages sent to them. Easier to manage one online identity than two.

My understanding is that people are complaining about Satoshi dice using BTC to communicate or something, this would help solve that problem, wouldn't it? Did I misunderstand something?

People would of course retain the ability to make NEW identities/use all of the other great features of BM, but if each Bitcoin address has a provably associated BM address I think that would be pretty convenient and foster adoption!

Thoughts?
full member
Activity: 224
Merit: 101
 Smiley I was waiting for such thing for sooo long!!! I just found it today when I did my quarterly search for "decentral email". YAY!

And thanks that it already has a portable mode. (Attachments would be awesome, of course.)
sr. member
Activity: 249
Merit: 251
I'm continuing to observe the BitMessage client occasionally use 100% cpu without my doing anything.  I left it running overnight and woke up to find it like this.  In fact it was unresponsive and had to be killed using process explorer.

This is the latest version 0.2.6.
Thank you for reporting the problem. I'll keep it in mind to see if a cause can be identified.
legendary
Activity: 916
Merit: 1003
I'm continuing to observe the BitMessage client occasionally use 100% cpu without my doing anything.  I left it running overnight and woke up to find it like this.  In fact it was unresponsive and had to be killed using process explorer.

This is the latest version 0.2.6.
staff
Activity: 4270
Merit: 1209
I support freedom of choice
A web-based Bitmessage client should never exist. Just like that article talks about, there is no way to securely deliver the code. If HTTPS is secure and trustworthy then Bitmessage need-not exist. I argue that it isn't.
I was thinking about something like this as chrome app: https://chrome.google.com/webstore/detail/cryptocat/gonbigodpnfghidmnphnadhepmbabhij
And working in backround, as this: https://chrome.google.com/webstore/detail/chat-for-google/nckgahadagoaajjgafhacjanaoiihapd
Pages:
Jump to: