5 ms·

I just came back from China. I could not message my loved ones, check my mail, update my social media, or listen to internet radio. This was not Facebook's faul

I just came back from China. I could not message my loved ones, check my mail, update my social media, or listen to internet radio. This was not Facebook's fault, it was not even China's fault (though they were the ones blocking me). It was the fault of the way we use existing technology. So Facebook is banned. Fine. That shouldn't stop me from being able to at least send data somewhere that will end up at Facebook. And I should have been able to send an e-mail that bounced through intermediary relays without connecting to Gmail's servers. A lot of the web that _could_ be peer to peer isn't. And it's not because it is difficult, but because we simply haven't decided to use it that way. Email is the oldest, most successful distributed decentralized application. It does not require complex consensus. It does not use distributed databases. And it just works. You don't even need to be connected to the end server. There's no reason the web couldn't work like this. It is feasible today. On top of this, where your web content is stored should be simple: it should be stored on a device in your pocket, and cached and distributed on servers around the world, just like email already is. Again, no complex algorithms are needed for this. A very simple design could accomplish this with existing technology, and perhaps an extension to a protocol's RFC and an update to some core software.

matthewmacleod 9y ago
This was not Facebook's fault, it was not even China's fault Absolute shite. It's 100% China's fault. Given the widespread, damaging, lingering ramifications of the lack of security on the Internet in general, the absolute last thing we should be doing is promoting the idea that this information and these connections should be less secure. I am baffled that anybody could even propose this idea.
- peterwwillis 9y ago
  Well, I propose it because 1) HTTPS is a tiny aspect of what makes one secure on the web, 2) I just want to talk to my loved ones on the web, 3) I don't care what China does with my data, 4) "lingering ramifications" of this are very manageable. Maybe I have a skewed point of view, but after being involved in computer security for 15 years, I literally have two cases at all where I care about the data I'm sending over the internet: logging into my bank, and making a purchase with a credit card. Virtually everything else, I don't care about.
freeone3000 9y ago
SSL isn't some slap-on after the fact that does nothing - it renders traffic unreadable to anyone in the middle, and unalterable. The fact that China deems any content it cannot read as immediately forbidden says way more about China than it does about "our current state of tech".
- peterwwillis 9y ago
  The great firewall does allow secure TLS connections that it cannot read or modify. It just doesn't allow them to specific targets, or using specific protocols.
icebraining 9y ago
If the interesting things ran over email, China would start blocking those too. Who cares where the intermediary server is, if the email headers contain the final destination? Plus, they would probably block any direct access to intermediary servers outside the country.
- peterwwillis 9y ago
  They can block all intermediary servers outside the country, that's fine. E-mail is an entire system, the way Facebook is a system. They would either need to block individual addresses, or the entire system. The former is the only feasible option with e-mail, otherwise basically you couldn't do business with the outside world, and China likes doing business. With the web, or chat apps, they block the whole target system and provide an alternative, which is somewhat reasonable. You could make this more difficult by designing the technology to make it more difficult to design alternate systems, and easier for it to work in a way that is acceptable to China and still supports existing users. This isn't a difficult or bad thing. We've had this forever on the web, where you could see a page was not using a secured connection (saying "this page is secure" just because the connection is secure is a big overstatement). It's mostly a recent development that privacy zealots have been shoving this "secure or nothing" idea down everyone's throats.
  - icebraining 9y ago
    They would either need to block individual addresses, or the entire system Well, no, they could (and would) block domains, like "*@facebook.com". Which would leave you exactly where you are now. If what you want is an a non-encrypted web where China can analyze and choose the specific content it wants to block, then fine, but I don't see what the whole email-like system has to do with it.
    - peterwwillis 9y ago
      Yes and no. You could block the Facebook domain, but due to things like alternate domains, aliases and forwarders, you can easily pass data to different domains and still have your data reach the intended recipient. The target can be made ephemeral just by configuring your mail server account. I don't think China wants to block specific content because that's a lot more complicated an approach than blocking target systems. By making the web more e-mail like, you can pass data through the web, and it can reach the intended target without that target being discovered. You could in essence borrow the Tor method, but for plaintext communications over HTTP. You could use the same approach for web data by decoupling the addresses from the content. Federated services and pre-authorized applications is one way to connect disparate content and networks to each other. Another would be a way to submit data to a network in general, and allow others to retrieve it, for example with some pre-shared key or authorized service. You could make this work across different types of data and networks. For example, I should be able to send an e-mail to some address, and an intended recipient could check their Facebook messages and see the data, even though the mail address was something arbitrary.

matthewmacleod 9y ago

This was not Facebook's fault, it was not even China's fault Absolute shite. It's 100% China's fault. Given the widespread, damaging, lingering ramifications of the lack of security on the Internet in general, the absolute last thing we should be doing is promoting the idea that this information and these connections should be less secure. I am baffled that anybody could even propose this idea.

peterwwillis 9y ago
Well, I propose it because 1) HTTPS is a tiny aspect of what makes one secure on the web, 2) I just want to talk to my loved ones on the web, 3) I don't care what China does with my data, 4) "lingering ramifications" of this are very manageable. Maybe I have a skewed point of view, but after being involved in computer security for 15 years, I literally have two cases at all where I care about the data I'm sending over the internet: logging into my bank, and making a purchase with a credit card. Virtually everything else, I don't care about.

freeone3000 9y ago

SSL isn't some slap-on after the fact that does nothing - it renders traffic unreadable to anyone in the middle, and unalterable. The fact that China deems any content it cannot read as immediately forbidden says way more about China than it does about "our current state of tech".

peterwwillis 9y ago
The great firewall does allow secure TLS connections that it cannot read or modify. It just doesn't allow them to specific targets, or using specific protocols.

icebraining 9y ago

If the interesting things ran over email, China would start blocking those too. Who cares where the intermediary server is, if the email headers contain the final destination? Plus, they would probably block any direct access to intermediary servers outside the country.

peterwwillis 9y ago
They can block all intermediary servers outside the country, that's fine. E-mail is an entire system, the way Facebook is a system. They would either need to block individual addresses, or the entire system. The former is the only feasible option with e-mail, otherwise basically you couldn't do business with the outside world, and China likes doing business. With the web, or chat apps, they block the whole target system and provide an alternative, which is somewhat reasonable. You could make this more difficult by designing the technology to make it more difficult to design alternate systems, and easier for it to work in a way that is acceptable to China and still supports existing users. This isn't a difficult or bad thing. We've had this forever on the web, where you could see a page was not using a secured connection (saying "this page is secure" just because the connection is secure is a big overstatement). It's mostly a recent development that privacy zealots have been shoving this "secure or nothing" idea down everyone's throats.
- icebraining 9y ago
  They would either need to block individual addresses, or the entire system Well, no, they could (and would) block domains, like "*@facebook.com". Which would leave you exactly where you are now. If what you want is an a non-encrypted web where China can analyze and choose the specific content it wants to block, then fine, but I don't see what the whole email-like system has to do with it.
  - peterwwillis 9y ago
    Yes and no. You could block the Facebook domain, but due to things like alternate domains, aliases and forwarders, you can easily pass data to different domains and still have your data reach the intended recipient. The target can be made ephemeral just by configuring your mail server account. I don't think China wants to block specific content because that's a lot more complicated an approach than blocking target systems. By making the web more e-mail like, you can pass data through the web, and it can reach the intended target without that target being discovered. You could in essence borrow the Tor method, but for plaintext communications over HTTP. You could use the same approach for web data by decoupling the addresses from the content. Federated services and pre-authorized applications is one way to connect disparate content and networks to each other. Another would be a way to submit data to a network in general, and allow others to retrieve it, for example with some pre-shared key or authorized service. You could make this work across different types of data and networks. For example, I should be able to send an e-mail to some address, and an intended recipient could check their Facebook messages and see the data, even though the mail address was something arbitrary.