Linked by Thom Holwerda on Thu 6th Feb 2014 22:20 UTC
Internet & Networking

Speaking with Wired editor David Rowan at an event launching the magazine's March issue, Tim Berners-Lee said that although part of this is about keeping an eye on for-profit internet monopolies such as search engines and social networks, the greatest danger is the emergence of a balkanised web.

"I want a web that's open, works internationally, works as well as possible and is not nation-based," Berners-Lee told the audience, which included Martha Lane Fox, Jake Davis (AKA Topiary) and Lily Cole. He suggested one example to the contrary: "What I don't want is a web where the Brazilian government has every social network's data stored on servers on Brazilian soil. That would make it so difficult to set one up."

A government never gives up a power it already has. The control it currently has over the web will not be relinquished.

Thread beginning with comment 582598
To view parent comment, click here.
To read all comments associated with this story, please click here.
snowbender
Member since:
2006-05-04

It's not much of a social network if you can only find people there who you have manually added. You see, if it worked like that all the parties that wanted to see each other would always have to manually establish the connection first, and that kind of removes a lot of the "social" from it. Also, with large friend networks it would quickly become a major drag on the networks, what with hundreds or even thousands of connections.


When I try to find back an old connection, I typically type his or her name in Google and see if I can find him or her. I don't log in on facebook and check there, then log in on Google+ and check there, and then once again login on LinkedIn and check there.

I assume that on a social network, one typically wants some "public" information available for everyone and other information only available for people that are trusted. The public information would be discovered by search engines.

Isn't it like that nowadays too...? If you want to discover information about people that are more privacy-minded, you'll have to first manually establish a connection with those people?


"Ideally you would be able to find thousands of people or companies willing to host "social network" server.

And now your data would again be at the mercy of others.
"

I think you get me wrong. I basically want federation to work. That means that every individual can host his own server with his own data, or can create an account on someone else's (individual or company) server. But every individual is free to choose a server that he or she trusts. The same way that everyone is free to chose an email account on a server that you trust. You dislike Google, then you are free to chose an email account with Microsoft for example, or run your own email server. None of those options is gonna limit in any way who you can email to. (or actually, nowadays it can limit you because of the rules that companies like Google and Microsoft use to limit spam....)

Aside from that: right from the moment you share any piece of data with anyone, you are at the mercy of others. You can take all the precautions you want, but as soon as you send data out, you're at the mercy of others.

When you send email, you are at the mercy of the people that maintain that server you send the email too. The same goes for XMPP. Those servers can also contain software to collect information on you, even though you don't have an account in there. For example if you email a lot to people that do have a google or hotmail address....


The protocol is actually XMPP, not Jabber. Jabber is a misnomer.

Well... I think Jabber is the original name of the protocol, but you are right.

Reply Parent Score: 4

WereCatf Member since:
2006-02-15

When I try to find back an old connection, I typically type his or her name in Google and see if I can find him or her. I don't log in on facebook and check there, then log in on Google+ and check there, and then once again login on LinkedIn and check there.


Well, normally people just go to Facebook and search there, ie. within the service they want to add someone to. Googling for people generally results in a whole lot of irrelevant stuff, blog posts, comments, whatnot, and takes more time than just searching within the social-network that you're using.

Reply Parent Score: 3

Alfman Member since:
2011-01-28

Soulbender,

I don't think we need a P2P social network. I don't think that will work. What we do need, is a decentralized social network. This needs to be similar to the way email works.


I wouldn't rule it out. It's the age old debate between having intelligence at the client level or at the server level. What would Tim Lee think? ;)


Thinking down the line though, P2P has a few advantages in terms of scalability since everyone contributes to the network. Otherwise we'd need to buy a lot more servers. It's easy to *say* the server's are the user's responsibility, however I suspect that would immediately turn off millions of would-be users who don't have a server and don't want to pay even a modest fee to try it. As much as I'd like to be selfless and let everybody use my servers freely, a flash flood of new users would immediately cripple me. Who's going to open their wallet to pay for the dynamic scalability that we'd need for the server based solution? I don't want the project to be at risk of collapsing each year due to $20k in reoccurring bills that I cannot afford to cover. Also, as you noted, google/facebook might take the federated solution for themselves and break it for others on purpose.

Putting these issues aside, I don't know that P2P and federated are mutually exclusive. It should be possible to build a P2P and federated hybrid network. The federated servers could use the same protocol and be totally transparent to other P2P nodes, only with a different configuration and policy. You could configure your client to delegate communications via your server instead of the P2P network. This way, we'd have the option to run a server to store/distribute data (say a corporation), but people without a server could still join as a P2P node.

Would you have concerns over a hybrid model?

Edit: OSnews is not really meant for brainstorming, and it's going to cut us off soon. However I really like the discussion we have going, would anyone else participate if I setup another channel to exchange ideas?

Edited 2014-02-08 18:08 UTC

Reply Parent Score: 3

snowbender Member since:
2006-05-04

Thinking down the line though, P2P has a few advantages in terms of scalability since everyone contributes to the network. Otherwise we'd need to buy a lot more servers. <snip>


You are definitely right in the sense that P2P effectively brings advantages.

But... I have to say you really like challenges, don't you? I think that going P2P is bringing a lot of extra complexity with it. That is more a gut feeling than anything else. There are a lot of things that seem harder for me to implement in a P2P based model. How would you define a "person" (or "persona") in that model? How would you be able to identify a "persona"?

I personally would like to base everything on public-private key cryptography. This would allow people to identify themselves by signing messages (or "stuff"). It would also allow to encrypt messages using the public key of the recipient.

Another thing on a P2P network: what about data persistence? When I switch off my laptop, are people still gonna be able to read my messages, my comments? Or how would that work?

... Would you have concerns over a hybrid model?


To be honest, the thing that I personally find most important is that the whole system is based on open protocols and does not promote lock-in. So that means I am free to chose the software I like, run it on any system I like and still be able to communicate with anyone else in the network, no matter what system or software that other person uses. So, if you pull it off on a P2P based system, I would be happy.

Federated system with real servers would of course be nice. But when designing a P2P system, I would take into account "federation" during design, but probably first focus on having a working pure P2P system. And then focus on federation in a second step.

Just to be realistic though, like "Werecatf" already mentioned, there's already several projects that tried to build an "open" social network and none of them succeeded in really becoming "visible".

Edit: OSnews is not really meant for brainstorming, and it's going to cut us off soon. However I really like the discussion we have going, would anyone else participate if I setup another channel to exchange ideas?


I would definitely like to share ideas about this a bit more!

Reply Parent Score: 3

zima Member since:
2005-07-06

Soulbender,

Actually, you replied to a post by snowbender. ;)

Reply Parent Score: 2