It looks more like a Twitter replacement rather than a Facebook.
Diaspora: One Month In
71–80 of 89 posts
Re: Diaspora: One Month In
#72It's disappointing to me to see them choose regular HTTP for communication. I think XMPP PubSub or just regular XMPP messages would work better, and you could use the existing federated jabber networks. With this, they are probably going to run into NAT traversal issues, etc.
I actually spoke to these guys directly about this at a party. Only for a few minutes cause I am sure everyone and their mother wants to bike shed with them. But yeah, xmpp might have been the better choice. Google certainly thinks so.
Protocol is all they have to get really right out of the gate.
Re: Diaspora: One Month In
#73So, if I'm reading this right, the servers will be totally distributed? I don't care about a Facebook replacement so much as a totally distributed file-sharing system where trust and recommendations are based on my real-life friends. It seems like that could happen with this. Also, I think file-sharing is a great hook for the everyday Joe. And it's somewhere Facebook can't easily go.
This was already done by Facebook and I assumed removed for liability reasons. See:
http://en.wikipedia.org/wiki/Wirehog
Though I do agree that the distributed nature provided you have encryption will make it very hard to catch people sharing files. Though if you stream the videos/images/books from your site to read/watch them you can get caught at that point(If ISPs complied with CopyRight enforcement strategies.)
Re: Diaspora: One Month In
#74It's disappointing to me to see them choose regular HTTP for communication. I think XMPP PubSub or just regular XMPP messages would work better, and you could use the existing federated jabber networks. With this, they are probably going to run into NAT traversal issues, etc.
If it can't be hosted on the crappiest major shared host provider you could name, it isn't a fit for their goals.
Which means it's a total mystery why they're using WebSocket, but yeah.
Re: Diaspora: One Month In
#75I actually like this refreshing new approach to writing open source software: keep it closed for several months & code like madmen. If they can get UX people involved in the closed stage we might even see an open source product with an elegant ui for once!
Yes! It shouldn't be too hard for them to find some good UX people in the middle of SF and still flush with Kickstarter funds; they'll just need to know what to look for. Of course, there's still the issue of actually countering Facebook with a more organic social structure. If you will excuse my carting out of this fine dead horse once again: http://interuserface.net/2010/05/how-diaspora-can-succeed-mo...
Indeed, they've already found good UI people, in the very building they work in: they're being hosted for the summer by Pivotal Labs.
Re: Diaspora: One Month In
#76What I want to see: DRM for the people! Incorporate White Box Cryptography and the whole suite of tricks so that users can share certain info then unshare it. It wouldn't have to be perfect to be of tremendous benefit to many. (Girlfriend revenge sites won't be thrilled.)
Can you describe how you envision this working? Why would you choose to use a client that correctly implemented the drm features? How could a free/open source client even accomplish the drm?
Re: Diaspora: One Month In
#77What I want to see: DRM for the people! Incorporate White Box Cryptography and the whole suite of tricks so that users can share certain info then unshare it. It wouldn't have to be perfect to be of tremendous benefit to many. (Girlfriend revenge sites won't be thrilled.)
Can you describe how you envision this working? Why would you choose to use a client that correctly implemented the drm features? How could a free/open source client even accomplish the drm?
I'd only implement this for completely closed platforms. All bets are off for hacked and jailbroken devices. The point isn't to provide absolute security. It's to provide an audit trail and some measure of security -- enough to foil an average consumer, but not enough to stand up to concerted attack.
Why would you choose to use a client that correctly implemented the drm features?
To avoid prosecution under the DMCA. To avoid having to jailbreak your device. Because the effort of hacking isn't worth it. Current streaming DRM technologies have a good-enough track record, in part because what they protect isn't that valuable.
Big companies don't have access to perfect DRM, but it does useful things for them. Why shouldn't individuals have access to the same tools to protect their own data against such companies? In particular, big corporations are just as vulnerable to legal remedies as individuals.
Re: Diaspora: One Month In
#78Earlier quoted context omitted.
I'm not quite sure what makes you say "the 4 founders...have very little technical knowledge and have outsourced the majority of their code." Here's what I read, After getting settled in ole San Fran, our first day at Pivotal Labs was June 7th. Not only has Pivotal lent us desks and monitors for the summer, they push us daily to drive development from the interface and focus on the experience, rather than providing j…
Pivotal Labs doesn't work for free. If you go to Pivotal's front page, what they do is pretty much do all the coding for a startup. They take a business person's ideas and then code them up and implement them. From there they then teach the business people either how to maintain it or how to hire a dev team that can (or work with the dev team they've hired to learn how to work with the codebase). Its just what pivota…
You may be right, but that's an awful lot of speculation. As for "Pivotal Labs doesn't work for free," the publicity of "helping the famous kids dream big" may be more than enough payment. (That's assuming we want to be ultra-cynical, which I have mixed feelings about frankly.)
Re: Diaspora: One Month In
#79... so when you post a message, it's like sending 300-400 messages to up to 300-400 servers? Uh, yeah, that'll scale.
IRC servers can have lots of clients that receive messages, and they have no problems sending thousands of messages to connected clients. Diaspora's nodes seems to be both server and client, but the mechanics are the same. When it receives a message, it's just like a IRC client, and when it sends a message it is like an IRC server. Unless each physical server is overloaded with busy nodes, I think it will scale well.…
OTOH, if you have 20,000 "friends" maybe Diaspora should force you to buy a more powerful server. Social networking is one thing and social marketing is another.
Re: Diaspora: One Month In
#80Earlier quoted context omitted.
Today's concerns for security are not concerns for keeping our information private. Rather, they are concerns that we will maintain control over our private information, to be able to share it at our own discretion. Facebook removes much of this control from the user.
no evidence to show that facebook users see it that way, or care. Also FWIW, I think the name "diaspora" is terrible I cannot see it working at all.
The name can be changed, what I want is the technology.
Show me the code and let me develop on it. Let me change the interface, add features, and do analysis on my social graph. Depending on the license they release the code under, there could potentially be other services that use the technology and have a nicer name than Diaspora.
Facebook users might even take notice if Diaspora has a "dislike" button and access to their Facebook data via Diaspora.