[streptomycin]

app.net looks more attractive, until the same shit starts happening there in 5 years. The problem is walled garden social networks. It's a fundamentally bad idea for everyone involved, except the people who own it and can turn a profit. StatusNet/Identica is a much better solution to this problem than yet another walled garden. Free, open source, and federated. How many times does the same shit need to happen before people realize the only long-term solution?

[MBCook]

The possible difference there is that app.net has both a business plan (sell access to post to the service) as well as a feedback model if they do something stupid (people will stop paying them).

[streptomycin]

Sure, that's a difference, but a rather incremental difference IMHO. Not nearly as big a difference as federation and open source, which is what we really need to break out of the walled gardens.

[Firehed]

How many successful open-source platforms are there? I can't think of any. Unlike Free software, platforms have a maintenance cost (servers), and that cost cannot be paid by donated man-hours.

The only thing close that I can think of is BitTorrent due to the distrubuted infrastructure, but that's quite impractical to duplicate in this context (how can browsing the content contribute back to hosting and storage automatically?)

In an ideal world I agree, but that's not where we live.

[streptomycin]

There are many. Email is a particularly notable example. Luckily, the technology sector wasn't run by "entrepreneurs" back when email was invented, or we would have wound up with walled gardens there too (Hotmail users can only email other Hotmail users and contacts can't be exported, etc).

[nostrademons]

But it was. Back then, you had "online services" - Prodigy, Compuserve, GEnie, and AOL. And for the most part, users on one of those services could only mail other users on the same service.

Email and the Internet won because the sector was competitive enough that no one service had enough market clout to not offer e-mail. Once some people got e-mail accounts, other people wanted to e-mail them, and threatened to move off the services that didn't support it in favor of services that did. That scared the online services, so rather than lose, they all started offering e-mail themselves, until they lost anyway.

If you want open social-networking platforms, the solution is to make the market competitive enough that no company can afford not to be open. Usually that'd involve choosing to use the smaller players until they're big enough that no one company is gigantic.

[peteri]

We did have walled gardens. I had accounts on Compuserve, bix, cix, MCIMail with a college account that did mail with bang path addressing over uucp (this is in 1986/7).

Most of the UKs email over uucp came in via the ukc gateway and IIRC I think I was charged something like 4p/1K for email that was international at that point. From memory my usenet feed came via a US bank in the City and was over a telebit trailblazer modem.

By 1990 I'm pretty sure that cix,MCIMail and Compuserve all at least had gateways to SMTP mail and within a few years it all became more more transparent and free.

[Firehed]

Interesting point, although I don't consider email a platform like I do Facebook or Twitter. While there's no central database, there's a means of determining where the right server to talk to lives attached to the username.

We'd need some sort of DNS-for-handles in order to implement something similar for a new protocol, unless you want your handle tied to a domain (in which case you've just remade email with a RESTful API and a length limit)

[streptomycin]

Look into StatusNet, it's a solved problem.

[oracuk]

It would seem to me that open standards are the answer to walled gardens rather than 'Free, open source and federated'.

I don't know of any standards-based activity discussing a twitter-like capability but multiple interoperable implementations, open and closed source, free and for money, provides an 'ecosystem'.

I would contrast an ecosystem to a platform. A platform is owned or controlled centrally whereas a successful ecosystem isn't. A well-managed platform ultimately provides the most benefits to the platform owner whereas a successful ecosystem provides the benefits to the most successful entities within it, a playing field to compete within.

[streptomycin]

Well if the standard is open and popular, there will be free open source implementations, and they will be important. And if the standard is open but not federated, it's not very useful.

[pablasso]

The whole point with app.net is that this doesn't happens. They don't mind that you use the data any way because you're paying for the platform, not the client.

[streptomycin]

That's not "the whole point". You're missing the part where it's yet another proprietary walled garden designed to maximize profits. That's not really a great model for end users.

[brianwillis]

The way you maximise profits is by making your customers happy. Twitter's customers are advertisers, not users. App.net's customers and users are the same group of people.

[rythie]

You gain customers by making them happy, maximising profits often does the opposite, but you hope most won't leave as you make the service worse/more expensive.

[streptomycin]

I know. All I'm saying is that's not the whole point. The walled garden is a fundamental aspect of both of those strategies, and in both cases the walled garden works against the best interests of the users.

[epo]

Anyone who trots out the tired old cliche about walled gardens is just spouting rhetoric. Anyone who also advocates open source is can be ignored out of hand.

The issues here are open standards, open protocols and the small matter of who pays for the upkeep of the infrastructure necessary to provide the service. Tired, Maoist-style sloganeering is merely making a noise.

[streptomycin]

I mentioned "federated" too, in case you missed it. You are right, that's the most important part. Yet another reason to support things like StatusNet over walled gardens like Twitter and app.net.

[MatthewPhillips]

App.net is going to support ActivityStrea.ms

[streptomycin]

I've never heard of ActivityStrea.ms, and their website is sufficiently obtuse that I still can't tell what it is. Can you enlighten me? Are you saying that app.net will support some kind of federation, maybe, hopefully?