App.net to support activitystrea.ms, pubsubhubbub, Webfinger, feeds
daltoncaldwell.com
daltoncaldwell.com
I've been on the site for a good week now (@kristian) and there's some incredible people there. Great conversation, great progress on determining just what this project is. That's not a bad thing — the nature of a project like this is that it is evolving constantly.
In that vein, I encourage people to check out an issue filed at the App.net API page on Github.
https://github.com/appdotnet/api-spec/issues/33
The basic idea is a reworking of the API into something more extensible. If I'm understanding correctly (I'm pretty new at this stuff), the API at this point resembles the use case of something like Twitter: users have many posts, posts have text, a date, etc (Rails associations, anyone?). This issue proposes that the access control on those posts be variable, to fit an infinite amount of use cases. A couple examples are Twitter-style DMs (posts visible between two users), mailing lists (posts visible between specific, but multiple users), etc.
I think the thing that is causing App.net problems is that people think they are funding a Twitter clone. The fact is that the basic system of "users" with "things" goes a lot further than Twitter. It's email, it's chat, it's notifications, it's whatever you want it to be. And that's what's fascinating — we're funding an extensible piece of the next phase of the Internet — something decentralized and more or less living and breathing.
So here comes the part where I tell you to fund it. But I'm not going to. It's your call. I'm a huge fan of the service already and I can tell you that within the last two days, we've had a mobile web app, native iOS app, and streaming web app pop up out of nowhere. It's a crazy active community, and now's the time to get in. If you want to fund it, you probably know by now where to do that. There's my 2 cents (though arguably that was like 80 cents).
How so? Has anything changed in that regard?
edit: crowdfund
This sucks. A distributed social network would be incredibly cool. On the other hand we've known we've needed one for a long time but nobody can seem to build on that catches on. Maybe there's a reason for that.
Here's a question for the app.net leadership. If app.net ever became successful, and someone came along with a real distributed social network and tried to Padlister your Craigslist, what would you do? Does your commitment to open APIs extend that far?
This post sounds more negative than I intended. Marginal improvement is still improvement! I'm actually starting to get kind of excited to see what happens.
There's an awful lot to be said for the focus of doing just one thing, well. I've never once thought to myself that the thing that Twitter really needed was complex ACL settings.
Unless there exists the ability for people to have their own App.net servers written in their own languages, then this is not decentralized. The recent trends towards web-app silos for our data is relatively new in the history of the Internet. If you want something decentralized, you need multiple clients and servers. You need the ability to set it up on your own hardware. You need the ability to make a new implementation if you don't like the old one. You don't need a single company running everything.
I'm not a backer but I'd rather they spend time developing their MVP and building infrastructure so they don't have the same issues that were so pervasive in Twitter's early years (Fail Whale every hour, anyone?).
As for PubSubHubBub, all they have to do is add a tag to the feed with the Hub's url and send a POST request to the Hub when the feed has updates; that's it.
For the effort is takes (or should take, assuming they have a decent architecture), and the advantage of being automatically compatible with a bunch of existing applications, it'd seem strange not to do it.
I don't see the point in this trying to be everything to everyone sort of deal.
But for RSS applications: considering they already have more than 14000 Twitter recipes, supporting IFTTT seems an excellent reason for implementing RSS. Particularly since it should take a few dozen lines of code, if that, in any decent language.
I like what they're planning to build, I just want that to be a layer built on the open internet...
The video is "I, I, me, me". And less about "we".
Decentralised would be awesome, with some sort of non-profit organisation accepting donations and paying bills.
It could be all decentralized in the same way that email is - with DNS records. And once the protocol is defined, there could be many implementations. What would be needed is:
1) a server to server protocol (SMTP)
2) a client to server protocol (perhaps a web interface would suffice, but really a client protocol would be better). (IMAP/POP3)
3) Implementations of a server and a client (and a web interface for the server). (sendmail, mutt/thunderbird/etc...)
Here's hoping that sometime Sunday night circa 11:55pm Larry Paige hits pause on the Gangnam Style/Kanye remix he's been marathoning, Chromes his way to app.net and kicks in the last 75 grand to fully fund it. Like a boss.
> App.net will only be funded if at least $500,000 is pledged by MONDAY, AUGUST 13 at 11:59PM PDT.
Why is this? They aren't going through Kickstarter, it's their own fundraising system and a deadline chosen by the App.net folks themselves. It's just an arbitrarily chosen deadline? A self imposed constraint that they can seemingly modify at will?
They made their own rules, but now they have to stick to them.
I'm not an open source zealot but I've been saying from the beginning that it's the one piece that's missing here.