Recent searches
Search options
@elplatt @evan@prodromou.pub @cwebber @tsyesika
At the time we started OStatus, and to some extent when AP started, there was still a big culture of polling feeds.
For example, feed readers would poll the same feeds hundreds of times per day, because they could never be sure when something would be posted.
The general consensus was that this was a big mistake and that push was the right way to do it.
@elplatt @evan@prodromou.pub @cwebber @tsyesika
@rabble and @kellan gave a really important talk at OSCON 2008 called "Beyond REST? Building Data Services with XMPP PubSub".
They talked about how the feed-polling architecture was costing an arm and a leg for publishers with only tiny benefit to users.
I was in the audience and it was a really important inspiration for later work.
@elplatt @evan@prodromou.pub @cwebber @tsyesika @rabble @kellan
Robert Kaye did a pretty good write-up on the talk here:
http://radar.oreilly.com/2008/07/oscon-day-1-beyond-rest-buildi.html
@elplatt I think the XMPP model helped out here, btw. ActivityPub is like XMPP + Atom over REST, if you squint.
@elplatt If I remember correctly from the Social Web Working Group, some of the initial prototypes from SoLiD were polling-based. I can't tell you if that's still the case or not.
I've had more than one "why don't you just..." reply on this thread.
Random Replier, I'm not willing to debate why your alternate poll-based design is better or not.
ActivityPub is the standard for social networking on the web. It works how I want it to work.
If you think you have a better idea, build a prototype and test it. The onus is on you to make it better or different.
And show your numbers, like Rabble and Kellan did.