Hi guys, I've had a quick chat with mk about a hubski API, and the consensus is that I'm going to start to spec out a RESTful API for Hubski client apps to use. I'm looking for anyone with experience developing and designing RESTful services to give me a hand with it, so that we can get the best experience for client apps as possible. I'll post any updates I have to the state of the API design under the #api tag from now on, so that I don't clog up #hubski. Let me know if you're interested in helping out!
Hubski Enhancement Suite anyone? But seriously, an API sounds awesome (and I don't actually think a HES would be necessary). All I would like to see other than mobile apps is a browser extension similar to reddit companion that gives links back to the comments page from an article/link.
I don't think HES would be necessary, since mk listens to the community :-)
While a full blown HES wouldn't be necessary, I think there are a few features that would be nice in a browser extension.
Like your script which can be found under tools? Sure, some features would be nice, but they will mostly be quite personal. Maybe when the API is done someone will hack some stuff together...
I don't have any experience developing APIs but I'm keen to learn.
If you go to https://github.com/Baltoli/hubski-api-spec you can see the API spec as we develop it. Watching how we (user92 and I, so far) design it should be very informative for you. If you've got any questions about APIs in general, feel free to ask, and I'm more than happy to help you.
Looks good! Any particular reason you are using POST only? If I want to update a comment or something it would be nice if the API supported a PUT request rather than POST. PUT would be idempotent also so if the first request failed I could reissue the same request and not suffer any ill effect... Looks good so far, keep up the good work.
Awesome, watching already. I'm working on a Win 8 app currently, hopefully I'll be able to use what you guys are developing.
In HTTP, when something sends you back a reply to a response, it includes a status code to let you know how your request got on. The "200 range" of HTTP status codes is used to indicate that your request was successful, and I wrote "200 range" only because I couldn't remember offhand what status code it would use (201, 202 etc). Does that help you?
What kind of data does Hubski keep? I'd love to be able to do some data mining as it grows/changes.
Off the top of my head: -Total number of users
-Posts/Comments in a single day
-Active users in a single day (as in, those who posted or commented)
-Lurkers per day (people who "vote" but don't post/comment)
-Post listing on feeds (not sure how I'd figure this one out. I guess basically a klout rating for posts? Not sure this is a good idea. I'll think about it) Right now I'm mostly curious how/at what speed hubski is growing, and what that growth looks like in terms of content and use of hubski. It will definitely be interesting to see what users become the "face" of Hubski, and if that occurs. I know that this is a little purist of me, but I think there's got to be a better way of doing this than the "popular" tool, since that runs into the same problems of just promoting what's already popular and snowballing.