Keepin it simpleSaturday, December 27, 2008 by Dave Winer.
Then, enter FriendFeed and its API, which does a bit more than Twitter's, but then FriendFeed does a lot more than Twitter. There were some things inexplicably missing from FriendFeed's API, I lobbied for them, but they either haven't appeared, or when they did, they didn't do what I asked for. I don't know or care why, that's not what this post is about. Rather it's to say in one place what I've learned about FriendFeed-like services, and leave behind the notes, either for FriendFeed itself or for a comparable service. 1. FriendFeed should both import and export OPML subscription lists. The attributes specified on opml.org are necessary and sufficient for it to work with all other feed reader software, as far as I know, because there was a minimal set of attributes at the beginning, when Radio 8 implemented OPML import/export. I was able to create a simple utility that exports a user's OPML from FriendFeed, but for it to really work, it should export the addresses of the feeds the user is subscribed to, not the addresses of the FriendFeed users, which is all I can access through the API. It should be possible for the user to completely disconnect from friendfeed.com and take their subscriptions with them. This is another instance of "people come back to places that send them away" -- if you give people complete freedom to leave, they feel more comfortable about staying, building their presence on your service that may come in the future.
3. RSS description elements seem to be a big problem for FriendFeed, but I don't understand why. It's true that they are used for two different purposes: In the classic way, as the description of a longer article, or to contain the full text of an article. They can contain encoded markup. So, imho, this is how they should deal with descriptions. Say the maximum length of a comment in FF is 1024 characters (I'm not sure what the actual limit is, but it doesn't matter). First, strip all markup and then if the resulting string is longer than 1024, truncate it to 1021 characters and add three dots at the end to indicate that there's more. I don't see what else they need to do. It could be I'm missing something, of course -- Murphy's Law, etc. In the last two cases, to get the behavior I've wanted I've had to code to the API, which seems very wrong, when there are feed-based ways to do both things. Low-tech is always the right way to go, imho. There are many people who can create feeds who can't program to an API, and they shouldn't have to for things that can be done with feeds. I know that FF has proposed richer mechanisms for change notification, I'm not going to comment on those at this time. But first, before going the complex route, support the common language already used in the market you're entering. You'll find the natives more friendly if you do, imho. |
"The protoblogger." - NY Times.
"The father of modern-day content distribution." - PC World.
One of BusinessWeek's 25 Most Influential People on the Web. "Helped popularize blogging, podcasting and RSS." - Time.
"The father of blogging and RSS." - BBC.
"RSS was born in 1997 out of the confluence of Dave Winer's 'Really Simple Syndication' technology, used to push out blog updates, and Netscape's 'Rich Site Summary', which allowed users to create custom Netscape home pages with regularly updated data flows." - Tim O'Reilly.
My most recent trivia on Twitter. |