JSON, and the Piping Bots
When it first started, it seemed like a novelty act – a really cool one, sure, but more ‘interesting’ rather than ‘immediately useful’. It was slow, and a bit buggy (it seemed). The number of different modules you could pipe together was somewhat limited. I looked around, set up a couple RSS feeds for myself, and then didn’t push on it much harder.
Since then, their team seems to have been really busy. Not only has the reliability and performance improved (at least to a degree), but they’ve added a steady stream of new modules.
But this one, posted a few days ago, looks like a significant upgrade: “Power Your Own Module.”
Basically, they’ve added a module that can make calls to ‘web services’ (think: web-accessible applications) using a simple structured language called JSON.
And that, kids, is pretty much it. I don’t know how fast it works, or how reliably — I haven’t kicked the tires yet, and I don’t know what the performance issues are. But! In principle, this looks like the Final Module. No more waiting for them to add something a new feature or function, you can just write it yourself if you want.
Um: Sweet.
August 17, 2007 at 8:18 am
Pipes! These are actually “interactive data aggregators, manipulators” and online data mashup mashers. Could we get this straight puhlease.
August 17, 2007 at 8:42 am
Technically, almost everything is an “interactive data aggregator/manipulator”.
Actually (one of) the direct ancestor(s) here is the old old concept of ‘pipes’ in Unix. You know, “cat foo.txt | grep “bar” | sort | uniq” etc.
August 17, 2007 at 10:41 am
Yeah, “interactive data aggregation/manipulation” sounds like a good definition of sentience.
Om mani padme hum.