Boris Mann

Open Source. Community. Decentralized Web. Building dev tools at Fission. Cooks & eats.


Personal Info Cloud

Personal Info Cloud. That's the closest I can come to for what I'm thinking, and as I search a bit I see that is one of the faces of Thomas Vander Wal, perhaps best known for coining the term "folksonomy". We don't say that so much anymore, having devolved into tagging madness.

It seems to be that, even with the "aids" of RSS to more quickly skim through information, and tagging to perhaps help winnow through the particular areas we are interested in, we are reaching a kind of nexus point of signal to noise. I need my filters back. The information -- now in podcasting, photosharing, video blogging goodness in HD quality with surround sound -- is rolling towards us and we are being crushed. I probably shouldn't add something about spam (splogs, etc. etc.).

I had hoped that would be the foundation for something that could help solve this. In part, because its technology1 was based on a real time infrastructure -- publish and subscribe -- instead of this polling and pinging mess we are faced with "regular" RSS. In my ideal world, PubSub's rich boolean search interface would be available, to easily create field-level searches trawling the entire web -- e.g. return all blog posts with the word "Drupal" in the <title> field. That's an incoming stream that could be vast indeed, but it's what PubSub did best: search the future.

And next is feedback. I would tag, or bookmark, or favourite or "Shift-S" for Google Reader fans, and stick it in my Personal Info Cloud. That could then become the basis for retrospective search aka what Google does really well right now, but filtered by my own sources of information. Ideally, I could expand my cloud to a Local region (geographically, my Placeblogger friends) or my trusted network (tie into the public sources exposed by my communities). I say communities rather than the simpler family, friends, colleagues, etc., because I want to stress that there are going to be more than one cloud connected in.

It's all overlapping networks of networks.

Sure, we've got bits and pieces around. Rollyo and Eurekster's Swicki are little bits of this, allowing people to create curated, pre-populated search collections based on domains. Even Wikipedia is getting in the game with Wikia / WikiSearch2.

I keep tumbling a littler deeper into this rabbit hole. Lijit, on my list of "really must dig into this deeper", has a map your world page with...surprise! a picture of a Personal Info Cloud. That ties back into discussions with Ton Ziljstra at BarCamp Brussels.

And here I am, manually re-entering my little bits of identity into different services, looking for the Holy Grail. Perhaps convincing some friends that this time, maybe just this once, won't they also please sign up and enter their information into just one more system, 'cause then it'll be better for everyone.

So. Info Clouds. Personal, Local, Community. Search between communities, each member a pivot into another cloud. Maybe aggregating searches. I didn't really cover ratings and feedback loops, as well as the magic nature-vs-nurture algorithm that gives you a serendipity feed, that will always be magically full of things outside your echo chamber, and you can decide whether to let them in or "thumbs down" them away.

Welcome to 2007. I'm still excited by what's out there. Lots of day-to-day get-stuff-done work that needs to be done. But let's keep talking about and building the big stuff. 

1 it's still around...I think it will be used for something less social in its next iteration, although I think the right technology is there underneath to do this social thing in real time. And real time is basically mobile search.

2 There's some confusion about different services with similar names and different relations to Jimmy Wales / Wikipedia. You'll need to trawl through the comments on this TechCrunch post to figure it out.