A group of standardized technologies are emerging that will evolve social networking from destinations we visit into something bigger – a federated address book that makes every single web site that chooses to adopt them entirely social.

Not only that but a grand vision for recommendations:

Amazon will show you what your Gmail address book friends have publicly said about a product and/or its category in any one of thousands of online communities. Finally, to help you further Amazon will offer an aggregated view of your friends’ friends opinions in a way that protects their identity.

That’s all very well, but not all my friends are equal. In fact, I have several distinct groups of ‘friends’, and to be frank, my Gmail address book isn’t one of those sets. My Gmail address book is people I email or, wider, people who have emailed me, which doesn’t imply anything about whether I care about their opinion when buying books.

The article gets it right on one level, but over-generalises on another. Yes, peers, friendship and implied recommendations are definitely The Future. Selling people a brand and a great vision is still going to be part of that, but backed up with thumbs up from people we have some reason to trust. Getting the first routes into that trust network is going to be fun; do you target the influencers and wait for their cult following to adopt their ways, or try a grassroots approach, almost as Amazon Vine have, selecting people from disparate networks and hoping each will spread the word?

And what do my friends — even my Facebook friends, who I went to university or worked with, or my Twitter friends, who post cool stuff and are doing cool things — know about my taste in American feminist fantasy literature, anyway? I don’t want recommendations from people I know, I want recommendations from people who like the same stuff as me, and my friendship groups at the moment are defined by technology, interest and business — not my reading material.

This would’ve been a different story a few years ago when I was heavily into Pratchett communities, but I’m not the only one who no longer defines their social space online by one specific shared interest. Plus, we can just switch gears: what on earth would the Pratchett readers have to impart to me about the shoes I’m browsing on Amazon? And yet there’s a reason to trust friends/social networks for some stuff. I’d love to see what my Twitter friends think of books on business or technology, or what Pratchett readers think of the George R.R. Martin series. Keep them within their domains and it’s fine.

But how do you define the domains and create boundaries? Putting together manual groups of ‘experts’ or ‘people whose opinion I trust on fantasy/nonfiction/books/DVDs/shoes/electronics/*’ is a pain. That’s why nobody does it. Generating automatic groups of people based on similar browsing habits, similar buying habits and similar, although not necessarily the same, social networks is more likely to be the answer. None of this OpenID OpenSocial ‘here’s my addressbook and tell me what all these people I email about dubious transactions in Nigeria think’ stuff.

Let’s not get too carried away, eh?

Plista’s presentations grabbed my interest from the start, since Dominik (speaking, above) outlined the information overload problem in stylish detail. Of course, it’s a pretty big problem and there are plenty of ways to solve it — Plista focuses on the recommendations approach.

The key components seem to be collaborative filtering and use of social graphs to build up a recommendations system that works across different sites; this allows for some interesting user tracking and the creation of central information profiles that can then be tailored to different output. A really interesting idea was the preference Doppelganger, someone who likes the same stuff you like — this is something I’ve vaguely touched on in the past.

While someone might like the same things as me, so I can use their behaviour to predict my own, I’m actually quite interested in the social graphs around that person and how they resemble mine; how the influence web bends around them compared to mine; and how that person’s key information sources differ from mine (are they reading a blog I’m not?). I guess this is more meta-informational than based in product recommendations, but there’s such a huge potential for discovery here. On the other hand, we once again run into privacy issues. My preference profile is going to be very diverse, do I really want people interested in the same technologies as me told which bands I like? To what extent are we really similar? And how do the different preference dimensions interact with each other? (Maybe people who use Open Source software prefer organic, eco-friendly products. Maybe they don’t.)

But back to Plista. An engaging presentation and a product that looks interesting, especially due to its cross-domain, supra-web type nature. An information overload company worth watching, even if it isn’t solving the problem in the same way I am.

(A bit more about Plista at CNet and Crunchbase.)