Some interesting thought’s on Alan Steel’s blog on the evolution of open alternatives to walled garden social networks.

alan’s blog: The open alternative to LinkedIn

Following Phil’s lead, I’ve added XFN support to my Links to other people’s blogs at right. Phil is right on the money in describing the difference between XFN vs. similar things like FOAF: if someone as lazy as me can make it work, then it’s got a chance at big-time success on the web.

It’s too bad they called it XFN, though. Unlike Phil, I didn’t mark any of my contacts as “friend” or even “coworker”. It’s not that I don’t like them… I
never had any use for Friendster and the like, and I have no use for it in this context either. I have been a big fan of LinkedIn, though - all my Links at right are labeled with rel=”colleague met” which is precisely the relationship that I express by linking someone on LinkedIn…

While blogs may be the way this linking gets initiated, it doesn’t seem like the logical place for XFN to evolve. It would be way cooler to have a site that would allow me to post my professional profile there but open it up to XFN linking (and make it point-and-click simple to create XFN links). And in true Internet style, it doesn’t have to be just one site: many sites could support this with different “flavors” - the mass-market Yahoo! Profiles vs. a more professional-oriented site, etc.. One would simply store one’s profile at the site that most closely reflects your life on the Internet (more professional vs. more personal) and they could all link together.

The missing ability right now, to my knowledge, is the ability to do a full text search of XFN relationships out to some number of degrees, but this problem will be solved.

When it is solved, the search will be far more compelling than what social networks currently offer. Instead of search the users essentially static profile, you get to search the living content they author on the web.

You can imagine decoupling the search problem from the communication problem.
Once you have an open market for each of the different problems, the market can devise an optimal menu of solutions to each of them.

An XFN search engine would allow me to figure out out that somebody interesting is, say, 2 degrees from me. I would then a variety of ways that I might seek to get in touch with that person.

I could talk in an ad hoc fashion to the colleague in the middle via email, I could use a linked-in style relay system, I could use an email reputation system we both trust, I could connect through a shared community, I could pay a small fee to get past their spam filter if they like that system, etc, etc.

It’s a great weakness that the social networks attempt to impose a single, members-only solution to the communication problem.

Network visualization is a topic of interest to me– allowing users to interactively explore large networks.

One of the best tools out there is called Thinkmap; is unfortunately closed source and too expensive for most organizations. A nice demo of it is at the
Thinkmap Visual Thesaurus

Some related open source projects are:

FoafCorp Visualization.

foafCORP is an example of using foaf in an interesting visualisation application. It turns out that there is a lot of cross-pollination amongst the boards of directors of large companies in America. By using foaf, these relationships can be described in a manner that is easy for a computer to understand. All this demo does is provide a User Interface to that data.

LJNet

LJNet is an interactive visualization of LiveJournal.com members’ social networks. It shows the friends and friends of friends of any given LJ member in an aesthetic and easy to read network - the first image below shows how the friends of our first sample member (b1uebutterf1y) are displayed around her icon (sorted into cliques), their interonnections amongst themselves are shown by curves, and any of their friends not already in the primary ellipse are placed in parabolas further out from them (repeated 2nd degree friends are repeated in multiple parabolas).

Social Circles intends to partially reveal the social networks that emerge in mailing lists. The idea was to visualize in near real-time the social hierarchies and the main subjects they address. When subscribing to a mailing you never know who the principals are, how many people are listening or what subjects they are talking about… At a glance it allows an easy way of grasping the whole situation by highlighting who is participating, who is “visually” central to that group, and displaying the topics everyone is talking about. How does the list structure itself? Is it moderated? Is it chaotic?

I have added XFN annotations to my blogroll; you notice the new logo to the right:

XFN (short for XHTML Friends Network) is a really simple way to annotate hyperlinks with information about personal relationships. It took me all of five minutes to add it with no tools support whatsoever.

For example, if want to say that Alan Steele a friend and a coworker, I just add the attribute rel=”friend coworker” to the hyperlink to his blog:

< a rel="friend coworker" href="http://www.drizzle.com/~asteele/blog.html">Alan Steele < /a >

Search engines are already starting to crawl this information and provide more meaningful relationship based results. For example, there is a search engine called Rubhub that has indexed these relationships:

See my relationships on Rubhub

Rubhub is a very basic version of what could be quite a powerful relationship map built on top of hyperlink annotations.

From here, you can imagine a virtuous cycle [1] of increased usage, search engine and web services support, better tools, and a more sophisticated relationship schema. Eventually stronger notions of identity, signing and trust will be possible on the open internet in a way never before possible.

The social network would span the open internet instead of being a walled garden and expose itself as a web services platform on top of which interesting applications could be built.

---

[1] You could tell the same “virtuous cycle” story with FOAF or many other technologies. The critical point is that the initial standard has to be so simple and easy that an interested core of people can do quick and dirty adoption without tools, as was the case with the the web itself.

AkaSig - Web scraping with python

“This series of messages introduces my current hacks that automate web sites crawling and data extraction from HTML pages. Here is part One : how to crawl complex web sites with Python. The next part will deal with data extraction from the retrieved web pages, involving much HTML cleansing and parsing.”

Useful tips, tricks, and tools for anyone who scrapes the web using python.

We’re pleased to welcome Joe Goldberg to the Jobster team.

Joe joins us from the Amazon personalization team. Joe has a great track record in designing innovative features for collaborative filtering and personalized suggestions, and in building web services that work at a massive scale.

More on the topic of social networks as walled gardens from the Cornucopt Dev Log.

“Systems like Friendster, Orkut, LinkedIn, MySpace, Tribe.net and on and on all have inflated opinions of their own utility, and most are closed to outside use anyway. They don’t even make good use of the rest of the net themselves; they’re walled gardens.

No, the social networking systems I’m interested in 1) haven’t been built yet, and 2) will have, in true small-pieces-loosely-joined fashion, deliberately low utility by themselves.”

If you rethink a social network as a web service designed in a modular fashion to interoperate with other web services, the walls of the garden dissolve.

For example, I could my extended network as a filter on incoming email– without having to reinvent email or convert everyone I know to my social networking religion.

Or as Jon Udell notes in the on-demand blogosphere,

“What if the blogs we read didn’t just scroll past us in our RSS inboxes? What if we could consult the wisdom of our networks of bloggers on demand, in realtime, relative to topics of current interest? And what if we could consult their networks too? “

The barriers to social network as web service are significant. Existing social networks want to to be a web destination that serves advertising, and are resistant to the idea of someone else applying a different skin and integrating additional functionality into their platform.

In upcoming posts I’ll give more thoughts on how such a service might attain critical mass and traction.

The Internet interprets walled gardens as damage and routes around them. (Unless the Internet is feeling grumpy. Then it just smashes through them.)

Social networks such as Linked In, Orkut, and Friendster have shown the value of being able to search your network of relationships, and control visibility and communication based on degrees of separation.

At the end of the day, though, all of these services are fundamentally walled gardens. Their limited features and scale make them vulnerable to open alternatives.

The social networks want you to take your already existing networks and recreate them in the context of a walled garden which they control and monetize. They seek to reinvent the networks of relationship already manifested online in the web, in people’s address books, in the blogosphere, in mailing lists, and so forth.

Walled gardens survive when there is no open alternative. The open internet and the companies that support it have proved ferociously creative at providing open alternatives when there’s a clear value proposition. Like life, the Internet finds a way.

The internet is learning how to take the social network value proposition and make it apply at scale.

The Foaf Project, an initial attempt at enabling relationship mapping on the open internet, was too geeky and has failed to achieve mass adoption. XFN is a lighter weight standard that attempts to accomplish the same thing using rel tags on hyperlinks, and rides on the momentum of blogs and blogrolls.

Now the big players on the open internet are getting involved in mapping the human internet. Google, for example, is investing in email, blogging, discussion groups, instant messaging, and of course social networking.

At Jobster, our philosophy is that the scale and power of the open internet will win out in the end over walled gardens. We apply this philosophy to the problem of finding better ways to connect companies and hidden talent.

Zdnet reports a growing trend toward mobile search built on top of SMS. For example, 4Info is expected to announce a $7.5M funding round from US ventures. Google has its own SMS service as does Yahoo.

For me, mobile search on a rich device like the Blackberry is an absolutely compelling application. I still need to be convinced about the utility of T9 search on a device with a tiny screen. Let’s assume, though, that SMS search becomes compelling or enough people migrate to richer devices.

The key questions for companies in this space are the revenue model and competitive differentiation. I believe the thinking that has been applied to these questions has been overly constrained by past assumptions about desktop search.

An additional constraint, as always, has been the shortsighted greed of the carriers themselves. I like this quote from an O’Reilly article titled 2005 is Year of Local Mobile Search

The beauty of a application like Berry 411 is that it that allows me to encourage charitable contributions, provide a lot of value to many people with minimal effort, and focus purely on delivering a usable application while sidestepping all of the difficult questions of monetization.

I believe a lot of innovation in mobile devices will bubble up from the grassroots, as in the early days of PCs– if carriers and the data providers don’t kill it.

Before

After

Thanks to our friends at Google Local, Canadian Berry 411 users now enjoy the same Google local results that the US does, including driving directions.

Driving directions are working for everyone again. (When the very cool Google maps were introduced, they broke Berry 411 driving directions.) Because Google maps are PC only, Berry 411 still links to Mapquest.