Entries from May 2008 ↓

1996 PBS documentary by Robert Cringley on the history of Personal Computing (via).

Summary:

Advogato is a community blog for free software developers, founded in 1999 as a testbed for ideas on attack-resistant trust metrics. The site now has 13k registered users, of whom over 3000 are ranked with one of the “Apprentice”, “Journeyer”, or “Master” certifications. Though I neglected the maintenance of the site for many years, it has retained an active community, and is seeing significant new life since it was handed over to the new maintainer, Steven Rainwater. By the exponential-growth standards of the dot-com boom, Advogato has been only a modest success. Yet, the experience of the site over the years contains a number of lessons. First and foremost, attack-resistant trust metrics do work. The site succeeds in being remarkably spam-free, as well as completely open to the worldwide community of free software developers, and achieves these goals without needing a huge amount of manual input to delete spammers. Thus, the main lesson is that trust metrics do work, but they need to be applied with care. Experience with the site teaches the importance of choosing and implementing the appropriate trust metric for the assumptions at hand. There is widespread “cert inflation,” where many users are ranked higher than the guidelines would recommend. The trust metrics also did not bring a flow of very high quality articles to the front page. Another important lesson is that openness and transparency work. The workings of the trust metric (including the complete source code) is public. Thus, Advogato strongly refutes the prevailing wisdom that secrecy is needed for spam protection. This lesson is similar to the ineffectiveness of “security through obscurity”. Lastly, I’ll spend some time discussing why Advogato failed to catch fire in the public’s imagination, despite its qualities. Possible factors include lack of promotion, and fact that the trust metrics were never tested against real money

More on this:

http://www.levien.com/
http://www.levien.com/thesis/compact.pdf
(via)

Getting married as an example:

• you do it just once (or a few times in life)
• complex
• (lots of issues and interdependencies)

Jonathan Zittrain promotes self-regulation when it comes to searchengines and privacy (which I think is naive).
(via)

From “Dive into” fame, interviewed by Dion Almaer. On Google Doctype, via.

From German-language Elektrischer Reporter, an excellent source of interesting video interviews (a lot of them with English-speakers, so also interesting if you don’t understand German).

Google Friendconnect is :

• a Google-hosted social network
• on a site-by-site basis

Webmasters can:

• have their own social network
• have user management and apps packaged as opensocial widgets provided by Google
• Get reports on user signup and usage

Users can:

• sign in with several credential: Google account, Yahoo account Facebook, AIM…and OpenID - other sites coming (any OpenSocial container probably?)
• “link” your identity provider (from the above list) with “friend containers”: Facebook, Googletalk, or some open social containers like Hi5, Orkut etc (on Frienconnect level)
• Publish activity back to their social network

First reactions on Twitter and blogs were enthousiastic about the “openness”. Why?

• I guess because of the OpenID and other ways of authenticating?
• The fact that you decide on your way of authenticating on a site-by-site basis

Still….

• Google totally owns user-site signups, relationships, account linking…
• thus even more of your online behaviour
• (and gets yet another piece of tracking javascript all over the web?)

What I would be interested in:

• Self-hosted Friendconnect - or at least the Friendconnect provider of your choice
• So there would be an open implementation, like Shindig is an open OpenSocial Container implementation

Even more interesing:

• social network functionality inserted by the user’s browser, not by the site owner
• (so think Greasemonkey inserting opensocial apps communicating with the Friendconnect implementation of your choice)

Update: This quote from Techcrunch nails it:

Basically, what Friend Connect does is gather this data from big social networks in whatever way they make available and then presents it in a uniform way to third party sites. It also works as a pass-through between those third party sites and the big repositories of social data. This eliminates any programming hassles on the part of small Websites that want to tap into these social networks, but it also positions Google as the central switch connecting all of these different identity systems.

These are the Campfire (announcement) videos available on youtube - less interesting however (and no part2):

Part 1

Part 3

Part 4

More of these interviews with Foocamp Attendees in the Oreilly Youtube account.

Presented at at the Emerging Communications (eComm) conference in Mountain View, California, on Wednesday 12th March 2008. David is only 21 yrs old according to the speaker???