WordPress trust syndication revisited: F2F plugin

This is a followup to my Syndicating trust? Mediawiki, WordPress and OpenID post. I now have a simple implementation that exports data from WordPress: the F2F plugin. Also some experiments with consuming aggregates of this information from multiple sources.

FOAF has always had a bias towards describing social things that are shown rather than merely stated; this is particularly so in matters of trust. One way of showing basic confidence in others, is by accepting their comments on your blog or Web site. F2F is an experiment in syndicating information about these kinds of everyday public events. With F2F, others can share and re-use this sort of information too; or deal with it in aggregate to spread the risk and bring more evidence into their trust-related decisions. Or they might just use it to find interesting people’s blogs.

OpenID is a technology that lets people authenticate by showing they control some URL. WordPress blogs that use the OpenID plugin slowly accumulate a catalogue of URLs when people leave comments that are approved or rejected. In my previous post I showed how I was using the list of approved OpenIDs from my blog to help configure the administrative groups on the FOAF wiki.

This may all raise more questions than it answers. What level of detail is appropriate? are numbers useful, or just lists? in what circumstances is it sensible or risky to merge such data? is there a reasonable use for both ‘accept’ lists and ‘unaccept’ lists? What can we do with a list of OpenID URLs once we’ve got it? How do we know when two bits of trust ‘evidence’ actually share a common source? How do we find this information from the homepage of a blog?

If you install the F2F plugin (and have been using the OpenID plugin long enough to have accumulated a database table of OpenIDs associated with submitted comments), you can experiment with this. Basically it will generate HTML in RDFa format describing a list of people . See the F2F Wiki page for details and examples.

The script is pretty raw, but today it all improved a fair bit with help from Ed Summers, Daniel Krech and Morten Frederiksen. Ed and Daniel helped me get started with consuming this RDFa and SPARQL in the latest version of the rdflib Python library. Morten rewrote my initial nasty hack, so that it used WordPress Shortcodes instead of hardcoding a URL path. This means that any page containing a certain string – f2f in chunky brackets – will get the OpenID list added to it. I’ll try that now, right here in this post. If it works, you’ll get a list of URLs below. Also thanks to Gerald Oskoboiny for discussions on this and reputation-related aggregation ideas; see his page on reputation and trust for lost more related ideas and sites. See also Peter Williams’ feedback on the foaf-dev list.

Next steps? I’d be happy to have a few more installations of this, to get some testbed data. Ideally from an overlapping community so the datasets are linked, though that’s not essential. Ed has a copy installed currently too. I’ll also update the scripts I use to manage the FOAF MediaWiki admin groups, to load data from RDFa blogs; mine and others if people volunteer relevant data. It would be great to have exports from other software too, eg. Drupal or MediaWiki.


Published by danbri

Click here to type

Join the Conversation


  1. @Robin, I had to approve your post first. Now it appears in the public site, and your URL is listed.

    What exactly is https://www.google.com/accounts/o8/id ?

    There have been anon OpenID services around – eg http://www.jkg.in/103-openid-a-spam-heaven/ – and I’m sure an “all or nothing” approach to deciding whether OpenIDs are good or bad is too simplistic. I expect we’ll end up with hybrid approaches and a lot of experimentation, as more datasets are exposed…

  2. (from danbri)

    Well, I just logged in using that Google URI. I can’t see what the full link is without submitting this comment, so … let’s see. I assume they append something, and the shared URI is just an OpenID 2.0 thing, like putting just yahoo.com to trigger discovery of your full individual OpenID. Let’s see.

    BTW what I see now: “Logged in as google.com/accounts/o8…. Log out?”

  3. Logged in, not listed..what I misunderstood?
    I’m interesting in test your F2F plugin.. I need others specification in addition to these ones on wiki pages.. could you provided it? thanks in advance!

  4. re “Logged in, not listed..what I misunderstood?”

    There is one more step: the list is a list of everyone who has had a comment accepted. I didn’t say it explicitly, but this involves a human (me!) reviewing all new comments. Once the comment is accepted, you show up. So you’re on the list now. The idea is that we can collect lists from lots of blogs, and use that to help people automate the comment approval process, so that spam is blocked but ‘known’ people are accepted…

Leave a comment