Mike’s Digital Laboratory

A conference ontology for a semantic wiki

Although I was not able to attend the The 7th International Semantic Web Conference ISWC 2008 this year, I felt I was there in spirit. For the past several weeks I have been helping the good people at Ontoprise put together a “Wiki on a stick” to give away at the conference. If you attended you may have picked one of these up. This is very cool, it’s a USB memory stick with a semantic wiki (SMW+) on it. When you plug it in, it launches a customized version of SMW+ with all the conference events and speaker information. I had a lot of fun building the ontology for this. I still can’t get over how fast I can develop a semantically enabled wiki with these tools. If you didn’t get one don’t despair, you can see the online version as longs as it remains up.

Also, it’s not too late to enter the The Ontoprise knowledge engineering contest and win an iPod. Check it out and while you are there you can explore the rest of the SMW Forum site and learn all about SMW+.

OpenCyc has just been released under the creative commons license.

Now it is even easier to use the rich and diverse collection of real-world concepts in OpenCyc to bring meaning to your semantic web applications! The full OpenCyc content is now available both as downloadable OWL ontologies as well as via semantic web endpoints (i.e., permanent URIs). These URIs return RDF representations of each Cyc concept as well as a human-readable version when accessed via a Web Browser.

Truly impressive, from what I’ve recently learned about this project it represents decades of work and now it’s available for everyone to share for free.  Kudos Cycorp!  I believe this is an extremely valuable gift to mankind. Here’s a quote from Z-Blog;  “OpenCyc is probably currently bast available and open ontology of the world.”

I think it will be interesting to see what people build with this and other massive linked open data stores. The first thing that comes to mind for me is how this fairly rigorously disciplined ontology, which was built by very serious professionals over a long time might be combined with the more informal but dynamically evolving ontologies of similar scale. For example something like as DBpedia. DBpedia reflects the informally developed, but possibly equally massive knowledge base harvested from Wikipedia.

One thing I learned while developing an internal corporate semantic wiki, was that the mix of informal and formal/dicisplined ontology development was of value almost from the start.  perhaps we’ll get to see someone repeat the same idea with these massive kowlege stores on a much bigger scale.  I can only guess that the folks at Freebase and other similar projects are already hard at work doing just that.

A good read: Paul W writes on automating CMS metadata

In a previous post I asked the question, “What if a web service could automatically provide the CMS metadata when you go to check-in a new topic?” In this post I’ll discuss why you would want to do that, some of the candidate technologies, and what is necessary to make it real.

The interesting back story to Paul’s new blog is this;  while Paul writes about CMS metadata I was on the phone with him yesterday brainstorming about how the heck we are going to get more metadata into his blog?  Being a new blog we have to pump up the Google juice as much as we can. While the best way of course is to keep writing and connecting with people it doesn’t hurt to add metadata to your blog.  Here’s the rub; If you run your own server, like I do, you have much more  control of your blogging software. I run Wordpress so I can experiment with the latest tools like Tagaroo or whatever comes along next that provides the Natural Language processing needed to suggest new tags that I can feed ito the semantic web. With Paul’s blog and others that use a hosted blogging service like Wordpress.com, Typepad or Blogger, you get only what the service provides you.  It’s not your car so you can’t add those hot rod components. So in this case something like Tagaroo won’t work.

So the conversation starts to diverge at this point. But I ask which CMS systems are open and extensible extensible enough to allow for the  next generation “beefed up” semantic gizmos to be bolted on as needed.  For example simple tagging out of the box is good but rich RDF metadata is even better. So as better NLP services come along I’d like to be able to swap out the old parts and put in better ones.  In my case I’ve already bolted on a plugin to generate SIOC metadata which gives me real RDF and FOAF on my site,  and every post is processed with NLP from Calais (Tagaroo) Not to mention a new version of Calias is now out that I can be immediately plugged into.  But I should be clear my point is that It’s not that these are better or worse than any other semantic solutions for my blog, it’s that I have the choice.

This leads to a broader and possibly more improtant conversation around open archtiectures and how they often come along with open source software. This why back a few years ago I switched from Movabale Type to Wordpress. The Wordpress plug in archtiecture is open and anyone with a good idea can get in on the game and we can all play.

My next experiment is to “plug in ” to the Semantic Web. Tonight I installed some new tools from sioc-project.org to enable a SIOC feed from my blog.

What is SIOC? The SIOC initiative (Semantically-Interlinked Online Communities) aims to enable the integration of online community information. SIOC provides a Semantic Web ontology for representing rich data from the Social Web in RDF.

First I installed Semantic Radar for Firefox. If you’d like to experiment with SIOC, and other RDF feeds, I’d recommend this tool. It’s easy to install and it automatically detects semantic data at any site you visit. Secondly if you leave the default options set every RDF feed you encounter will be sent to http://pingthesemanticweb.com/ where it is collected and stored in a larger repository. (One of many that make up the emerging semantic web)

Ok, cool so now I can detect RDF. Next I wanted to actually generate RDF from my web site with two goals in mind. One to generate an RDF feed representing my blog and all my posts. And two, to produce RDF (FOAF) feed representing the primary Author of the site, namely me.

To accomplish this I installed the SIOC exporter for Wordpress. This little tool is the product of those Deri.org genius minds Uldis Bojārs and Sergio Fernández. Installation is easy, I just dropped in couple of files in my Wordpress plugin directory and poof it’s done. Hey it works with no fiddling. If you are looking at my site now with Semantic Radar installed you should see some RDF feeds and should easily be able to browse them. If not you can try this link to the SIOC browser. Ok, so I said it works without fiddling. But of courae I had to fiddle. First of all if you haven’t noticed by now I’ve changed my site name. In reality I have actually restored it to what it was two years ago just before my blog fell into the great state of disrepair that lasted so long. (But we don’t talk about that) What’s important about that is that I wanted a “title” and “description” in the RDF based on the site name and tag line. I actually wanted the description to read “Mike Axelrod’s Weblog and not the title. Anyway, that did not quite work. Instead I found the title being repeted in the description field. Aha! It’s not a bug it’s an opportunity. So a code hacking we go. Here is the fix if you want to try this:

In the file sioc-include.php change line 605:

before:

$rdf .= “\n\t” . ‘<dc:description>Weblog: ‘ . htmlspecialchars(get_bloginfo(’name‘)) . ‘</dc:description>’;

After:

$rdf .= “\n\t” . ‘<dc:description>Weblog: ‘ . htmlspecialchars(get_bloginfo(’description‘)) . ‘</dc:description>’;

That’s as messy as it gets, and it’s an optional step to take. The only other messing about for me was to fine tune my semantic content. For example I discovered the “bio” section of my Wordpress user account is used to create the description field for the RDF that represents me as an individual (SIOC/FOAF data). After fixing that up by adding a little bit of a bio I felt satisfied.

Ok, looks good, it works, but what’s this all mean? What’s the big picture here? Well I believe it means that this data about me and my site is now harvestable and shareable in a standard format that plugs into any ontology. And of course it plugs into the great “Web ontology” that is the Semantic web, the web of “meaning”. Looks like it’s here and it’s real and I am now part of it.

Huzzah!

Update 8/13/08: Apparently I forgot to comment on this. The following day my ranking in Google shot up to the number 2 spot when searching on my own name. Prior to this experiment I had to go back several pages of Google results to find my own blog. So Apparently Google is paying close attention to RDF tags. I checked again today and my site is now the number one search result when searching on my name. Nice, SEO persistence. I like it.

So now I have a notebook bulging with new things to try. I think I’m in a state of semantic shock. I debated (yes, with myself) whether I should make one big list of stuff I learned from the conference and place it in one big blog post, or just tackle things one at a time. One at a time wins. So today I installed Tagaroo on this web site and will try it out on a number of posts including this one. The idea is that Tagaroo sends your post off to the tagaroo server, analyzes what you write and sends back suggested tags to use to add semantic metadata to your post. The semantic tagging is then available via RDF and the my site becomes part of the semantic web.

Although it’s all quite a blur my notes seem to indicate I caught Thomas Tague’s presentation on Calais home to many projects (including Tagaroo). Calais is backed by the publishing giant Thomson Reuters. What will they do with all this semantic data from the worlds bloggers who choose to use Tagaroo. Only good things they promise. The data will be kept confidential and they claim they are not selling it. Ok, I’m a trusting sort of fellow so here it goes.

Oh, did I mention Tagaroo will also offer suggested images from Flickr to add to your post. Here’s one I like: