” We were given an imperative from our business leaders to provide a friendlier and more collaborative front end on our metadata…

… This in turn led us to MediaWiki and the rich and diverse array of add-ons developed by the Open Source community, particularly its semantic extensions.”

I was happy to play a small part in this project last summer.   I suppose I was more of kibitzer than anything else, but It was enjoyable to solve a few problems for the team as well as get an inside look into this very interesting project.

Looking ahead I’ve been thinking about how to offer more services around Semantic Wikis and what potential customers might want when engaging a consultant in this area.  It’s an area worth pondering because building applications with SWM changes the game of software development a bit.  For one things it’s a lot faster to develop SMW applications and it’s a bit easier. However it still requires good old fashioned thinking.  Also it requires combining skill sets from a variety of traditional areas in software development.  In a phrase I could say “traditional software engineering skills still apply”.

What do I mean by this?  Well here’s a quick  list off the top of my head of the skills I’ve been using and sometimes  redefining (or at least refining) as I build or help others build these semantic applications in a wiki environment.

• A good head for requirements analysis.
• Techniques for building semantic models to jump start the design.
• A standard approach to prototyping that establishes the basic framework of the application
• Inclusion of traditional web site design techniques to cover all the bases of building a usable application.
• And a complete development methodology that is agile and includes the customer on a regular basis.

Does this approach work?  So far it seems to be. The only other thing I think I can add to this at the moment is that when developing Semantic MediaWiki applications there seems to be a recurring need for what I can only call “applied cleverness”.  What do I mean by this? Well it’s hard to say without being specific, but lets just say that the rising tide of new web 2.0/3.0 tools and techniques that are available today are continually inspiring me to combine these tools in new ways that (I believe) have never been done before.  There is a tremendous amount of “invention” going on today in this industry.  And it’s a heck of a lot of fun to be a part of it.  But now I’m wandering into a new topic so I’ll stop here and save this for another post yet to come.

“Project Wonderland is a 100% Java and open source toolkit for creating collaborative 3D virtual worlds. Within those worlds, users can communicate with high-fidelity, immersive audio, share live desktop applications and documents and conduct real business. Wonderland is completely extensible; developers and graphic artists can extend its functionality to create entire new worlds and new features in existing worlds.”

I just spent the last half hour watching a set of video’s that provide a technical overview for Project Wonderland. I must say I’m impressed. In it’s current state (version 0.4) it lo0ks fully functional and ready to support real time 3d  avatar based interaction complete with a feature rich audio bridge system.  The back end is provided by the open source MMO game engine Project Darkstar.

The system looks good enough today to use for small group collaborative applications and the future 0.5 version promises enhanced scalability and better graphics and avatar support.

I think the most interesting outcome for me was the realization that the global tweet stream might actually have some hidden potential value for the semantic web.  what I once thought was a fairly useless communal stream of consciousness, the endless stream of twitter messages about silly things and foolishness, might actually contain hidden semantic gems.

It turns out there is a common subset of twitter messages that follow the same pattern of semantic metadata.  For example many messages are commonly in the form of;  “Mike Axelrod is going to the Rochester Museum and Science center today http://www.rmsc.org/“  This twitter message actually contains some fairly decent semantic data.  It’s in a nice concise triple form (subject, predicate object) it’s just not quite consumable by machines, yet (that is to say, not in the RDF, OWL)  Hint, Hint  folks, we could have something interesting here.  As it stands if we just filter for all the tweets that have URLS we might have some interesting semantic data to mine.  Tweets are so short that if they do contain an url typically it will be just one.  This guarantees us a semantic triple every time. The sender, the comment and the url.

Ok on to the second part of this post. So how do I actually fit this whole tweeting, twittering foolishness into my lifestyle?  Well here’s what I came up with so far.  First of all  I only want one point of entry for my tweets.  One client, one ping to rule them all. So naturally I’m gravitating to ping.fm.  This tool allows me to route a single status update  (tweet) to all the social networks I use.  This would be LinkedIn, Facebook, and (it seems) twitter now.   Next I want a better client experince than what a web page can offer.  It’s just not conducive to the way I work to have to navigate to an url when I want to post an update. additionally some of the clients have a very concise format and play well on my desktop.   And finally I want control over where tweets go.  Sometimes to linkedIn, sometimes to Facebook, sometimes to both and sometimes to my WordPress blog.  So ping.fm gives me this fine grained behavior with one exception. To route tweets to my wordpress blog I had to install the ping.fm wordpress plugin, which I’m happy to say works like a charm. (<–look over there in my sidebar and you’ll see it.)

Now there is only one missing piece of the puzzle. Of all the front ends to ping.fm I tried. None of them had all the features I wanted.  Twitterific is wonderfully simple but only works with Twitter, Tweetdeck works well but only gives me the choice to send to Twitter or Facebook.  Twhirl is getting much closer to what I need and has almost everything except one critical feature. The ability to easily pick which ping.fm group I wan to send to. The last one I tried today does this! It’s called MePing but alas it does almost nothing else, compared to the others I tried. (It’s very beta) Well I’m not giving up and I’m sure very quickly one of these clients will fit my full set of requests. I have the feeling we are in a horse race right now and if you are reading this post 6 months from now many of the above mentioned clients will have everything you need (or will have dropped out of the market.)

There is plenty more to explore with twitter intrgration to websites. I certainly won’t be spending a great deal of time looking at all of this but I do believe there may be some value to discover, if applied correctly.

…a community platform supporting professional social networks across local government and the public sector. It provides a secure environment for knowledge development and sharing through online communities of practice.

I recently was invited to join a Scottish group on “shared representation using semantic technologies”.  I noticed this site has recently won the 2008 eGovernment national award in the UK. And rightly so, it is a very good example of a site for people who care about people.  This collaborative tool features forums, library (document management), wiki, news feeds (RSS) and user blogs.  All entries are tagged so one can navigate and search more effectively.  A member can start up a  new community of practice (CoP). To do this the site provides a “create new CoP wizard” that asks the creator to agree a number of  principles and responsibilities for positive, effective collaboration, although a usability hurdle I can see how  it really would make one think about what they were about to start up. Apparently the site is a success as there seems to be  a very large number of groups collaborating on it.  I think it’s a good benchmark for social software to support communities of practice.

An ontology for rice.

Agropedia is a social software site for agriculture. A first glance reveals blogs, forums, wiki and semantic web enabled repositories for knowledge about potatoes and peas!

Form the home page:

agropedia is an agriculture knowledge repository of universal meta models and localized content for a variety of users with appropriate interfaces built in collaborative mode in multiple languages. agropedia aims to develop a comprehensive digital content framework, platform, and tools in support of agricultural extension and outreach. In other words, it aspires to be a one stop shop for any information, pedagogic or practical knowledge related to extension services in Indian agriculture – an audiovisual encyclopedia, to enchant, educate and transform the process of digital content creation and organization completely.

Hmmm…. Well I like the part about “enchanting”. However it’s still very rough around the edges. I also noticed they are using CMap in a collaborative mode to create the ontologies. Browsing the graphic representations of knowledge map for rice reveals the difficult challenge of displaying large ontology models. It’s just difficult to navigate, but it does work. Of course my 24 inch iMac display makes it easier, I’m sure it’s a significant difference on an average display. On the other hand large displays are becoming more and more common. One of our students the other day showed up in the lab with his 24 inch iMac in tow as his “portable computer”. I believe he had one of these (or something like it)

Problem-o-pedia

One thing I just noticed is that apparently there are more than one “agropedia”s on net.  This may be a sign of the possibly worn out naming convention of fillintheblankpedia as a way to name a site.  Having used this convention myselff (to name “Excellupedia”) I can testify that we had some qualms about it, but went ahead with it anyway.  An interesting argument against this naming convention is that these sites can become (are) so much more than encylopedias about this or that…  It’s possible that with sites, like this one, we are building a new sort of thing that we don’t really have a good generic name for yet. (socialsoftwarewikiizedforumblogthingamijig  just wont do.) Or are we struggling with the possibility that we may be redefining what an encylopedia is.  Are we?  perhaps.