Reflection: IKS Project Requirements Workshop

The IKS workshop was surprisingly great! Salzburg Research and the IKS team managed it to bring complete different CMS free- and OpenSource-software communities together.

Raphael Ritz and Jens Klein (me) represented Plone, Zope and Python in the discussion. Raphael did this primary from a scientific point of view while I brought in needs of a small to medium enterprises as founding member of BlueDynamics Alliance and managing director of Klein & Partner KEG.

We had a bunch of JAVA CMS members like OpenCMS, Jaiha and more, a bunch of PHP folks, like Midgard, Drupal, fCMS and others. We had people from all over Europe, India and Canada there. This were the 18 people from open source communities (community members). The participating projects core-members are from 7 research partners and 6 industrial partners. At all we have been about 60 persons. We started at thursday with participants self-introductions. Unfortunatly only community-members introduced themselfs. Then we head over to dinner at Krimplstätter: Tafelspitz, home brewn beer and 60 semantic people - it was an interesting evening with lots of valuable talking.

IKS project is a Integrated Project of the European Union funded for 4 years, initiated and coordinated by Salzburg Research. Integrated Projects are the biggest possible projects in the field of european ICT projects. IKS means Interactive Knowledge Stack. While this means all and nothing lets explain a bit from my understanding what it is all about.

IKS deals with idea to empower small and medium european enterprises in the CMS market to integrate, use and sell semantic web technologies. As all technology is free and opensource this is for sure available to non-europeans too. Most important three goals the eu commission want to achieve is: impact, impact and impact.

But: What is semantic web? First its a well known buzzword for more than 10 years now. Its also known as an academic, non-real-world approach to deal with content. On the other hand we all know: Our current way to deal with information is not enough. In several of our projects we added additional fields containing keywords from controlled vocabularies or onthologies. We used those additional information with Plone-collections (stored dynamics searches) to offer the consumer of the information dynamic horizontal views and the content-soup - and at all it worked fine. But we don't have it generic. The same procedure like every year? Yes the same procedure like every year.

Having a more generic way to give the content a meaning is indeed often a customer need. But customer - better say customers content-creators - do not like to edit additional information! Here is the conflict and main reason why semantic web isn't used.

At Friday we started with brainstorming: our requirements, use-cases, user-stories and ideas were asked for. Several ideas and requirements came up, most of them related to semantic web. But even for us Plone-people trivial (and off-topic) things like security, transaction-wareness and process support came up. Here you can see the power of Plone in several fields - we may need to communicate this better (even outside any semantic context). After the brainstorming session we had 5 pinboards full of cards. One major point I brought up (I think I dropped in about 5) was to make the IKS software stack architecture independent from its implemenation and offer everything as services (vs. libraries). Its needed by the heterogenos technology mix out there in the real world. All outcomes of the workshop will be documented and published soon. I'll provided the link as an new post here commenting them from my point of view.

We identified some fields where action can be done very soon. One of them is a semantic search engine based on Lucene/SOLR. Bertrand from the industry partners already started working on this. Since Plone can easily be integrated into SOLR its also easy to do this for a semantic SOLR. Then the documents need to be enriched with semantic metadata. At least the data we have can be exposed. This is goal of the Google Summer of Code project Raphael mentors and Matthew works on. If all is combined we be able put Plone Content into Semantic SOLR and query it on a semantic layer. If Semantic SOLR supports SPARQL, we can use the python-sparql client in Plone to ask complex questions on a semantic layer for Plone Content. We're not so far away from it.

Another short-term goal is a semantic wysiwig editor. Current editors like KUPU do not enrich the markup with semantics but with formatting, like classes and tags. A semantic editor would do this an the base of well-defined microformats or RDFa.

I'am looking forward to test the outcomes of the short-term goals as soon as possible.