So why is this a success then? Well, many complaints means that there has been a lot of activity in the source code, and this means that many people started committing their additions and changes. And this was actually the stated main goal of the 1st KiWi Programming Camp: help the KiWis and friends learn how to work with the code so that they can do their own modifications, particularly for implementing the enabling technologies and the use cases.

But the KiWi Programming Camp was also very successful in its second objective: starting to build up an Open Source community around KiWi. Although we sent out the invitation on very short notice (4 days before the camp started), we were joined by two additional external developers during the week who wanted to learn how to extend the system for their own purposes: Christoph Lange from Jacobs University of Bremen builds a Semantic Wiki for mathematical knowledge called “SWiM” and Michael Schneider from ArtAround, a small company in Salzburg, wants to build a community platform for art on top of KiWi.

Besides this, the programming camp was also a great place for spinning new ideas and moving the project forward. Since I did not want to disturb the social dynamics between the young researchers through my presence, I only joined from time to time, and every time I entered the room there have been lively discussions, people arguing at the flip chart, people programming together. In fact, I think we did not have many weeks in the project where we did such big conceptual steps forward like this week.

Among other issues, we managed to move forward the issue of annotating text fragments, the issue of semantic forms which are linked to the textual content, the recommendation functionalities, as well as compound content items. We also had quite some interesting discussions on search and querying, on reasoning, and on information extraction.

So, thanks for all who attended (i.e. Fred, Karsten, Marek, Klara, Jakub, Alex, Christoph, Michael) and particularly to my team in Salzburg (Rolf, Steffi, Szaby, Mihai, Thomas) for preparing it and helping the others get into the system. I am really looking forward to the 2nd KiWi Programming Camp, which will probably take place in autumn and be open to everyone interested to participate!

Now, we should probably try to get to “Hudson build is stable again”, so that we can prepare for the next prerelease end of March.

Wednesday morning started with me presenting a look back (see also this blogpost) and a look forward into the next year of KiWi. Just to briefly summarise what we are up to in the coming months: a KiWi 1.0 release with the core system finished, then start of the implementation of reasoning & querying, reason maintenance, information extraction, and personalisation, as well as the two use case applications. As part of the implementation of the use case at Sun, we are also going to integrate Sun’s Community Equity into the system. In parallel, as part of our dissemination strategy, we now start building up a Open Source Community around KiWi (see also the post about the first KiWi programming camp), because we believe that KiWi as a platform can be useful for many Open Source developers who want to build social software platforms.

After this introduction, I gave a demonstration of the current KiWi system as we released it two weeks ago. If you want to try this out without downloading, simply visit the Showcase. Szaby then followed with a presentation of the rationale behind the KiWi user interface, summarising a meeting we had beginning of February. We had quite some discussions on some of the aspects, particularly the “stream of activities”, which we consider a major part of any social software system. Rolf introduced into the KiWi architecture and extension mechanism, and Steffi concluded the morning session with a presentation of the transaction and revision system she is currently working on, one of the core components of the KiWi system.

Wednesday afternoon was dedicated to the enabling technologies. Jakub started with presenting his ideas for a rule-based reasoning language based on tags, and he showed us some of the challenges arising in social media (e.g. concerning negation). Jakub was also concerned with how reason maintenance can help to explain to the user why certain things happen (like the system being inconsistent). Klara followed with a presentation of the label-keyword search functionality she envisions. Label-keyword search will allow to query content not only based on text but also on structure and meta-data. Pavel and Marek then showed how information extraction using an extension of latent semantic analysis could help to automatically annotate content and suggest tags. And Peter D. and Fred discussed various algorithms and their exemplary implementation that can be used for recommending related content items to the user based on tags and context as part of their work on personalisation. Eventually, you will find this and more in the planned Dashboard application. We concluded the day with a dinner in a nearby restaurant where we could introduce our guests into the Austrian kitchen (Griesnockerlsuppe, …).

Thursday morning started with Julia presenting the KiWi Dissemination Plan and strategy for the next year. This inspired quite some discussion on target groups and relevant conferences so that the session took 90 minutes instead of the planned 45 and I had to move my presentation of the Open Source strategy (which is part of Dissemination) to Friday. The second part of the morning was on financial issues, a boring but important topic for the project.

Thursday afternoon was time for the two use cases: Peter Reiser and Josef described to the broader audience the refocussing of the Sun use case to addressing the whole Sun Intranet (SunSpace) instead of only the Netbeans community. Of major importance will be the integration with Sun’s Community Equity algorithm for “measuring” the social value of one’s participation – could be a really cool feature even outside the use case. While this all seems like a big change, we realised that it was only a refocussing regarding the content and functionality we had planned from the very beginning.

Daniel, Fred, Peter Dolog, and Peter Axel Nielsen (lots of Peters in the project…) then described advances in the Logica use case. From that session (and the session on Friday morning) it became a lot more clear to us what is actually needed at Logica: a tool for moving from informal content (project descriptions) to formal content (ERP system and QM system). I think that we might find here many interesting complementary issues to the Sun use case, most importantly the realisation of semantic forms as alternate views to content items and the connection of form fields with human-readable content marked up with annotations.

Thursday evening, we had our social event with a tour through Salzburg’s largest brewery (Stiegl), followed by a nice dinner at the brewery’s restaurant. I think that all participants really enjoyed this dinner after the intensive working day.

Friday morning, it was again time for me to present: first, I outlined the KiWi release roadmap and how we envision that the development of the KiWi system continues in the next months, particularly with respect to the integration of the enabling technologies. Afterwards, I briefly introduced into the KiWi Open Source strategy: the main message is that you do not build up a community by simply publishing the source code. Rather, it requires a lot of effort on our side. Main measures during the project are the KiWi Programming Camps (the first one is taking place this week), the KiWi Challenge, and the KiWi Practical Courses at universities.

In the second half of the morning session we had group work where we dicussed in four groups how the final KiWi system should look like, particularly regarding the two use cases and the enabling technologies. I think that most of us now have a much clearer picture of what is expected. Overall, we decided that we should have more such sessions in coming meetings.

Finally, Friday afternoon was dedicated to the General Assembly. The main issues to discuss were the refocussing of the use case at Sun and the planning of the review meeting in June. We concluded at around 4:30 in the afternoon, and are now looking forward to another busy week at the 1st KiWi Programming Camp!

Overall, a very successful meeting where we could really move forward. The best thing, however, about the meeting was definately the organisation and location (thanks, Julia!). In such an inspiring surrounding it was really easy to come together and exchange ideas, and get motivated for the year to come.