We are happy to announce that the paper about the incremental knowledge compilation algorithm designed for and implemented in KnowWE has been accepted at the The Sixth International Conference on Knowledge Capture (KCAP2011). The paper entitled “Incremental Knowledge Compilation for Textbased Closed-world Authoring” has been selected for presentation as one of 19 papers out of 80 submissions. We are looking forward to present the topic and our System KnowWE in Banff, Canada at the end of june.
KnowWE goes KCAP2011
March 31st, 2011KESE7 in Tenerife
March 31st, 2011The call for papers is out for the next edition of the Workshop on “Knowledge Engineering and Software Engineering“. This time, we will leave germany and go to Tenerife. So take a look at the call for papers and consider to submit a paper. KESE workshops have been always very interactive, intriguing, and of course a lot of fun!
This year we also have a “Call for Tool Presentations”. Watch out, the deadline for submissions is May 6, 2011.
4th birthday of KnowWE
February 21st, 2011
On Feb 12, 2011, we celebrated the fourth birthday KnowWE!
KnowWE stands for “Knowledge Wiki Environment” and it is a Semantic Wiki specialized on the inclusion of strong problem-solving methods. With the following version control checkin the “official life” of KnowWE started (please note, that some work started in 2006):
build_20070212_1445 pkluegl * initial commit * KWiki became KnowWE-Webapp
We all are very excited about the current developments of KnowWE, that started as a “fun idea”: Back then, we wondered how to make sense of some text parsers for importing knowledge bases from Word and Excel documents. These parsers were originally developed for d3web and KnowME (Knowledge Modeling Environment, a rich client application) to automate the knowledge acquisition task. It seemed easy to include these parsers into a standard wiki, and thus move a part of the knowledge capture into a web-based environment. After all, we realized that the existence of a wiki brought even more “candy” than just “textual knowledge acquisition”, for instance distributed knowledge engineering, configurable environments, and finally the knowledge formalization continuum. Read the rest of this entry »
The d3web Kernel as a city
October 9th, 2010Recently, I came across a very interesting visualization method for the assessment of source code. Since I am currently working on the improvement and quality assurance of the d3web sources, I got interested…
The method is called CodeCity and was developed by Richard Wettel. The primary metaphor is the visualization of the source code as a “city” containing many different buildings.
Every building represents a class, whereas the size of the building is corresponding to the number of attributes; the height of the building visualizes the number of methods. The class LOC is displayed by a color: dark blue means a little LOC number, where lighter blue colors represent more LOCs.
The buildings are painted in different colors, when a particular disharmony is detected for the corresponding class. I was surprised that for the d3web kernel the application yielded some charming results.
Footnote: Please note that the kernel only contains the very basic classes and features of d3web.
Altogether, most of the provided smells (e.g., brain class, brain method, feature envy, shotgun surgery) couldn’t be found at all. However, the algorithm detected four “god classes” (painted in red):
- KnowledgeBase (tall red in foreground)
- KnowledgeBaseManagement (very slim and tall red in the middle)
- DefaultBlackboard and DefaultInterview (both red on the right)
Based on the tasks of these classes, I can life with the results. Probably, we should consider to rework the class KnowledgeBase class some day. Hmm…
Also, two “data classes” were detected by the system, i.e., Score and PSConfig. Those are considered to be used as data containers, so I am still relaxed. 
Next I was interested in the recent evolution of the kernel. Thus, I imported the kernel code of 10.2009 (one year ago) and found… more skyscrapers and more color.
From first sight, the city looks larger, more complex, and less harmonious. In the middle, the red skyscraper represents D3WebCase, which was partitioned into some classes (among others, into the DefaultBlackboard and DefaultInterview from above). The class KnowledgeBase (taler red building on the left) remained in the new code, but XCLModel was refactored into a distinct project. The large, “fat” building in the back of the image is the “Property” class that recently was refactored dramatically. Just one small side note for all “insiders” of d3web: The orange building on the right is the MQDialogController…
All in all it looks like that we refactored the kernel into a relaxed and small village with only some large buildings…
d3web/KnowWE release “Hancock”
September 8th, 2010
Right now, we just released the first official version of the d3web and KnowWE system. The preview release “Hancock” (a little ugly, but having a “good” core) contains the essential packages of the d3web core API and the core packages of the KnowWE system. d3web is a free, open-source platform for knowledge-based systems (expert systems). The following problem-solving methods are supported: Scoring rules, abstraction rules, set-covering knowledge, indication rules, and cost-benefit indication.
KnowWE is a Semantic Wiki based the open-source Wiki engine JSPWiki and enhanced by the core libraries of d3web and semantic extensions.
Some weeks of heavy testing and bug fixing have passed and we are really happy with the result. After this preview release, the Beta (Moneypenny) is scheduled for December 2010, and the first release is planned for Spring 2011.
d3web/KnowWE is developed by the University Würzburg (Informatik VI, Prof. Dr. Puppe, ROKT group) and the denkbares GmbH. The main developers are (alphabetically): Joachim Baumeister, Volker Belli, Johannes Dienst, Max Diez, Martina Freiberg, Markus Friedrich, Sebastian Furth, Reinhard Hatko, Fabian Haupt, Alexander Legler, Stefan Mark, Marc-Oliver Ochlast, Jochen Reutelshöfer, Albrecht Striffler, and Florian Ziegler.
The download link of the release will be published in the next few days.
Happy Knowledge Engineering!
Cropping and Numbering PDF Papers
March 6th, 2010When writing, reviewing, or proof-reading papers in Springer style, I am always annoyed mainly about two technical issues:
- The margins of the page are so large that a two-page printing reduces the size of the fonts extremely. Thus, I often print those papers not in the “greener” two-pages-on-one-side mode.
- By default, no page numbers are given on the sheet, so I frequently have problems to sort the pages during reading. Yes, I am aware of the concept of staples, but…
Recently, I found a solution, that is platform-independent, free, and flexible (and also usable for non-Mac users not having these fantastic PDF tools already installed… ): I now use the pdfpages package of LaTeX to include the to-be-printed paper in a newly created PDF that resizes and numbers the included pages.
The HermesWiki: KnowWE Paper accepted at ESWC
February 25th, 2010
We proudly announce that the KnowWE paper decribing des HermesWiki project is accepted at the In-Use track at ESWC 2010.
The paper entitled “Taking OWL to Athens — Semantic Web Technology takes Ancient Greek History to Students” describes KnowWE beeing used as elearning platform in the domain of ancient greek history. The project is a cooperation of the ROKT group and the Department of Ancient History of the University of Würzburg.
Accepted papers of the KESE Workshop
July 29th, 2009Hello readers,
just a short note: The accepted papers and tool presentations of the KESE workshop (Knowledge Engineering and Software Engineering) are now online at https://ai.ia.agh.edu.pl/wiki/kese:kese2009.
This year the KESE Workshop is co-located with the 32nd German Conference on Artificial Intelligence in Paderborn (Sep 15th, 2009).
We selected seven contributions from the submissions and we think, it is gonna be an interesting event. We will start the workshop at lunch time, so that there will be a chance to visit one of the KI tutorials in the morning. Further, we are planning to have a social meeting the evening just before the workshop. So if you want to attend, then simply drop me a note.
Best wishes, joba
Hello readers,
just a short note: The accepted papers and tool presentations of the KESE workshop (Knowledge Engineering and Software Engineering) are now online at https://ai.ia.agh.edu.pl/wiki/kese:kese2009.
This year the KESE Workshop is co-located with the 32nd German Conference on Artificial Intelligence in Paderborn (Sep 15th, 2009).
We selected seven contributions from the submissions and we think, it is gonna be an interesting event. We will start the workshop at lunch time, so that there will be a chance to visit one of the KI tutorials in the morning. Further, we are planning to have a social meeting the evening just before the workshop. So if you want to attend, then simply drop me a note.
Best wishes, joba
ROCKIT with VIKAMINE 1.9!
July 24th, 2009
Dear All,
due to requests from the industry 
– there is now also a blog entry in addition to the announcement on the usual forums and mailing lists.
So, according to our 2-month release cycle version 1.9 of the semantic data mining environment VIKAMINE has been released on 23rd of July, 2009. See the download page for the new version. This release includes many bugfixes, algorithmic improvements, and provides new exciting features for handling continuous target variables. Additionally, VIKAMINE can now directly be integrated with a data warehouse.
So, go and download VIKAMINE NOW at the project site: http://www.sourceforge.net/projects/vikamine
For more information, consult the web site at: http://vikamine.sourceforge.net
ROCKIT with VIKAMINE 1.9!
Martin
CfP: KESE09 – deadline extended
June 17th, 2009Hello,
due to multiple requests we decided to extend the KESE deadline for submissions to July 1st, 2009.
The 5th workshop on “Knowledge Engineering and Software Engineering” will take place at the 32nd German Conference on Artificial Intelligence in Paderborn this year (Sep 15th, 2009).
We solicit interesting submissions regarding the intersecting field between knowledge engineering and software engineering. Further, we organize a special issue in the International Journal of Knowledge Engineering and Data Mining (IJKEDM) of selected and revised papers. The official call for papers is attached.
Yours, joba