I have just had an interesting conversation with Peter Krieg, CEO of Pile Systems. For an evolving trial at a computer science oriented understanding and description see Ralf Westphals blog.
Krieg seems to be in on the technical basis for sth. that I have learnt to crave for deeply in writing my dissertation. The basis for a system that lets you work with relations between "elements" in such a manner that you can lay different filters over a pieces of knowledge (a set of "elements" which are made up by relations). One could pick out individual "elements" (which are not really elements or objects because the whole thing is constituted by relations alone) that serve as focus for organizing the rest of the "elements" in a if you want "hierarchical" fashion.
This view relates to personal construct theory that has been used in knowledge management tools, e.g.at U Calgary http://pages.cpsc.ucalgary.ca/~gaines/pcp/ -- for a BRINT overkill of info on personal construct theory see here: http://www.brint.com/PCT.htm
The technology of the PILE system would (among other things) allow to organize and quickly access pieces of knowledge such that the contents and differences of perspectives on a field of knowledge are visible. One could for instance look for all articles which relate to a certain subject (such as creative destruction) from the point of strategic management, switch to a prespective on creative destruction based on innvation studies and then go to the economics perspective on Schumpeter's famous concept. One visual interface for such a system has been implemented in TheBrain, and similar to some functions implemented by Beat's Biblionetz.PILE presently has no visual interface, so this is a nice project for some computer science students.
Projects that work in similar directions have been planned and/or implemented by Gunnar Sömmestad (Literary Machine), Ted Nelson (Xanadu and ZigZag), and derived from that GzigZag by Tuomas Lukka and Fenfire by Benja Fallenstein and coworkers based on RDFs. The problem with these is either some limitation in the functionality (e.g. LM programmer Gunnar Sommestad is somewhat reluctant towards network visualizations of the text elements in LM) or some technical limits such as limited scalability in (G)zigzag (according to Peter Krieg). Now PILE is behind these in terms of usability for specific tasks since at present it is not more than a platform. However, if the claims of the PILE team are true and the system lives up to their expectations they have something really interesting. To validate the claim a mathematical (or practical) proof is needed though.
Application areas probably include all areas where connections respectively patterns of connections between pieces of information are important or need to be figured out such as literature exploration and management for scientists, knowledge management in business and science, bioinformatics.
Some links - which I could not properly include on my Apple / Safari ;((
Fallenstein and Lukka's RDF based hypertext system http://fenfire.org/
Sommestad's Lietrary Machine http://www.sommestad.com/lm.htm
ZigZag implementation http://gzigzag.sourceforge.net/nutshell.html
Vannevar Bush's Article on Memex http://www.theatlantic.com/unbound/flashbks/computer/bushf.htm
No comments:
Post a Comment