Wednesday, June 05, 2002

PeterV's XFML meta data project is approaching 1.0 with a recent expression in RDF (via XSL). Lots of intriguing potentials: Practical experimentation in existing websites could bring some of the semantic web ideas down to Earth. Mozilla is very good at working with RDF, despite a the warts, and it could integrate the act of browsing with tagging. XFML is meant to be a shared category/ontology and  instance format. The category names are also distributed. This interlinking of "instanceOf" declarations is an important for reasons quite larger than the stated intentions of the project. The proposed uses cases include categorical syndication and the ability to have multiple taxonomies for a content set -- enabling faceted browsing UIs. 9:26:36 PM    comment

HP does an excellent Kazaa Usability Study [PDF].  Lots of good points: "[Kazaa] presumes that users have perfect knowledge of what kinds of files that are contained in those folders": Know your user. Expection extraction during a usability test and focused interviewing can extract this type of detail.  Recursion is difficult to convey and hierarchies are empirically shown to be a cognitive challenge.  The tree map approach can really nail this issue. Most of the other issues fall into the heuristic "keep the user informed of system status". System status, in this case, relates directly to what is being fetched and what's being offered for fetching. The study is also worthy of note for another aspect. It intermingles evaluation techniques very effectively, diagnosing the problem in expert review, proving the existence of the problem by sampling the existing users configuration, and exploring it in depth with traditional testing. 6:52:37 PM    comment