ISWC Day 1
by Taowei Wang
Sat in the Semantic Web Policy Workshop today, where Vlad had a presentation. There are a couple of papers speaking of constrained natural language as a device to allow (computer savvy) people to author policies. One paper claimed that prior to this GUI/CNL approach, no one other than the policy buffs can write a policy (policy language barrier, domain-specific knowledge barrier). However, with CNL approach, typical computer science students can write/specify policies. They placed the verification of the correctness of the policy on the users: let users look at the NL generation of the policy to see if they are correct. However, even with just boolean combinations on constraints of the policies, NL is very misleading. Therefore, relying on human verification using NL is chancy at best.
Of course, this is where Vlad’s work comes in — a framework for automatic policy verification at policy-creation time. Of course, people have other ideas.
The rule-based people want to increase expressivity (more so than OWL-DL) on rule-based policy engines so they can easily express more ‘reactional, behavioral’ types of policies. Vlad argued that this makes the already messy rules system even more to analyze. Some snapshots: Vlad talking + Vlad holding his ground on the minipanel
Met lots of former Mindswappers on site: Aditya, Bernardo, Bijan… good fun’s had by all.
In the late afternoon, I attended the Scalable Semantic Web Knowledge Base Systems. There were lots of bench marking opinions — how to choose reasoners for applications, etc.. People are getting increasingly interested in incremental reasoning, and it’s been mentioned explicitly in one of the talks.
There are also lots of arguments about what to benchmark, what to measure, and how to foster a community-driven benchmark building process. Very interesting.
Tomorrow: SWUI
Dave
