This document contains an use cases for a generic Trusting System. For instance, it describe a way to assign a reputation to an Accessibility Analizer in a distributed analisys system.
Accessibility is a discipline that studies how to grant access to a web resource for all the possibile users. For each page, web site, or generally web resource on the network, accessibility can be measured by the conformance to the Web Content Accessibility Guidelines. Each guideline consists of a set of checkpoint, that can be satisfied, not satisfied or Not Available when the subject of the checkpoint is not present in the analized resource. The analisys should be stored using che EARL format, based on RDF/XML.
Each person could make an analisys of a web resources, so a generic accessibility analisys could rate a resource in a good way or not. It could be correct or not. We are looking for a way to determine if an analisys is correct or not, without checks it point by point. A reasonable way to do that is to give a reputation to each analisys author.
So we could rate an author checking some of its analisys, to build a direct reputation graph. So, an author A can check the single checkpoint c of the analisys done by B on the web resource r. If A agree with this, his trust in B about the checkpoint c increases. If A disagree, his trust in B about c decreases.
direct trust of A in B about c = number of analisys such that A agrees B on c / number of analisys of B checked by A