Hi Andreas et al, Let's not just fix the WebID feature, but also use it for greater purposes in Stanbol :) Several potential use-cases of WebID integration in Stanbol were formulated in this mail-thread with the help of Rupert, Fabian etc.
Some interesting usecases discussed in this mail thread are as below. I think one or more items can be done in addition to fixing the WebID feature in Clerezza as the GSoC project. Would like to get suggestions from everyone as much as possible and prepare the proposal including your ideas. 1. WebID of the user can be used by the enhancement engines to provide more relevant and contextual enhancement results to the user. 2. WebID based authentication and userstore :Why doesn't Stanbol have a user-management module at the moment ? I think a user-store is a valid requirement for any CMS. 3. Entity Disambiguation for Persons using WebID : A WebID has a lot of linked-data which provides contextual information about the person which could highly useful for Entity disambiguation in the enhancement phase. 4. Knowledge population. Related to #3. WebID can be used for entity-disambiguation and also to introduce new entities to the knowledge base formulated from the WebID 5. Social Profile : WebID can be used as a portable Profile across multiple SNSs and the data collected can be used to store personalized content in the CMS. 6. A Federated to Sign-On using WebID : not directly relevant to Stanbol , but a valid use case in general Thoughts are welcome! Thanks, Dileepa On Thu, Apr 25, 2013 at 5:23 PM, Andreas Kuckartz <[email protected]>wrote: > Hi Dileepa and others, > > in general I do think that this can be part of a GSoC project. > > A project involving both, the Apache Stanbol and Apache Clerezza > communities, should not be an unsolvable problem as those are > overlapping. I also would not be against someone from Apache Clerezza > mentoring this alone if that is necessary for any reasons (especially > formal ones). > > But I fear that "only" fixing a feature which already worked in the past > would not be enough to convince the evaluators of the proposal - > independent of the amount of work involved (and I have no idea how much > that would be). > > What can be done realistically in addition to fixing this feature in > Clerezza? > > Can you describe what WebID integration in Stanbol would look like so > that proposal evaluators can easily understand it? (I might have missed > that in the other mails) > > Cheers, > Andreas > --- > > Dileepa Jayakody: > > Thanks Rupert, Reto for your ideas.. > > > > I went through Apache Clerazza project by checking out the source from > the > > git repo at [1].I think the webID integration code that Reto refers to is > > at [2]? Please correct me if I'm wrong. How about we make it working > again? > > Regarding the ssl http problems maybe we can ask Eclipse Communication > > Framework [3](ECF) guys to collaborate. > > > > As you have listed down revoking the WebID in Clerazza and in-effect > > Stanbol can bring so many benefits to it as a semantic access-control and > > authentication framework. WebID integration can be very useful in the > > entity disambiguation as well. > > > > I would love to know if this can become a possible GSoC project. > > > > Thanks, > > Dileepa > > > > [1] https://github.com/apache/clerezza > > [2] > https://github.com/apache/clerezza/tree/trunk/platform.security.foafssl > > [3] http://wiki.eclipse.org/ECF > > > > > > > > On Thu, Apr 25, 2013 at 2:05 PM, Rupert Westenthaler < > > [email protected]> wrote: > > > >>> @Rupert: With regard to your opinion on using EntityHub as an index > >> rather > >>> than the primary storage for the knowledge base, can you please > elaborate > >>> more on why do you think so? Maybe Wiki-Link or Freebase [3] is a good > >>> option for such a knowledge base? > >> > >> The Entityhub is designed to play the role of an index & local cache. > >> It was not designed to manage knowledge. Because of that I would > >> recommend to use the knowledge base that does best fit your data > >> (Triple Store, Graph Data Base, SQL, NoSQL ...) and then use the CRUD > >> interface of the Entityhub to create ManagedSites with those data > >> needed for EntiyLinking and Similarity queries. > >> > >> Note that Enhancement Engines can also directly interact with the > >> knowledge base (e.g. if the want to modify the knowledge base such as > >> adding a mention example or creating an unknown Entity). However the > >> Entityhub ensures really fast lookups as required for EntityLinking. > >> > >> best > >> Rupert > >> > >> -- > >> | Rupert Westenthaler [email protected] > >> | Bodenlehenstraße 11 ++43-699-11108907 > >> | A-5500 Bischofshofen > >> > > >
