Pundit hosted his first Hackday in Pisa.

Two days where hackers joined their forces to envision and realize new and interesting developments for Pundit.

Pundit User Experience

Christian Morbidoni, Christian Stein, Friedrich Schmidgall, Alessio Piccioli, Giulio Andreini, Michele Barbera, Chiara Aiola

The main problem addressed: users need a quick and understandable way of creating a bit number of annotation with the same “pattern”. E.g. Assigning text fragments to a specific catogory, asserting who is the creator of a text.

Solution investigated: Annotation templates >> preconfigured annotation facilities that allows users to produe semantically reach annotations (in the form of one or multiple RDF triples) by simply selecting a page fragment.

Pundit selectors and Dandelion.eu

Martino Pizzol, Simone Fonda

Martino created two brand new selectors for two dandelion.eu datasets: the Geo and POI (points of interest).

The two selectors lets you use geographical entities (such as cities, regions, municipalities and more) and points of interest (like the leaning tower of Pisa) in a full Linked Open Data fashion!

The EuroGEO is an easy-to-use API with administrative & geo relations of countries, cities & regions in Europe from trusted, official sources. The POI holds tens of thousand of places in Italy from OpenStreetmap, Wikipedia, Geonames and our partners.

Eager to know more about them? Check dandelion.eu out!

Vocabulary live editing

Romeo Zitarosa, Riccardo Tasso, Stefano Baldocchi, Tommaso Agnoloni

The group started off by extending korbo API to make it able to import SKOS taxonomies. A custom hack permits anyone to translate such taxonomies in korbo-ready data.

Then they moved towards a dojo 1.9 implementation of a simple tree visualizer, using dojo’s built-in store, memory, ObjectStoreModule etc.

Chrome and Firefox bookmarklet extension

Michele Nucci, Marco Grassi, Susanna Martinelli

The team developed two add-ons for Chrome and Firefox, to easily work with Pundit with multiple web pages in multiple browser tabs.

The first working version is hosted at GitHub: https://github.com/marcograssi/PunditBookmarklet

The Chrome extension allow users to:

  • load Pundit on a web page and have Pundit loaded automatically from that moment on into any tab you open

  • configure the options of the extension to choose what version of Pundit you want to use

  • Right clicking on any image in a web page opens the image in a separate annotation environment, built on Feed.ThePund.it. This works even if Pundit is not opened.

The Firefox plugin is not yet fully working.

ask.thePund.it desiderata

Christian Stein, Simone Fonda, Martino Pizzol, Christian Morbidoni

Ask the pundit should provide multiple ways of visualizing a notebook:

  1. Annotation centric: knowledge (triples) is grouped in containers representing the annotation event (a user creates a set of triples on a given page at a orecise time).
  2. Entity centric: knowledge is grouped with respect to the resource(s) involved in triples.
    • A first naive grouping strategy:
    • All the triples with resource A as subject, predicate or object goes into group G(A).
    • Groups are sorted from the most cited resources (e.g. the number of times the resource appears as subj/pre/obj in a notebok)
  3. Other:
    • Generic RDF browsers as alternative viweing modes?  (e.g. LodLive). Should be customized. Relations shown should be filtered based on user preferences.
    • Vertical visualizations (e.g. Edgemaps) directly accessible from aks (add-ons). Vertical vis works on specific kinds of annotations, usually following the same triple pattern.

In principle the view modes can be extended to multiple notebooks or in general to a set of RDF annotations. Future: study this better (e.g. scalability, social/sharing models).

The idea is to provide a concise view of  each annotation in the annotation centric mode and of each entity in the entity centric mode. clicking “more” a detailed view is shown including all the triples.

Multiple triple sets (annotations or entities) are arranged in multiple (e.g. 3) columns. Multiple sets can be expanded t the same time so they can be compared. With this respect a useful feature could be the possibility of draggin and dropping triple sets to move them into a grid.