wiki:PimOntology

Version 8 (modified by Leo Sauermann <leo.sauermann@…>, 18 years ago) (diff)

--

Naming of the Ontology

When we talk about personal mental models and personal ontologies, we use these words:

  • a personal mental model of person Paul: Paul's PIMO. A PIMO instance
  • personal mental models in general (the topic of personal mental models): Personal Information Models (PIMO)
  • the PIMO is defined using the: PIMO-language ?
  • the PIMO-ontology defines: much. all of this. see this tech report: todo.
  • the initial mental model is provided as part of the PIMO-ontology. It is the "PIMO-upper".
  • one entity in the mental model (be it document, concept, etc): Thing
  • one instance in the mental model. Project "AcmeSales" is a: Concept
  • documents or other information items that are stored in computers: Resource
  • the relation between a document and a concept: the concept occurs in the document. The document is an occurrence of the concept.

before, we had these names in discussion but now focus on above words:

  • PIMO: mywikitology, PIM-upper, TOPIC-PIM, SKOS, SGNOS, pim upper, pim mid, pim basic, topic, subject
  • why did you take PIMO-language and not PIMO-Ontology? Because people should understand that this is a language that persons can use to say their mental models to a computer. And PIMO-Ontology is hard to pronounce.

then replace them throughout this wiki

PIMO Language (short:PIMO-lang)

The PIMO-language is used to represent the personal mental model of the user. Places, people, times, projects or other subjects of interest are part of the mental model of a person, the PimOntology allows the user to enter these now into the system, when needed. We do not want to capture all mental activity, an unnecessary task, and instead focus on those concepts that are there anyway: today we find them in a tag, foldername, file-name, bookmark name, etc. The technique behind the PIM-Ontology is comparable to the Topic Maps standard or to the SKOS Ontology.

In gnowsis, the PimOntology consists of two parts:

  • The ontology language - a way to express the mental model and connect it to files, folders, web-sites
  • An basic initial mental model - a few pre-defined concepts so that you don't have to start from scratch

The PimOntology is used in several parts of gnowsis and the mental model expressed in it helps to integrate applications. The user's primary way to access the PimOntology is the WikitologyGui and the sidebar shown in [Miniquire]. For the programmer the PimOntology can be accessed directly as RDF graph or through an API. It is wiser to use the API, it handles more complicated tasks like connecting subjects to folders better. See WikitologyDeveloping.

The PIM ontology language was created in the EPOS project by the EPOS team, Ludger Van Elst was responsible for this task. The RDF representation of the ontology has to be integrated to the gnowsis project, and adapted for our use cases.

Initial Mental Model - PIM upper

The initial mental model was also created in the EPOS project. First a survey of existing desktop structures was made and then the initial ontology was expressed. It contains basic classes for:

  • Person
  • Place
  • Topic
  • Project

It will be integrated to the gnowsis project when needed.

Keeping the personal mental models up to date

After the gnowsis is installed, and the initial mental model is available, the mental model will evolve. First of all, we extract structures from native folder structures on your hard-disk and email folders, then these structures may be synchronised. See PimGenerationFromNative for the details. The process of keeping the mental model up-to-date leads also to so-called "suggested" classes. The synchronisation service may find things that are not 100% sure to be represented in the pimo. These are only "suggested" until the user confirms them.

The suggested classes are either kept in a seperate model or are reified or tagged somehow as only suggested. They could have a property "suggested" or put into a sub-graph called "suggested".