Skip to content
benel edited this page Jun 25, 2012 · 27 revisions

The name "Porphyry" refers to one of the most ancient graphical knowledge models.

With Porphyry software, users can explore corpora through several concurrent categorisation systems.

Functional features

Used in qualitative document analysis, Porphyry assists the analyst in experimental reading and in comparing different analyses.

Technical features

Porphyry complies with the Hypertopic model but has a few specificities:

  • First, because topic networks in Porphyry are done by the users themselves rather than by computer scientists, we decided to keep the relations between topics simple and use only one kind of them: “topic-inclusion” (which forms a “directed acyclic graph”). Despite its simplicity, this relation type can be used for hyponymy (e.g. “philosopher is a kind of human”), meronymy (e.g. “head is a part of human”), instantiation (e.g. “Socrates is a philosopher”), and even for the most advanced patterns (associations, association links, association roles, etc.).
  • Secondly, because we are more interested in the “sense making” process than in its results, Porphyry traces every situation when a context between two information objects is tied or untied by someone.
  • Last but not least, because items must be shared by different author groups and even accessible to readers to follow authors’ interpretation trails, these items must be available in the system. Therefore, the entities handled by Porphyry are documentary fragments, sources (images and plain texts) and folders.

Genesis

Porphyry would not exist without the transdisciplinary seminar of ARTCADHi (dir. Andrea Iacovella). Researchers in history, historiography, archaeology, and art history came with their document corpora and their related research topics. The transdisciplinary study of corpora and research practices lead to the design of theoretical models.

First, archaeologists had been early adopters of databases and expert systems. They had experienced the limits of impersonal “data” and “facts”. The archaeologists in our team proposed instead the notion of “source”: from ancient texts, to artifacts photographs and even to researchers’ articles, all of those are sources (i.e. authored “discourses”).

Secondly, “viewpoints” was a way to bring intersubjectivity (interpretations conflicts), whose need in archaeological information systems had arisen for long without being addressed.

Thirdly, the evolving topic network built on top of document fragments was inspired by the semiotic analysis of excavations reports partially based on the “interpretative semantics”.

Lastly, the idea to consider historical knowledge as a process is the root of historiography. Indeed, through these discussions, the knowledge engineers of our team had to move from the “ontological approach” they followed before to a new one they called an “hermeneutical approach”.

Credits

Project Management: Aurélien Bénel

  • v0 (1998)
    • Mockup of a browsing system for digital libraries (Aurélien Bénel).
  • v1 (1999)
    • Algorithm for multidimensional browsing (Aurélien Bénel).
    • User interface for multidimensional browsing (Franck Eyraud).
  • v2 (2000)
    • Annotation system and document server (Mehdi Lababidi) .
    • Structure server (Thomas Buisson).
    • User interface enhanced (Aurélien Bénel).
  • v3 (2001-2002)
    • Distributed servers (Thomas Buisson).
    • First use of Web services (Rémi Huynh, Olivier Martin, Élodie Tasia, Rodolphe Vatré, Jocelyn Viallon)
    • User interface enhanced (Aurélien Bénel).
  • v4 (2003-2006)
    • Patent search on annotation and multidimensional browsing (Caroline Djambian).
    • State of the art in open-source licenses and community management (Michel Nux)
    • Refactoring (Guillaume Deshors, Julien Gossa, Baptiste Meurant)
    • User interface enhanced (Aurélien Bénel).
  • v5 (2006-2010)
    • Compliance with Hypertopic protocol v1 (Aurélien Bénel)
      • compatibility with Agorae/Argos,
      • compatibility with Cassandre.
    • User interface ergonomics and responsiveness (Aurélien Bénel).
    • Tag-cloud based visualization (Aurélien Bénel, Chao Zhou).
  • v6 (2010-2012)
    • Compliance with Hypertopic protocol v2 (Aurélien Bénel)
      • load performance,
      • syndication of different services (Argos, Cassandre, Steatite...).

References

Clone this wiki locally