RObjects

1
R-Objects

• By
• Cartic Ramakrishnan
• LSDIS-UGA

2
R-Objects PepperTM

• PepperTM is a software the supports individual
  and (potentially distributed ) group decision
  making.
• It is constructed using peer-to-peer network
  technology and broadly extensible semantic
  information model.

3
PepperTM architecture overview
4
Features

• When PepperTM is run for the first time
  http//www.r-objects.com is accessed and through
  Java Web Start pepper is down loaded and
  installed on the users machine.
• PepperTM has a configurable, semantic, meta-model.

5
Distributed data management in PepperTM

• Data managed by PepperTM can come from several
  sources
• Native PepperTM data created through its user
  interface.
• The user can also drag and drop files into
  PepperTM which is then parsed using a library of
  type specific parsers (called enablers).

6
Distributed data management in PepperTM

• Data brought from outside the PepperTM system has
  somewhat unusual characteristics
• The original file is tracked.
• Fine-grained (paragraph level) event model will
  notice the change in the original document and
  forward the change to all affected user interface
  components and to subscribers of its content on
  the p2p network.

7
Distributed data management in PepperTM

• Pepper does not just shuffle around opaque data
  blocks. The particular enabler (type specific
  parser) breaks the imported file into individual
  chunks of information (PeppercornsTM) and
  semantic relationships between these
  PeppercornsTM.
• Imagine theses PeppercornsTM and their
  relationship similar to a filess DOM (Document
  Object Model).

8
Distributed data management in PepperTM

• PeppercornsTM are file/format independent data.
• All PeppercornsTM together for a uniform semantic
  object network.
• Two PeppercornsTM can be related only if the
  meta-model allows for them to be related.

9
Distributed data management in PepperTM

• Unlike basic XML, Peppers meta-model is not
  limited to tree structures but is a specially
  developed variation of a powerful, extended
  entity(object)-relationship(association)-attribute
  meta-modeling technique.
• Supports
• Typed (meta-)entities
• Typed (meta-)relations
• Single and multiple inheritance
• Inheritance of meta-relationships
• Multiplicity constraints for meta-relationships
  etc.

10
Distributed data management in PepperTM

• Peppercorns and relationships are stored in an
  object repository with advanced object oriented
  p2p data replication features based on a JXTA
  transport.

11
Viewlets

• Viewlets are user interface components. Pepper is
  driven by a directory of such UI components. The
  set of viewlets is an open-ended list.
• Viewlets announce what data type they can work on
  and as a result when a user selects a Peppercorn
  in the Pepper GUI the UI dynamically configures
  itself based on appropriateness of certain
  interface components for the selected object.

12
What does Pepper do?

• The semantic meta-model is the component of the
  system that reflect the usage scenario.
• Semantic meta-model in Pepper can be explained
  using IBIS meta-model as an example.
• IBIS is a structured discussion, decision making,
  and facilitation technique originally develops in
  the 1970s.

13
What does Pepper do?

• Pepper provides a collaborative and decentralized
  platform to enable individuals and/or
  organizations to use the structured method of
  IBIS to support decision making.
• The core concepts of IBIS are
• Question
• Idea/proposal far an answer
• Argument in favor
• Argument against

14
IBIS Meta-model in Pepper
15
Usage scenario
16
Usage scenario