1 Introduction
The original idea of Logical Information Systems (LIS) comes from
Olivier Ridoux, who was unsatisfied by hierarchical file systems in
particular, and hierarchical data organizations in general. He was not
satisfied either by databases because they lack flexibility in the
description of objects, integration with the operating system and
other applications, and navigation capability to help non-expert
users. His basic idea was to combine the expressivity of querying, by
the use of logics, and the practicality of navigation as a way to
suggest query increments.
When in 1999 I looked for a research subject for my master thesis, I
was immediately convinced by these ideas, and eager to work on them. I
was lucky to do my PhD on Logical Information Systems under
supervision of Olivier Ridoux, and defended it in October
2002 [Fer02, FR04]. Its main result is to found LIS on
logics and Formal Concept Analysis (FCA) [Wil82, GW99].
In parallel to theoretical works I have developped a LIS prototype,
Camelis, that implements most of the ideas of my PhD thesis (and
more), and is now mature enough, I think, to be distributed. However
it is still a continuously evolving research prototype, and the main
reason for distributing it, beside its possible usefulness, is to get
some feedback about LIS. This document is about the 4th public version
of Camelis, whose interface has a lot of improvements.
A file system track has been started by Yoann Padioleau, who defended
his PhD in February 2005. The main result is LISFS (a.k.a. LFS),
a LIS implementation as a file system plus the ability to navigate
into parts of files [PR03, Pad05]. This makes LIS benefits
to existing applications without modification. Camelis and Lisfs are not in concurrency as they stand at different levels of
use. On the contrary it is intended that the two tracks converge and
finally fuse at some horizon. Besides, there exist two other
implementations of LIS: Geolis applies to geographical data and
has a web interface [BFRQ08], and Odalisque applies to
OWL-DL ontologies [AF08].
The aim of this manual is firstly to explain the main concepts of LIS,
and secondly to document the various displays and commands of Camelis. The central concept is the context, which comes from
the theory of FCA. In short a context is the combination of a logic, a set of objects, and a mapping from objects to
logical formulas. Section 2 and Section 3
respectively explain what is a logic, and what is an object. Further
sections present operations that can be performed on a context, and
how they are achieved through the Camelis interface: these
operations are browsing, importing and exporting
data, updating, and acting on objects. Section 8
shows how persistency is achieved from one session to another.
Camelis is a very generic system in that the logic and the type
of imported/exported data can be changed at will.
Section 9 presents a general purpose instance of Camelis (called a Camelis application) that can handle all
sorts of files, and especially JPEG pictures, MP3 music files, BibTeX
bibliography files, etc.
For more information about LIS, there is a web page at http://www.irisa.fr/LIS that contains links to people, papers, talks,
and software.