Introduction
EVAT software allows the visualization and navigation for trees. As a
consequence, it allows some operations of comparison between trees as finding
common subtrees. Find must be understand here as presenting common things with
same colors. Overall implemented tools are strongly based on intrinsic
combinatorial parameters with as few as possible references to syntactical data.
The main idea is to combine automatic methods coming
from enumerative combinatorics and the expertise of the eyes for the
construction of a valuated navigation (this deal was initiated by Mandelbroot when
he points “le retour de l'oeil, seul modèle mathématique dont dispose celui qui
cherche à saisir l'inconnu et l'incontrôlable, chassé par Laplace”).
One new feature shown in EVAT is a focus-context approach in relation with the drawing.
It means that zoom are not done using the classic fisheye algorithm. The drawing is recomputed using
the selection and the parameter that the user have choosen.
EVAT was constructed under the framework Tulip developed at LaBRI.
Material and limits
EVAT is running under Redhat9.0. It is suitable to hold a 512Mb memory and a graphical card with Open GL 3D accelerator.
EVAT can handle in a same session several huge trees. For example, for 550 000 nodes, the size of the used memory is 144Mb.
Knowing the Tulip framework, it is easy to extend EVAT.
Terminology
- Parameters - values assigned to a node coming with the files or calculates within EVAT.
- Attributes - values used in the visualization of trees : size, position, color of the nodes
- Regular Expression - value in use with QT some simple rules are :
- ".townsend" means one character followed by townsend
- ".*townsend" means any string followed by townsend
- "^a.*jpg" means a chain beginning by a followed by any string followed by "jpg"
For more details, see the QT documentation