Système de représentation d'interfaces centré sur l'usager

Dubé, Pierre-Samuel (2005). Système de représentation d'interfaces centré sur l'usager. Mémoire de maîtrise électronique, Montréal, École de technologie supérieure.

Prévisualisation

PDF Télécharger (7MB) | Prévisualisation

Résumé

Le système Activity-oriented Interface Representation ( AIR ) est un outil d'aide à la conception et à l'évaluation des interfaces usager. Du côté de la conception, il permet de spécifier le détail des interactions entre un usager et une interface à l'aide d'un langage. Pour ce qui est de l'évaluation, un outil ou modèle permet de prédire le temps d'exécution d'un usager hautement familier avec les interfaces graphiques avec une précision d'au plus 26%, ce qui est acceptable en modélisation analytique. Ce modèle ne tient compte que des opérations explicites de l'usager, soit celles qui on un effet direct sur l'interface. Le modèle fût mis au point et sa précision validée par le biais de tests avec sujets humains dans des conditions contrôlées. Le système AIR comprend des outils logiciels pour faciliter l'utilisation du langage et du modèle.

Titre traduit

Activity-oriented interface representation system

Résumé traduit

This research project introduces the Activity-oriented Interface Representation (AIR) system which aims at specifying interactions between a user and an interface. This system covers two points. First of all, we developed the AIR language allowing task execution specification at a detailed level. In order to describe a given task, a script has to be written with this language. For each step of a script, the description includes four elements: 1) User action. 2) System internal state. 3) Interface feedback. 4) Duration. Each step can be specified at different levels from an abstract description ( e.g. move an icon, open an application) to a description at the elementary operation level (e.g. typing of a keyboard key, pointing movement). During the course of the project, we also developed a software tool to write AIR scripts in text and HTML ( « Hypertext Markup Language ») formats and to verify their respective syntax.

Secondly, we developed a tool to predict user execution time (for tasks on graphical interfaces) from AIR scripts. The prediction is done on three overt operations: 1) Typing of a keyboard key (K). 2) Pointing movement (P). 3) Mouse button click (C). We first transform an AIR script into a sequence of overt operations without considering system state change and interface feedback. Such a sequence is called a KPC model. For each operation, we then multiply its number of occurrences in the sequence by an estimate of its mean execution time. Thus, we are able to predict time taken by a user in his execution of an AIR script. The KPC model assumes that users are skilled. Skilled users are very familiar with graphical interface usage which is the case for most computer users these days. Our hypothesis states that even when users are not really familiar (novice state) with a given interface, they will learn fast how to use it because of the automatisms that they developed with time. The KPC model was tuned and validated via an experimental study on 20 skilled subjects. The first half of the subjects was used to determine the estimates of the K, P and C mean execution times. The other half helped in determining the prediction accuracy. From the obtained results, we can affirm that it is possible to predict skilled user (novice or trained) execution time from a KPC model. For skilled novice users, we obtained an error in prediction not exceeding 26 % which is acceptable according to the standards in analytical modeling. According to the same results, it has been shown that skilled users, novice or trained, can execute a task using multiple efficient strategies. In other terms, there is no optimal strategy in executing a given task. We propose that in the near future, it will be possible to record sequences actually executed by skilled users in order to build AIR scripts in a semi-automatic fashion.

Type de document:	Mémoire ou thèse (Mémoire de maîtrise électronique)
Renseignements supplémentaires:	"Mémoire présenté à l'École de technologie supérieure comme exigence partielle à l'obtention de la maîtrise en génie électrique". Bibliogr.: f. [250]-255. Chap. 1. Revue de la littérature -- Chap. 2. Langage de spécification AIR -- Chap. 3. Outils logiciels AIR -- Chap. 4. Modèle KPC -- Chap. 5. Mise au point et validation expérimentale du KPC -- Chap. 6. Discussion.
Mots-clés libres:	Activite, Air, Analytique, Evaluation, Graphique, Interaction, Interface, Kpc, Langage, Modele, Modelisation, Oriente, Representation, Specification, Syteme, Usabilite, Usager, Utilisateur.
Directeur de mémoire/thèse:	Directeur de mémoire/thèse Fimbel, Éric
Programme:	Maîtrise en ingénierie > Génie électrique
Date de dépôt:	31 janv. 2011 21:11
Dernière modification:	01 nov. 2016 01:06
URI:	https://espace.etsmtl.ca/id/eprint/374

Gestion Actions (Identification requise)

Dernière vérification avant le dépôt