Using sketch recognition for capturing developer's mental models - CRISTAL-CARBON Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

Using sketch recognition for capturing developer's mental models

Résumé

The purpose of agile practices is to optimize engineering processes. Beside it, software documentation often suffers from the priority given to fast and successive deliveries of new functionalities. As a consequence, incomplete documentation and graphical representation make it difficult for a developer to maintain and evolve. Sketching is an integral part of the software design and development. Indeed, among other stakeholders, developers use sketches to informally share knowledge about source code. Since sketches are often hand-drawn, written on paper or whiteboard without any additional technology tool, they are not considered as an artifact in agile method or traditional engineering process but as a disposable production. In this work, we focus on sketches containing Unified Modeling Language (UML) diagrams. To produce documentation or to exploit information from this kind of sketches, developers have to transcribe it in a UML case tool, what they see as a waste of time that hinders productivity. We argue that sketches might be worth considering as non-code artifacts. The fact is that developer or designer drop informally a lot of information about the software, which is unusable. Our goal is to verify that simply capturing sketches can provide useful information for later use and then improve the modeling process efficiency. In this paper, we present a preliminary approach that consists in automatically capture information from the sketches using image processing and pattern recognition. We propose a fledgling prototype that demonstrates the proposal's viability. Then, as a future work, we plan to put in the hands of the developers a finalized version of our prototype and study the added value of our proposal.
Fichier principal
Vignette du fichier
hufamo_paper_4.pdf (2.88 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01951959 , version 1 (11-12-2018)

Identifiants

  • HAL Id : hal-01951959 , version 1

Citer

Tatiana De-Wyse, Emmanuel Renaux, José Mennesson. Using sketch recognition for capturing developer's mental models. The 3rd Workshop on Human Factors in Modeling, Oct 2018, Copenhagen, Denmark. ⟨hal-01951959⟩
286 Consultations
141 Téléchargements

Partager

Gmail Facebook X LinkedIn More