Gilberto Borrego
Instituto Tecnológico de Sonora, Computación y diseño, Faculty Member
An important challenge in Agile Global Software Development (AGSD) is architectural knowledge vaporization, i.e., the loss of technical knowledge due to a lack of documentation. In a previous work we identified that technical knowledge is... more
An important challenge in Agile Global Software Development (AGSD) is architectural knowledge vaporization, i.e., the loss of technical knowledge due to a lack of documentation. In a previous work we identified that technical knowledge is usually available in unstructured textual electronic means (e.g. chat, mail, blogs, etc. – also known as
UTEMs) used by AGSD workers during development, although not necessarily easily accessible. Also, we proposed the concept of Knowledge Condensation as a means to recover knowledge from UTEMs, by introducing a structure to their contents along with means to retrieve knowledge. In this work we present our first step towards a Knowledge Condensation tool, through a prototype tagging mechanism to structure
architectural knowledge during UTEM interactions. We evaluated the prototype and obtained the following results: (1) a high perception of usefulness and a high projected intention of use; and (2) the identification of 6 tagging behavior profiles from participants. These tagging behaviors can be grouped as (i) participants who tag as expected, (ii) participants who tag more than expected, and (iii) participants who tag less than expected. These results suggest that it is important to cope with “little or no tagging” and “excessive tagging” behaviors to implement knowledge condensation based on UTEM tagging. Furthermore, we identified that tagging could be a good option
to structure UTEM interactions, and to retrieve architectural knowledge from UTEM logs. Future work includes improving the tagging mechanism prototype based on the participants’ suggestions and overall evaluation results, and implementing a searcher prototype to further evaluate the concept of knowledge condensation based on tagging of UTEM interactions.
UTEMs) used by AGSD workers during development, although not necessarily easily accessible. Also, we proposed the concept of Knowledge Condensation as a means to recover knowledge from UTEMs, by introducing a structure to their contents along with means to retrieve knowledge. In this work we present our first step towards a Knowledge Condensation tool, through a prototype tagging mechanism to structure
architectural knowledge during UTEM interactions. We evaluated the prototype and obtained the following results: (1) a high perception of usefulness and a high projected intention of use; and (2) the identification of 6 tagging behavior profiles from participants. These tagging behaviors can be grouped as (i) participants who tag as expected, (ii) participants who tag more than expected, and (iii) participants who tag less than expected. These results suggest that it is important to cope with “little or no tagging” and “excessive tagging” behaviors to implement knowledge condensation based on UTEM tagging. Furthermore, we identified that tagging could be a good option
to structure UTEM interactions, and to retrieve architectural knowledge from UTEM logs. Future work includes improving the tagging mechanism prototype based on the participants’ suggestions and overall evaluation results, and implementing a searcher prototype to further evaluate the concept of knowledge condensation based on tagging of UTEM interactions.
Research Interests:
— Agile Global Software Development (AGSD) is a reality, since nowadays software products are required to get into the market with more speed than before. This situation has pushed Global Software Development (GSD) companies to adopt... more
— Agile Global Software Development (AGSD) is a reality, since nowadays software products are required to get into the market with more speed than before. This situation has pushed Global Software Development (GSD) companies to adopt lighter ways to develop software (Agile Software Development-ASD) to satisfy market demands. However, AGSD companies have encountered increased technical debt and architectural knowledge (AK) vaporization, mainly because the inherent differences between ASD and GSD, especially in documentation handling. This paper describes a research project in which it is proposed to exploit the AK that is recorded in unstructured textual electronic media (UTEM) generally used in AGSD, in order to decrease the problems of software evolution and maintenance, caused by a lack of adequate AK. Until now, the preliminary results of this research show that this approach could be feasible in AGSD environments.
Research Interests:
—Many companies have adopted agile software development (ASD), mainly due to it can handle scarce requirements. However, some unsolved challenges exist in ASD, particularly in global software development (GSD) companies (known as AGSD).... more
—Many companies have adopted agile software development (ASD), mainly due to it can handle scarce requirements. However, some unsolved challenges exist in ASD, particularly in global software development (GSD) companies (known as AGSD). These challenges include ASD lax documentation contrasted by the methodological standardization required in GSD, due to its inherent distances. Lax documentation leads to documentation debt and architectural knowledge (AK) vaporization, which cause negative effects on the development process and on the product itself. In order to reduce these effects, we propose Scrumconix, a hybrid method that uses a lightweight approach to document in AGSD environments, which also aims to decrease the effect of linguistic and cultural distances. In addition, we present preliminary results of Scrumconix implementation in a Mexican AGSD company.
