As técnicas de elicitação e de documentação de requisitos nos métodos ágeis
Angélica T. S. Calazans, Roberto Avila Paldes, Eloisa T. S. Masson and Fernando De A. Guimarães
The principles of Agile Methods should be taken into account to create new RE approaches that are also more agile.
The present literature on Agile RE takes focuses on the academic view on the subject. They would like to complement this view with new studies that take into account the industrial view.
They developed a case study on the development unit of a Higher Education organization.
This study started by listing from the literature, the applied elicitation and documentation techniques; and by developing the software developers' profile.
Then, they applied questionnaires in three rounds to understand the most used techniques. The study at first generated results that were not statistically significant, but new rounds have been performed to reach consensus on the results. In these other rounds, the list of techniques have been reordered, based on the results of previous rounds. With this method they managed to reached consensus.
Most important techniques according to their practice:
- elicitation: Interviews, brainstorm and user stories
- documentation: Tasks, user stories and prototypes.
They also discuss the changes in the results, i.e. techniques that started with less consensus and ended up with more; techniques that showed more stable consensus since the beginning.
Q&A:
Observation by Marcos and Marcela that a case study should focus on what is being currently used int he organization, and not what the software engineers think they should use.
The authors argument that they aimed at understanding what the effective techniques are. But Marcela thinks that another kinds of more exploratory study should be used to understand that.
*The question is what leads to the consensus reaching from one round to the other? Doesn't the reordering induces the responses? Is this good or bad?
The authors responded that in the first round, the participants responded alone; in the next ones, possibly, they talked to each other.
There was a comment about the appropriateness of the techniques according to the characteristics of the project. Perhaps new studies could try to understand if these characteristics (size, time, cost etc.) would change the response.
Mapeamento dos processos e artefatos da Engenharia de Requisitos para o eXtreme Programming
Ricardo Duarte Cardoso, Renata Brasil-Silva and Fábio Levy Siqueira
Objective: analyze how an agile method (namely, eXtreme Programming - XP) complies with the best practices of the traditional RE approaches.
They mapped the processes and artefacts of the ISO 29148 to XP's practices and artifacts. They applied another ISO standard to support the evaluation of how to compare two distinct processes.
For which item, they gave a grade: not comply, partially comply, highly comply and fully comply.
Threat: this mapping was made according to the understanding of the authors, which is highly subjective.
The author presented some examples to illustrate how this mapping was made.
Two charts were presented, showing their results: one for processes and another one for artifacts.
They concluded that the XP covers well the processes and the artifacts of the norm
The author thanked the company where he works (ProSimulador), which financed his attendance to CIbSE.
For the future, they want to investigate how the XP methodology can be improved to cover better the practices and artefacts that are not well covered today.
Q&A:
*I made an observation regarding the big threat of this kind of work, which is the subjectivity. Probably, if three other authors had made the same study, the result could be different. I suggested other kinds of empirical studies could be sought to make an analysis with a similar objective.
*Giovanni also made a comment on which XP documentation was used for the mapping, because it is difficult to find a standard XP (he used a book on XP 2.0); and about the profile of the participants. This has to be made explicit, because people from industry and academia may have different ideas on the mapping. He said that two of the authors are from the academia and he is from industry, all three of them are Requirements Engineering people.
Modelado de Requisitos de Seguridad para Big Data
Julio Moreno, Manuel Serrano and Eduardo Fernandez-Medina
Big data is not only about data volume.
Properties to consider when connecting two nodes of data: volume, velocity, variety, veracity and value
So how to connect a new data node to the existing big data cloud?
Their objective is to propose a Requirements Elicitation Framework - they propose that abstraction is necessary to solve this complex problem. So there should be an architectural model which considers the common elements of the data nodes being connected.
He presented a meta-model for Big Data Security Requirements.
Then, he presented an agile approach to make requirements analysis in this context. It starts with identifying the Big Data goals and scenarios. Then, the requirements are defined guided by: Big data goals, organizational context, regulations, scenarios and requirements artifacts.
He gives an example in which the security requirements are focused, having different types: analytical requirements, privacy and security, functionality, hardware etc. For analyzing these requirements, they create a table: for each of these requirements types, the five dimensions discussed before (volume, velocity, variety, veracity and value) are analyzed.
Q&A:
The model generated (i.e. the table) is a very abstract high-level view on requirements to guide the users on choosing and connecting data nodes.
In his literature review, he said that he noticed that many companies want to do Big Data, but they do not know exactly why. They just want to use what others are doing.
Angélica T. S. Calazans, Roberto Avila Paldes, Eloisa T. S. Masson and Fernando De A. Guimarães
The principles of Agile Methods should be taken into account to create new RE approaches that are also more agile.
The present literature on Agile RE takes focuses on the academic view on the subject. They would like to complement this view with new studies that take into account the industrial view.
They developed a case study on the development unit of a Higher Education organization.
This study started by listing from the literature, the applied elicitation and documentation techniques; and by developing the software developers' profile.
Then, they applied questionnaires in three rounds to understand the most used techniques. The study at first generated results that were not statistically significant, but new rounds have been performed to reach consensus on the results. In these other rounds, the list of techniques have been reordered, based on the results of previous rounds. With this method they managed to reached consensus.
Most important techniques according to their practice:
- elicitation: Interviews, brainstorm and user stories
- documentation: Tasks, user stories and prototypes.
They also discuss the changes in the results, i.e. techniques that started with less consensus and ended up with more; techniques that showed more stable consensus since the beginning.
Q&A:
Observation by Marcos and Marcela that a case study should focus on what is being currently used int he organization, and not what the software engineers think they should use.
The authors argument that they aimed at understanding what the effective techniques are. But Marcela thinks that another kinds of more exploratory study should be used to understand that.
*The question is what leads to the consensus reaching from one round to the other? Doesn't the reordering induces the responses? Is this good or bad?
The authors responded that in the first round, the participants responded alone; in the next ones, possibly, they talked to each other.
There was a comment about the appropriateness of the techniques according to the characteristics of the project. Perhaps new studies could try to understand if these characteristics (size, time, cost etc.) would change the response.
Mapeamento dos processos e artefatos da Engenharia de Requisitos para o eXtreme Programming
Ricardo Duarte Cardoso, Renata Brasil-Silva and Fábio Levy Siqueira
Objective: analyze how an agile method (namely, eXtreme Programming - XP) complies with the best practices of the traditional RE approaches.
They mapped the processes and artefacts of the ISO 29148 to XP's practices and artifacts. They applied another ISO standard to support the evaluation of how to compare two distinct processes.
For which item, they gave a grade: not comply, partially comply, highly comply and fully comply.
Threat: this mapping was made according to the understanding of the authors, which is highly subjective.
The author presented some examples to illustrate how this mapping was made.
Two charts were presented, showing their results: one for processes and another one for artifacts.
They concluded that the XP covers well the processes and the artifacts of the norm
The author thanked the company where he works (ProSimulador), which financed his attendance to CIbSE.
For the future, they want to investigate how the XP methodology can be improved to cover better the practices and artefacts that are not well covered today.
Q&A:
*I made an observation regarding the big threat of this kind of work, which is the subjectivity. Probably, if three other authors had made the same study, the result could be different. I suggested other kinds of empirical studies could be sought to make an analysis with a similar objective.
*Giovanni also made a comment on which XP documentation was used for the mapping, because it is difficult to find a standard XP (he used a book on XP 2.0); and about the profile of the participants. This has to be made explicit, because people from industry and academia may have different ideas on the mapping. He said that two of the authors are from the academia and he is from industry, all three of them are Requirements Engineering people.
Modelado de Requisitos de Seguridad para Big Data
Julio Moreno, Manuel Serrano and Eduardo Fernandez-Medina
Big data is not only about data volume.
Properties to consider when connecting two nodes of data: volume, velocity, variety, veracity and value
So how to connect a new data node to the existing big data cloud?
Their objective is to propose a Requirements Elicitation Framework - they propose that abstraction is necessary to solve this complex problem. So there should be an architectural model which considers the common elements of the data nodes being connected.
He presented a meta-model for Big Data Security Requirements.
Then, he presented an agile approach to make requirements analysis in this context. It starts with identifying the Big Data goals and scenarios. Then, the requirements are defined guided by: Big data goals, organizational context, regulations, scenarios and requirements artifacts.
He gives an example in which the security requirements are focused, having different types: analytical requirements, privacy and security, functionality, hardware etc. For analyzing these requirements, they create a table: for each of these requirements types, the five dimensions discussed before (volume, velocity, variety, veracity and value) are analyzed.
Q&A:
The model generated (i.e. the table) is a very abstract high-level view on requirements to guide the users on choosing and connecting data nodes.
In his literature review, he said that he noticed that many companies want to do Big Data, but they do not know exactly why. They just want to use what others are doing.
Nenhum comentário:
Postar um comentário