Till innehåll på sidan

An Investigative Eye on the Agile Manifest

Publicerad 2014-04-07

Identifying, creating, and developing appropriate work methods for R&D has long been on the agenda for scholars of engineering processes. E.g. concepts such as front-loading, different forms of concurrent engineering, and waterfall are now part of the standard toolbox.

Now, there is a new addition to the toolbox; so-called Agile methods have recently gained a lot of attention. A large number of developers are now shifting their devotion away from the “traditional” engineering methods to instead embrace the constantly changing character of a dynamic world. Emanating from software engineering, these methods, which go by various labels, are argued to distinguish themselves from traditional methods on a number of different areas. Accordingly, a fundamental assumption in the traditional view is that products are fully specifiable, predictable, and are built by meticulous and extensive planning, often using waterfall as a development model. Agile development, however, is developed by continuous design improvements and testing in an evolutionary manner allowing for faster feedback and change of the design. The former view is argued to be more rigid and formalised, relying on the traditional chain of command hierarchy as well as on heavy planning and late testing. This is contrasted by the flexible and organic manner, relying on informal collaboration in smaller teams as well as continuous development of the design and testing used in the Agile methods. A main goal of this shift is based on the idea to promote fast responses to changing environments and user requirements.

An example of an Agile method that is widely spread in both hardware and software settings is Scrum: consisting of cross-functional teams, daily stand-up meetings, sprints with demo or review and retrospective meetings at the end of each sprint. The sprints usually last a few weeks and focus on the delivery of a single feature of the product.

There is a lot of appealing substance in this line of thinking. However, the scientific validity of Agile is not ascertained. On-going research by Industrial Ph.D. student an Ericsson employee Maria Carmela Annosi, Professor Mats Magnusson, Researcher Jens Hemphälä, and Ph.D. student Susanne Nilsson at Integrated Product Development, has contributed to a more in-depth understanding of Agile and its effects. Although there are a high number of developers with very positive experiences deploying this method, some crucial tensions are emerging. These are tensions related to three main areas: (1) integrated products containing hardware and software in contrast to pure software, (2) safety critical development in contrast to perpetual beta release type development, and (3) development of radically innovative products in contrast to incremental adaptations of existing products.

In order to validate the emerging concerns about Agile, a quantitative study is currently being conducted at a large international R&D unit at Ericsson that has worked consistently with Agile for a number of years. This will be one of the largest studies of the effect of this rapidly spreading method. The results of this study will be available later this year.

A number of joint interests for the Machine Design department, and within KTH at large, are also emerging in this field. The interest in Agile is especially interesting due to the subject’s potential to integrate activities within our department and within KTH. Apart from the pure research efforts, the interest in Agile also spans other areas: industry relationships as well as education. Currently, there is an effort underway to raise the topic of Agile in ICES - The KTH Innovative Centre for Embedded Systems. The ambition is to establish the interest for a Focus group regarding Agile for ICES members within the year. Further, the topic at hand also relates to a number of our capstone courses where students currently are encouraged to work in Agile inspired ways. Very exiting things may emerge as a result of the intersection of these domains.

My hope is that these burgeoning initiatives will provide a garden where joint interests can grow. Please contact me for further dialogue.

Jens Hemphälä (hemphala@kth.se), Researcher at Integrated Product Development, KTH Department of Machine Design