Research on globally distributed software engineering - Distributed Scrum als oplossing voor outsourcingsproblematiek - Rini van Solingen 1 Who am I? Rini van Solingen, 39 years old, married to Patricia, 2 kids (Bo en Bas), lives in Zoetermeer, The Netherlands Technical informatics TU-Delft (MSc) Technology Management TU-Eindhoven (PhD) Head in the clouds, feet in the dirt (industry-science) Schlumberger, Fraunhofer IESE, CMG/Logica, Mavim, isense Prowareness, TUDelft, TUEindhoven, Stenden University Loves to lead, investigate, teach, publish and coach 20% TU-Delft, 80% isense Prowareness, 20% Logeerplezier December 27, 2010 2 1
The Delft Software Engineering Research Group (http://www.se.ewi.tudelft.nl/) Group size: ~25 fte, including ~15 full time researchers (PhD candidates, postdocs). Teaching: Programming and project skills of Delft computer science students ~20 MSc projects p/y in software engineering Research: Software evolution Reverse engineering Software testing Services Model-driven engineering Web engineering Embedded systems Globally distributed SE Close collaboration with industry 3 4 2
Adding a global perspective to SE Source: Carmel & Agarwal, IEEE Software 2001 5 Allen Curve 30 meter principle Probability of Communication Weely Communication 5 4 3 2 1 0 0 20 40 60 80 100 Distance (meters) Source: TJ Allen, The Organization and Architecture of Innovation, 2007 8 3
The Research Challenge for GDSE Our Quest for GDSE research: Creating the virtual 30 meters Deep understanding of GDSE issues Both theoretical and practical Real versus perceived distance Technological solutions are available and still sufficient additions feasible and required The virtual project space as dot-on-the-horizon 9 Although: the Virtual 30-meters. 10 4
Although: the Virtual 30-meters. 11 Virtual 30-meters: Traceability Requirements Source code related to task Task status 13 Test cases related to requirements 5
The TU-Delft GDSE Research Agenda Successfully deploying GDSE in practice Best-practices of companies, success-factors and fail-factors from practice are an important focus in my research Technological support for distributed sw-development Technology to support a decreased perception of distance and increased transparency Distributed agile development Agile methods increase the level of informal communication, coordination and control; Agile might solve many issues in distributed development Follow-the-Sun software development The GDSE stress-test ; if tools and technology work for FtS, they are likely to be applicable elsewhere too Teaching GDSE dynamics Prepare our current student population for a career in which GDSE plays a role December 27, 2010 14 14 Example 1: Best-practices for Distributed Scrum Scrum as solution to a problem 15 6
Example 2: Technological Support for distributed agile December 27, 2010 16 Example 3: experiment FtS development December 27, 2010 17 7
Example 4: Routing for FtS Http://www.follow-the-sun.org 18 Example 5: GDSE Game Running GDSE company Deciding Lab locations Defining Strategy Customer focus Projects vs Products Advance decision making Teaching GDSE dynamics 19 8
The TU-Delft GDSE Research Agenda Successfully deploying GDSE in practice Technological support for distributed sw-development Distributed agile development Follow-the-Sun software development Teaching GDSE dynamics December 27, 2010 20 20 Best-practices for distributed Scrum December 27, 2010 21 9
Best-practice 0: If Single Roof is Possible: Do It! Don t distributed if not necessary 22 Best-practice 1: First deploy Scrum locally before working distributed Deploy by the book : inspect and adapt empirically 23 10
Best-practice 2: Assign Scrum roles explicitly, define proxies and ensure alignment between these PO-role even more critical and crucial 24 Best-practice 3: One team in one rhythm Team members are located on all locations and work in same sprints 25 11
Best-practice 4: Meet Teams are not build up by themselves: actively establish personal relations 26 Best-practice 5: Impediment resolution and Retrospective effectiveness remains crucial Meet for retrospectives 27 12
Best-practice 6: Work at customer location at least between 10-20% of the time 28 Best-practice 7: Personal mindset is crucial: what did I do wrong? what can I do different? what can I do to help? 29 13
Best-practice 8: Don t focus on tools: discussion and interaction is more important 30 Best-practice 9: Communication & Awareness is no Automatism On this issue tools do help 31 14
Best-practice 10: Fail fast: improve empirically Both successes and failures are sources for learning 32 Scrum Management Book January 2011 (expected) The Power of Scrum An inspiring story about a revolutionary project management approach Authors: Jeff Sutherland Rini van Solingen Eelco Rustenburg December 27, 2010 33 15
Thank you for allowing me to talk about my passion! Questions? D.M.vanSolingen@tudelft.nl R.vanSolingen@prowareness.nl www.rinivansolingen.nl 34 16