Scrum > Training > Coaching

The German Scrum Open Space

The delicate flower “German scrum community” is steady growing. Some active German Scrum user will meet on the German Scrum OpenSpace in Munich. Everyone who is interested and would like to contribute is welcome.
You will find further information on the wiki. Or - much better - participate in the yahoo group “deutschescrum”.

Uncategorized

Agile Maturity at the XP 2007 / June 18-22 / Como

Together with Boris Gloger I will facilitate the Workshop “Agile Maturity” on the XP 2007. Read more on the conference program:

“Capability and maturity models and assessments are widely spread. Some companies apply them to improve their process quality and transparency; some just have to comply with their customer’s demands. It is commonly accepted that agile methods improve results and efficiency of development processes. (…)”

Everyone with Agile, CMMI or SPICE experienced will be welcome. We will together find an answer to the question: “How do those two mental models fit together?”

Uncategorized

Agile Quality Management Training

Plan
During the work as Scrum coaches we were very often faced with questions of QA people and Project Manager how the quality will be managed in agile projects. This led us to the idea for a new training. The goal of this training is that the participants understand that quality management is an inherent part of agile software development and not a separate discipline like in a traditional environment. We’ve planned that the training follows the 4 phases of the Deming Cycle: Plan, Do, Check and Act. In our opinion the Deming Cycle describes the heart of Agile which drives the inherent quality management at best.

Do
On February 14th Andreas Schliep and I ran the first Agile Quality Management (AQM) Training in Vienna with 14 participants. 4 had a QA or PM background. 10 were developers with different skill levels. The training was a mix of exercises and lectures including a live demonstration of a pair programming and TDD session.

Check
The last exercise of the training was a retrospective about the training. On the one hand the participant experienced in this exercise that a retrospective after each iteration is a very important act to assure quality. On the other hand Andreas and I could gather their feedback about the training. Necessary for the last phase in the Deming Cycle for this training:

Act
The feedback in the retrospective shows us that the participant understand now the inherent quality management concept of Agile. Otherwise some participant expect a more detailed answer to how maturity model like CMMI fits into Scrum and Agile. However, they understand that a certain CMMI level can be reached with Agile but Agile don’t and won’t guarantee this. For future training we will stick more to this issue.
The energy curve in the middle of the time line (see picture) shows us that we have to improve the part after the lunch. A concrete feedback to this phenomenon was, that the training day was too long. Therefore we will tighten the training for the next time. However, the lunch was obviously very good :-).

At this point I would like to thank Andreas that I could contribute to the training. He had the initial idea. Without his knowledge and creativity this training would never have been realized.

Time Line Retrospective AQM Training

Uncategorized

Pictures of the Chile Tour 2006

Scrum Checklists are available

checklists1

Update: The 2nd print of the Scrum Checklists is exclusively available at InfoQ since October 2006. More than 3000 people downloaded the PDF within just 6 weeks!

Uncategorized

Scrum Checklists

I wrote together with my colleges at Sprint-it Scrum Checklists mainly for our own use. After a while we noticed that this Checklists were very helpful for everyone using Scrum. Our Scrum Checklists are now timely available as a high quality print for the Agile2006.

Yuliya and I wrote an article about the idea of this Checklists and why they are so useful. Enjoy reading!

Download the Article here or at Sprint-it

Uncategorized

Model Driven Acceptance Tests

Last time I saw a fit test table, I realize that this table is nothing more than a Domain Specific Language. This DSL is interpreted by the fit framework in combination with the, from the developer implemented, fixtures. The question is now: why not use this table as model and generate the test code?

In practice we could use the openArchitectureWare Generator. We just have to implement a model adapter for the fitness table and to create a meta model correlated to it. For simplification we would create our own test condition table, for example in an XML format. But the idea of this table should stay the same. Each story, feature or backlog item should described with test conditions in this table.

Why we should go the model drive way? First: the effort is quite the same. Maybe implementing fixtures is easier than implementing code template for a unexperienced developer. On the other side our test suite is now independent from the Fit framework. For example: we need to run a acceptance test suite on an embedded device. This means we need very small test code and there is no TCP Socket to output the test results. A very specific test code with a simple output to a serialize interface could help.

The second big benefit to generate acceptance test code is more obviously. Why just generating the test code? Why not generate the code at once? Because usually on the beginning of a agile project, the stories are quite vague and therefore the architecture too. But later in a project, especially when the architecture is stable, we could use our generator know how and our AT model to start generating code artifacts. And, because we have already a acceptance test suite, this architecture is also already well tested.

o.k. - this is just an idea. But I’m relay interested if this is an useful way to combine TDD and MDSD. Maybe someone knows more or has some more good ideas about that issue? Feel free to add a comment.

Uncategorized

Evolution: the Killer Application

On my media server at home I use Ubuntu since one year and I’m very satisfied with this distribution. It’s a great playground for a technician like me and easy to use for my family. On my laptop I use Windows XP and think with the idea to replace it with a Ubuntu installation. But there are this dumb hardware problems especially with laptops. And I’m very glad that my working environment with Mozilla Thunderbird, GnuPG, Palm Treo 600, OpenOffice and many other special application is working well. But what I miss on this environment is a calendar application supporting the iCal format. Mozilla calendar is to buggy. Evolution is just running on Linux. But Evolution is so nice …

What I now have running on my laptop after one night playing around is the following. Windows XP as the basic OS. VMware Player to run a small Ubuntu with no Xserver and Cygwin with Xorg. Almost only to run Evolution. And I’m very surprised how uncomplicated this construction is. It took just 10 minutes and all my addresses and appointments were copied from my Treo to Evolution. Now my decision is clear: If Ubuntu “Dapper Drake” is released I’m going to switch to Ubunto on my laptop too.
Evolution: my killer application.

Evolution on Windows XP

Uncategorized

Buffers in Selected Backlogs

During the work with 15 new scrum teams this week, I learned a lot about backlog items used as a buffer. As I didn’t have so much time to understand every selected backlog I started searching for buffers in them and use the buffer as indicator.

If I found a buffer, I asked the team and the product owner if this buffer is related to the goal of the sprint. If yes, the BL was just lumpy and the PO didn’t do his work well or he didn’t understand scrum. I tried to overcome with the PO to remove the buffer and explained him that he always have the possibility to add features to the backlog and then could reprioritize it.

If the buffer was not related to the goal the team has obviously a second PO. E.g. the team has to handle bug fixing requests. These requests are very random and therefore they want some buffer to handle them in there BL. In this case I tried to overcome with the PO that he took the responsibility to add the incoming requests to the BL and reprioritized it. He also should become an agreement with the team, that he could do this with their commitment during the sprint.

In the case that I couldn’t find a solution with the involved people and the buffer was further more on the BL, it was for me a good idea to add this to the impediment list. Because of that it was obviously for the organization, that the team was slowed down. Therefore a buffer in a selected backlog can also have some positive aspect.

Uncategorized

Acceptance Tests FAQ

The work with product owner shows, that they have many question about acceptance test. This FAQ collects this questions and gives for each of them an answer as short as possible.

Download Acceptance Tests FAQ

Uncategorized