Agile2009 Take Aways

In addition to the usual "Made in China" conference trinkets that tend to disintegrate on the next day... ;-)

Most of the people there were not developers - quite a difference from Java One or Devoxx. This impacted how the sessions looked like - more talking, less demo's; more social, less technical.

Key points (tracks) of the conference for me (random order):

• Pomodoro technique. I attended a session about Pomodoro, it was interesting enough for me to try at home. My milleage 'varies' so far, but I really want to make it work. There is a simple but useful timer for Mac.
  
• A lot of stress on TDD. Many people talked about how great it is, how it helps them write good code, how it helps the design to emerge etc. I guess I have finally been convinced to the when.../should... style of writing unit tests.
  
  My favorite quote was "TDD (test first) is the only way to ensure your code is testable".
  
• Lot's of stress on individuals and interactions over processes and tools, despite visible presence of tools vendors, esp. continuous integration ones.
  
• I really wanted to learn about Kanban, but I rather failed to achieve that. I attended a session about Lean (and Lego bricks), it taught me a bit about Lean background in manufacturing, but not directly about using Kanban in software development.
  
  There was another session that was intended to prove that Kanban is better then Scrum that I haven't attended, but from what I heard the presenter exaggerated Scrum faults too much just to prove the thesis.
  
  Later I have read the article by Henrik Kniberg, it answered most of my questions.
  
• Agile is dead. Not literally, but many voices say that Agile already became mainstream, so it's time to bring Agile to higher level.
  
  Everybody already uses Agile for small non-mission critical systems for which Agile was originally advocated. Now it's time to bring Agile to big, mission-critical and corporate environments, or at least this is the subject the 'thinkers' should focus on.
  
• Game theory in agile contracts was an eye opener. The message to me was "stop being so narrow-minded and think". During this session it was applied to formulating an agile contract equally profitable to both parties. I was too married to my ideas of how the contract should look like that I either favored one party too much or provided too little incentives to get the best contract outcome.
  Of course this was just a simplified exercise, but it showed that out of the box thinking can be applied also for contracts.
  

The more I think about it, the more I agree with the concept that software development is not about computers and technology, it's about people.

maven-eclipse-plugin problem

Just short maven tip. Maybe someone will find it useful.
Few days ago I faced following problem when trying to generate eclipse project using mvn eclipse:eclipse:

Resource directory's path matches an existing source directory. Resources will be merged with the source directory src/main/resources ... Request to merge when ‘filtering’ is not identical.

The solution is to use version 2.6 of the maven-eclipse-plugin (version 2.7 will not work) so add the version tag explicitly to the pom or call maven:

mvn org.apache.maven.plugins:maven-eclipse-plugin:2.6:eclipse

Agile2009 Conference Debriefing

OK, so last week we spent in Chicago at Agile 2009 Conference.
Here is what I think about it.

Main facts

• 1350 attendees, mostly from the US (I estimate them to be 70 - 80% of all participants).
• more than 50% of sessions presented by a pair of speakers (pair-presenting). Unfortunately it was often problematic rather than beneficial.
• up to 22 parallel sessions
• most of the sessions took 90 minutes. Some took 45 minutes. Only a few 3 hours.
• every day sessions started at 9:00 and ended at 17:30. Then less official activities lasted till midnight or so

Main observations

• CI vendors crowd. I dare say CI tools vendors constituted with project management tools vendors vast majority of the exhibitors. Also, there was a lot of sessions related to Continuous Integration (comparing to our only single session about code review)
• max 1/3 of attendees were active developers (really technical folks), the rest: project managers, product managers, coaches, psychologists, business development managers, social sciences students and consultants,
• all really technical talks gathered relatively low attendance rate. On the contrary, social or project/product management sessions usually gathered 50+ people.
• Kanban gathers really a lot of interest (it's already an anecdotal war going on between Kanban and Scrum). However vast majority of people seem not to understand what Kanban is and how it applies to software development projects. I attended 2 sessions about Kanban and I am still confused. Just another hype? Another buzzword? Jeff Sutherland said: "Kanban is Scrum without the sprint constraint". Is it true?
  

The good

• a lot of interesting topics (with varying level of expertise expected) and excellent speakers
• good venue, hotel, food, service and the staff
• unique accumulation of sessions about social/cultural aspects of agile software development
• a lot of networking opportunities (these folks were in general keen to meet new people)
• our talk was rated at almost 4 out of 5 points. There were several positive comments (e.g. about the lack of marketing buzz from our side) and no negative ones. Still, I know that I still sucked as a speaker, but at least I can practice and learn
  

The bad

• wireless worked only in a few places (but it was really fast and reliable there)
• 90 minutes was too much time for a lot of the sessions (including ours). 45 minutes was sometimes not enough. I think that sticking to 30 minutes and 60 minutes for most of the sessions (excluding tutorials/clinics) would be much better.
• when you decided to go for 45 minute session, you had usually very few options after it, as most of other sessions were 90 minutes long and were already in their second half (especially as there were no breaks then).
• 22 parallel sessions were sometimes too much. I wish I had cloned myself into several copies. Fortunately Slawek naturally selected different sessions, so we could cover usually 2 tracks.

Interesting sessions and quotes (not necessarily verbatim)

• Alistair Cockburn: “if you ask me to give you advice about what to do to improve your agile distributed teams, my simple answer is: don't do distributed teams”.
• Esther Derby (a session about performance appraisals): “you are afraid of training your employees because they may leave when they are trained. But if you don't train them ... they will stay!”
• Jared M. Spool: “Crappy people have no problems with producing crappy products.", "Get rid of quality as a requirement and writing software becomes easy.”
• Giora Morein at a session about distributed agile: “change phones into planes”. The speaker (many years of agile experience) is convinced that only by meeting face to face and establishing personal ties, distributed teams can reasonably cooperate. It calls it “Ambassador”, with several subtypes including:
  • delegate
  • traveling star
  • people exchange
  Travel has to be voluntarily. I shared these hints:
  • sites must be located in interesting cities (touristic targets)
  • do not send two or more people to a remote team at the same time - they happen to create a ghetto (isolate and wander together) rather than immerse and socialize with others
• UX design is critical (gave Apple 75% market share over 7% of the next competitor in MP3 players market). The following 3 rules determine whether you care for UX (and thus can develop really awesome software):
  • at least every 6 weeks you (team member) observe for at least 2 hours end users actually using your software (feedback loop). You don't do it, you are screwed.
  • every employee can give you a vision of the product in 5 years from now
  • rewarding for failures (no failures -> nothing innovative, risk aversion companies fail)

Keynotes

There were two keynotes: delivered by Alistair Cockburn about next phase of agile adoption (as agile philosophy was melted down in the ecosystem and now agile is facing new challenges: how to apply to big projects in sophisticated organizations - something which agile methods originally did not (want to) address). Alistair is a great charismatic speaker but I did not find anything new, specially interesting or revolutionary in his talk, although, maybe surprisingly, I liked it.
Another keynote was presented by Jared M. Spool and treated about User Interface Engineering. This was a great, funny and energetic talk. Jared was shining. He emphasized how important user experience is and that it can be a key differentiator in company portfolio (e.g. Apple case). Key learnings were listed above.

Conclusion

This is really a good conference and I think it's worth to be there also next year (Nashville, August 2010).