I have participated at a few Open Space events, and facilitated some, but the crowd never exceeded 60. So I was a bit apprehensive, when it appeared that the Open Space Technology part at Valtech Days Paris 2007 would gather more than 150 participants. Here are some things we did and lessons we learned.
I am going to attend the Paris on Rails conference on December 10th. This is a rare treat: there are no other Ruby-related conference in France to my knowledge. Exciting!
With all the techniques that we have at our disposal, with all the education that developers got, still we see websites that handle internationalization poorly. I noticed the problem recently on a minor website, but even Flickr and/or Yahoo got it badly wrong.
I first got introduced to using hexagonal cards for facilitation at AgileOpen. I thought they were pretty neat, and I acquired two packs of them for use at the Valtech Days. Here are some notes on how to use them.
Continue reading
My quest for spreading the word about Hudson continues.
Continue reading
Vincent Massol has been kind enough to step up at the last minute as a presenter at the Valtech Days. His talk was about XWiki, a tool he is working on with his team at XWiki.com (he also hold a session during the Open Space Technology part). Also in the room was Ludovic Dubost, the CEO of XWiki.com.
My pictures are stored on flickr and I have also put my notes on the conference wiki: Intégration Continue, Outillage Agile, Recruter un bon profil Agile.
All notes from the Open Space sessions, links to blogs and pictures are available on http://valtechdays.pbwiki.com/.
Wow. I started writing this post on my way back from the Valtech Days conference, which I have contributed to organize, and all I can say is “wow”.
Concordion, a tool that was presented at CITCON Brussels a week ago, is finally out.
The approach is interesting and can be compared to FIT or FitNesse. Where FIT and FitNesse let analysts/testers give examples in the specifications documents as a series of inputs and outputs (in Word documents for FIT, and wiki pages for FitNesse), Concordion lets you write the examples in HTML, expecting tags to mark which value is for what purpose.
Example:
<p>
The greeting for user <span concordion:set="#firstName">Bob</span>will be:
<span concordion:assertEquals="greetingFor(#firstName)">Hello Bob!</span>
</p>
If you feel that this is making life hard for a large amount of values, fear not: there is a way to use HTML tables as well.
I think that the approach is interesting. It certainly helps making natural speech useful for integration tests. That said, I cannot really imagine analysts or testers writing HTML pages with tags. Word documents (or HTML documents edited with Word) or wiki pages seem much more plausible. Another thing is that the tooling for adding the appropriate tags seems limited. Can we really expect these people to add HTML tags (that is, they cannot easily use a wysiwyg tool for that), and give them values with such constraints? We’ll see with time, I guess.
Other differences between Fit/FitNesse and Concordion are discussed here.
The second edition of Paris on Rails will take place on December, 10th.
I’ve been wanting to learn Ruby for some time, but it’s only since I was offered the book at XP Days that I seriously started investing time in it. The conference will be cheap way to catch up with the rest of the Ruby crowd.
After doing it at AgileOpen, I again had the chance to demonstrate Hudson. The attendants had all proved their CI credentials, to say the least, as they included developers on AnthillPro, Build-o-matic, and CruiseControl, of course (how many CC developers in the room? at least 6, I’d say).
I think my point that Hudson was the easy, very usable alternative, went well, but they were more impressed with the combinations matrix feature (I personally think that the killer feature is the simpleness of the form where a job is defined).
At the evening drinks, one or two people asked whether I was part of the development team, which was flattering. I guess I should consider getting more involved, but there are already so many things I want to do, and so little time! How do all these guys do?!
I was very excited when Squirrel suggested a topic about mutation testing. I had looked at Jester & Jumble before, but gave up after a couple of hours, as they are not very easy to setup, and mostly not maintained.
The idea of those tools is to make changes to the source code (Jester) or the bytecode (Jumble), and check if the tests still pass. If they are well written, they should fail when the code is changed. If not, then the code is not covered properly. Ideally, you should only have one test failing; otherwise, it means that you have redundant tests.
Ivan, the creator of Jester, was present, but admitted that he had not even used it in years. After seeing the interest of the participants, he seems willing to give it more time. Hurray!
Check out my notes on the conference wiki.
I came back from Brussels this Sunday evening, after some visiting of the city. The conference has been a blast, especially networking with other participants. It has been great seeing again people that were there last year in London… and sometimes pick up the conversation where we had left it! I also talked with a few people that were at AgileOpen in the Netherlands last June. I can’t wait to see them all a third time in 2008.
Continue reading
The folks at developpez.com (the de facto leading French website for developers) have interviewed me about the Open Space Technology conference that we are having at the Valtech Days on Wednesday.
Just kidding, of course. Or am I?
Check out this article sent to me by a colleague. It is full of little gems that I cannot resist to share.
What about this line:
“Even some who are otherwise fans of open source cite the importance of good version control.”
Err… what is that supposed to mean? That many open-source developers do not care much about version control? That would be some news! Or maybe that good version control systems do not exist as open-source? But then, why quote Andy Hunt just one paragraph earlier: “excellent free, widely used, open-source solutions abound—CVS and SVN come to mind.â€
“Rajiv Delwadia, chief technologist at life-cycle planning and management application company VersionOne in Atlanta, said that most open-source tools are good enough for most teams. “But if you’re going to pay for anything, pay for good source control,†he added.”
OK, sensible enough… I guess… but wait for the kicker:
“(Delwadia’s team uses Microsoft’s Visual SourceSafe.)”
So… they are throwing away good money for VSS? One of the most hated tools on the market? When so many great free alternatives exist? Priceless, I say!
“doing incremental builds for production software (instead sending complete builds through QA on the way to release) is too risky, the company [Electric Cloud] says”
…uh? I guess the reporter got it exactly wrong. As we all now know, incremental builds are the right way of doing releases.
Maybe he really wanted to say that *partial* builds (say, one or 2 dependent components) were risky?
Ah well. I want to be sympathetic. Really. It’s not easy being a journalist when the tech guys you talk to all do not care much that all the information is properly communicated.
