Mauro wrote on 09/11/2007 10:39:18:

> I do think toy and example apps can be very useful but as an 
> complement rather than a substitute for unit/behaviour testing.

Absolutely! I was thinking of the toy projects as the equivalent of 
stories and automated scenarios - it always seemed ironic to me that while 
JBehave drove JBehave at a unit level, it never exercised its own behavior 
at a system level. I'm happy to do the hard work for this. We can knock up 
smaller examples for any stories that are needed for real-life projects 
(if the toys don't need them anyway).
 
> Looking forward to trashing out a roadmap for JB 2.x!

Sweet. :)

Dan's persuaded me that Mercurial really isn't that tricky to use, and 
he's promised to blog his experiences with it to soften the learning 
curve. We're setting up a repository, and hopefully I'll be spiking out 
the first of the toy projects later next week, just to get an idea of how 
easy it is and how the build etc might work. It's all open for discussion.

(I'm quite excited about the distributed SCM. I didn't explain well what I 
meant by checking in complete features to a stable branch. I want to avoid 
the situation where we can't release bugfixes, incremental improvements or 
new features because we're waiting for a particular feature to be usable. 
Does that make sense?)

We should definitely meet up for beer next week and chat.

Cheers,
Liz.

--
Elizabeth Keogh
[EMAIL PROTECTED]
http://sirenian.livejournal.com
http://jbehave.org

Reply via email to