RE: [scrumdevelopment] Re: Scrum in 2003
- Jens,It sounds like you will have better success if you can break down the backlog items into smaller pieces of useful functionality that are easier to grasp, and take a more evolutionary approach to design. Something like "must have perfect history" might be the vision, but you may want to do it in multiple stages (eg start with one type of history entry, or start by ignoring some aspects of the "extra dimension" that adds complexity in your case). Of course, you want to find subsets of the "perfect history" feature that can be meaningfully demonstrated to the customer. Each subset is easier to design and implement, so there will be less feeling of chaos and uncertainty.Your statement that "architectural issues have to be correct" is not shared by the XP community. Architecture can evolve, and at any point it only has to be good enough to handle the backlog items you have implemented.- Charlie-----Original Message-------- In firstname.lastname@example.org, deborah@h... wrote:
From: je@... [mailto:je@...]
Sent: December 17, 2003 8:19 AM
Subject: [scrumdevelopment] Re: Scrum in 2003
> This is the second time I've seen mention here of "chaos in week 1"
> and it surprises me... are the items in your backlog <= 16 hours,
> recommended? If so, shouldn't chaos last only hours at a time?
I do belive our chaos (complexity) is due to the fact that we have
not had proper backlogs. This would mean that things are not thought
thru, before, the making of the sprintlog.
For instance, in one of the sprints there were an activity which
read "must have perfect history". We all know this can be a
challenge, and this particular system had an extra dimension. The
team set aside hours for analysis, db modelling, etc. After 3 days
the team sees that there solution is not good enough. They have to
start over, putting more hours in the sprintlog, etc, and the team
worries. Archtectural issues are hard to estimate since they have to
be correct. If there had been a proper backlog, the history problem
would have been thought about before the sprint, informal discussions
had been held, etc. There chances of designing a good history system
quicker, would have increased.
In the next sprint we will have a proper backlog, and I am curious
how that will effect the sprint.