Hi,
What is your input on project/components organisation (using Jira/Greenhopper):
We roughly have 8-9 teams (grouped by IT knowledge) (backend,frontend,ops,qa,design etc etc) and around 14 projects..and by projects I mean standalone "modules" that have own release cycle, own versioning and sometimes interdepend (example: engine, payment system, mobile server api, website etc)
1. User centric projects - single project represent specific team (backend developers for example)
Pros as I see it:
- people can easily see what they work on. It'll be a bucket for all work that a team can handle
- possibly easier to follow stats and reports (still not into reporting and greenhopper specifics) but ultimately I'd like to use boards to clearly help devs see whats coming.
Cons as I see it:
- Single story need to have N number of copies..one for each project and relate those. Not sure if using relates to link will be possible to aquire good logging - for example how much time spent on QA, whats overal time per feature etc.
- more difficult release cycle - in this java team project will have N components (different things they work on) but those does not relate to each other and more often than not have own release cycles and will end up having Versions like somebackendlib-1.0.0 and anotherbackendlib-1.2.3 etc etc which might end up hard to follow.
2. or Product centric projects - single project represent specific product worked on (example Platform - which contains backend, frontend, design tasks etc)
Pros as I see it:
- single issues - assign multiple components (frontend/design etc). Still for complex stuff I'll need issue linking so that I can generate reports /stats.
Cons as I see it:
- people will jump between multiple projects to work on
- defining priorities across projects might be troublesome ( maybe Greenhopper will help nice here )
- stats and boards might be more difficult to comprehend create to answer questions like...what to be worked on now...
Would you rather make user centric projects or I am making all the wrong assumptions here.