These are questions that came up from the XpUniverseBeyondXpWorkshop
This is from Swee Lee:
For large distributed projects:
1. the analyis process is more complex and engaging and requires dedicated skilled people. The level of dedicated attention, coupled with potentially the lack of skills of customers in area, would typically perclude the use of a programmer.
2. The 1st iteration, for a 'mid conversion' project tends to be longer than 'typical'. As the team eplores and learn an existing system. I am unsure where to strike at to balance 'early' feedback vs 'usable' code.
3. A big picture is particularly crucial to guide the metaphor or maybe more appropriately help produce "a system of Names" as Martin F calls it. This is critical to avoid potentially wasteful refactoring.
4. I am stll unsure how you would facilitate Closure for you and your customer. Who has the mandate? Argubaly you have since all the base requirements for your client have been achieved this last round. The rest is to "packagise" the system. Is this right?
5. How is knowledge going to passed on to the support team?
6. I am a littel unsure too about how you could transform a custom product into a "package" product for sale when modularity is not part of the design(since the client would not need the concept), eg how you modularise into say : Fixed Assets, Invoicing, Debtors, Lease transactions, Reporting, etc. Or has these been expected and completed at the front of the project.
This is from Mark Kilby:
Borders of XP (future topics of discussion; where XP is not well-defined or may not work in current incarnation)
this is addressed on XpUniverseBeyondXpWorkshop
What we didn't discuss...
Our typically team composition
--AmrElssamadisy (8/6/01)