Re: p5ee xp

Guys:  I've been on vacation for a couple weeks, but I saw that there was an 
interesting exchange regarding XP and P5EE.  I'd like to point out that using 
XP in the context of open source development has been covered by several 
writers.  Perhaps they could inform your decisions.  Here are a few links:

* An article on requirements solicitation in open source projects
http://www.stsc.hill.af.mil/crosstalk/2000/jul/henderson.asp

* A PDF document discussing distributed development using XP
http://www.cs.wustl.edu/~mk1/DistributedXP.pdf

* A Wiki on combining XP and OS
http://c2.com/cgi/wiki?CombiningOpenSourceAndXp

* An article on using XP and OS by the infamous "chromatic"
http://linux.oreillynet.com/pub/a/linux/2001/05/04/xp_intro.html

* An advogato article that draws some obvious parallels between XP and OS
http://www.advogato.org/article/202.html

I'll save you some time and summarize:  While pure XP is almost impossible to 
use with OS, modified versions of it will work just fine.

It stands to reason that this would be so.  Under its surface, XP is really 
just a set of rules to play the software development game with.  These rules 
create dynamic processes that cause us to create flexible, robust software.  
Open Source software, or more accurately, _distributed_ Open Source software 
development, suffers from some problems of interpersonal communications and 
(on occasion) lack of a true customer, but can still benefit from XP's 
dynamics.

Also, in _A Practical Guide to Extreme Programming_, (ISBN 0-13-067482-6), 
the authors cover the problem of having no customer.  Essentially, the person 
or persons who originally conceived of the product must proxy the 
To-Be-Determined customers.  This represents a risk to theproject, as 
mentioned by others on this list, in that there will be less separation 
between the programmer team and the customer team.  But if you consider the 
fact that P5EE's customers will be _programmers_ themselves, this risk should 
be minimal as long as the "visionary" can keep an objective viewpoint.  What 
you need is a good test-bed application to prove P5EE's worth and usability, 
and I believe I saw that you'd started one.  This would also help you enforce 
your standardized interfaces as your internal architecture changes.

It would be rather silly to say that you must _always_ have a customer, since 
that would prevent software companies from creating software for sale.  Just 
because no one asked for it, doesn't mean no one wants it.

My own opinion is that you can only benefit from trying to use XP.  Not pure 
XP, but P5EE's version of XP.  Take what you can, and leave the rest.

jpt

• Re: p5ee xp by James Tillman