Extreme programming

[CountZero]

I just read an article (mandatory free registration, a la NY Times) in the Dallas Morning News this morning about a practice known as extreme programming (or "XP") where companies will make 2 programmers sit at the same keyboard and take turns talking out code and typing it out. Programmers choose their project from a wall of index cards. In addition to all this the programmers also have short daily meetings to demonstrate their progress.

Anyone ever seen/done/heard about this? I found it interesting but I'm wondering about what programmers think of it. The article neglected to mention that.

[JanErik]

I do it on my job (I work on a research project on XP, for my university).
Think it works quite well, actually. What I don´t know how to do, my coworker knows, and vice versa.
We are not making commercial code, just programming as lab-rats on XP.

[compu-tom]

I'm working as a developer and we are currently getting deeper into the XP thing. Standup meetings (AKA scrum) are very effective for a long time now. And pair programming (which is only one part of the whole story) can be very effective, too (but not every problem needs a PP solution).

I recommend you reading
- http://www.extremeprogramming.org/
- http://www.xprogramming.com/
- http://www.martinfowler.com/articles/newMethodology.html
to get an overview.

The most important thing is: all (your team members, management, and also the customers (if you do project work)) have to commit to work the XP way.

[Yarrick]

im currently taking a class in this particular subject. were programming in java, and doing cvs checkouts, pair programming, refactoring, testcases with junit and so on. next month were starting a project. seems nice so far.

the part i like the most i that the test cases are written before the code, so after i changed something i can simply run all tests and i if it all works. gives confidence.

[gsfgf]

That's what the computer club at my school does. Each group gets the same problwems w/ points assigned to them and the group w/ the most points at the end wins.

[swagr]

I read an interesting view a while ago.
I don't know where I found it, but the author made a good point. Most things that aren't questionable have no advocacy. (e.g. everyone agrees that eating , breathing, bathing and sleeping are important yet these activities have no advocacy groups).

Anyway, I think one of the reasons XP is questionable is that there exist examples of very successful non-XP projects.

I've been working on a clean-room drop-in replacement of an expensive propriatary 3rd party API. It's a solo project, and we already have a mature product based on the very API I'm re-implementing. The short story is: after a year it's functional and stable. The only real bug found in the last 3 months took a few hours to hunt down and fix. I'm ahead of a similar open source project that has been in the works for 3 or 4 years (I didn't copy any code). The product I've replaced has bugs that I don't have (althought I can emulate the bugs for grater compatability). So I think I've done OK. And I did it without XP.

Could I have done better with XP? Maybe.
Could XP developers do better using a different system? Maybe.

Don't get me wrong, I don't think that XP is "bad" or flawed, I'm even curious to try it. But beware of anyone "preaching" to you that you will definately see an increase in productivity or decrease in bugs.

Anyway, take a look at this for an alternative view:
http://www.softwarereality.com/lifecycle/xp/case_against_xp.jsp
_________________
RUN!