More discussion about MS and agile methodologies - ISerializable - Roy Osherove's Blog

More discussion about MS and agile methodologies

 
Jay hopes that my comments about adopting XP at Microsoft will never be heard.
 
My answer to that would be hopefully never. In his post, Roy ponders how the XP process would work on projects as large and vast as what Microsoft produces. Is he really serious? Can something like Word be developed using XP?  I can't possibly see how.
 
Well, Jay , XP has worked on a very large project over at Chrysler a few years back (many years back actually).
There's no reason to think that due process will not lead to success on even bigger ones. It's been working on many a projects around the world and for the past couple of years has been gaining more and more recognition as something that's definitely worth looking at and learning from. I myself have found that even taking *some* of the concepts it uses provides for huge benefits.
And you know what? I'm sure MS is doing Unit testing somewhere behind the iron curtain. Granted, it's not as “evolved” yet, but its definitely something every good developer will want to know about. And MS is filled with those.
 
More from Jay:
 
Microsoft needs to stay out of this area. They are best suited at making tools.
 
Making tools? how about Operating systems? programming languages?
 
I think it would become very difficult if Microsoft started shoving processes down my throat.
 
What have you been programming with so far in your N-tier applications? a few words: DNA, MSF, and parts of those are all in most of the development methodologies and architectures existing today. Like it or not.
 
And heaven forbid that they should start tieing features of the tools with processes.
 
Hardly. Tyeing in and facilitating certain abilities are two separate things. You won't have to do XP on Whidbey, but you could much easier than in previous versions.
 
 
Published Friday, June 04, 2004 1:07 PM by RoyOsherove
Filed under:

Comments

Friday, June 04, 2004 2:00 AM by TrackBack

# More from Roy on XP

Friday, June 04, 2004 7:32 AM by Martin Liversage

# re: More discussion about MS and agile methodologies

Actually the infamous C3 project at Chrysler was cancelled before it was complete. Reading the book Extreme Programming Refactored: The Case Against XP by Matt Stephens and Doug Rosenberg should provide some interesting insight into some of the problems with XP. I haven't actually read the book, but there is an interesting review of the book on /.: http://books.slashdot.org/article.pl?sid=04/03/24/2118215. I'm not saying that XP is bogus, but reading just the review of the book provided me with some quite interesting perspectives on XP.
Friday, June 04, 2004 7:38 AM by Roy Osherove

# re: More discussion about MS and agile methodologies

Martin: The cancellation reasons were, AFAIKm not related to how successfully it really was in terms of what project manager would consider successfully - on time and on target.
Still, the case against XP is worth a read. No one should go blindly into anything.
Friday, June 04, 2004 8:46 AM by Jay Glynn

# re: More discussion about MS and agile methodologies

The cancellation was due to the fact that they couldn't deliver the goods. If Beck and company want to call that a success, then more power to them. If it walks like a duck and talks like a duck, well you know the rest.

Operating systems, programming languages, Office etc are the tools that I use to accomplish what the business needs. Yes they are tools. The applications and systems that I produce are the tools that the business uses to gain (hopefully) an advantage over the competition.

DNA is another name for n-tier which is of course an architecture for designing systems which of course has absolutely nothing to do with the process of designing the system. XP is about the process. MSF is Microsoft's process for building software. It is an OK process, doesn't fit in our environment but then it is not built into their toolset. It isn't put in my face every time I start up Studio. While I think that some of the tools that are being offered in Team System are nice, Microsoft does not know nor can they know the type of environment that I develop in. Process should not be dicated by toolsets. They should be supportive of "any" process and not get in the way. As long as that remains true to form, then I don't have any issues.

BTW, my name is Jay, I don't know who John is, but he may be mad at you for misquoteing him ;-).
Friday, June 04, 2004 8:56 AM by Roy Osherove

# re: More discussion about MS and agile methodologies

Jay: OOps! fixed your name..

"MSF is Microsoft's process for building software. It is an OK process, doesn't fit in our environment but then it is not built into their toolset. "
So is XP. It won't be built into the toolset, but it will be accomodated by the toolset easily enough.

Friday, June 04, 2004 12:36 PM by Keith Rowe

# re: More discussion about MS and agile methodologies

Actually, Microsoft is doing a lot of work in process and tools right now. Last week, my team announced our new product, Visual Studio Team System. This provides a suite of tools targeted across the application development team - testers, architects, project managers and developers. You can see more at http://msdn.microsoft.com/vstudio/teamsystem/default.aspx.

As part of this effort, we are reworking the Microsoft Solution Framework. There will be both agile and formal versions of the methodology. These methodologies will direct the behavior of the tools in significant and interesting ways.

But, it will all be configurable. You'll be able to select which methodology you want to use on a given project. You'll be able to create new methodology templates that will drive the toolset the way you want. I'll be discussing this further in my blog in the months to come...
Friday, June 04, 2004 1:16 PM by nospamplease75@yahoo.com (Haacked)

# RE: More discussion about MS and agile methodologies

I read the XP Refactored book and I think it does a pretty good job of highlighting some of the faults of XP. Until that book, I was Gung-Ho XP, but alot of XP practitioners are waving it as the hammer to hit all nails.

XP has its time and place, but it isn't adequate for much larger projects as even some of its creators acknowledge (the book makes references to these comments). However, I think some of its practices have a place everywhere (Unit Testing, Refactoring).

XP also violates the principle that a methodology should try to make best practices (as much as possible) the path of least resistance, since that's what people take in general. XP requires extremely high discipline, and that's hard to maintain.

The book is worthwhile in helping to build a more pragmatic version of the XP methodology.