i wish the dev shop I work for did half those things...

Everything you've listed is extremely valuable, but often a few of the items are overlooked or not considered to be a part of a classic CI build. I believe this to be an example of local optima - many dev shops are focused too much on the dev cycle (which does need to be addressed) and forget about engaging the rest of the delivery process.

I think we need to stop just thinking about CI builds and start thinking more systemically - its really about developing a consistent and automated Deployment Pipeline.

Currently our shop has the following stages (each being a separate, and serial "build" w/in our TeamCity setup):

1) Development

2) Deployment

3) Testing

4) Review

 -> the artifacts from Review can be pushed as public bits

And while the stages could be better named to illustrate what they really do, the point is, at each step along the way we do particular things to get the code closer to a "ready to ship" state. You know what, I'll check w/my boss and if its cool, write up a blog post to explain more about our pipeline and why we believe its important to invest in such an apparatus.

This sounds great....would love to be doing even some of this, but where to start!!??  Can you point to any online step by step guides??

btw...running SourceGear vault as our SC provider.

Absolutely agree with the list.

In fact, that's what so far I've been having in my projects. That's the only way to be able to deal with complex development projects anyway.

Just a few "must have" items to add to your list:

* Generating documentation based on the code and updating section of web site to the latest

* updating (if forced) auto-update package deployment (so that customers will get their smart clients updated automatically, when they launch application

* Running code quality tests against the codebase (enforcing project-specific design patterns and guidelines)

* Grabbing and publishing performance statistics of some long-running method operations (i.e.: especially long-running DB queries)

Here are some more concepts (nice-to-have):

abdullin.com/.../some-integration-recipes-for-the-efficient-development.html

VS2010 and .NET Framework 4 In-Process Side by Side Part 2 - Common in-proc SxS scenarios Achieving Speedups with Small Parallel Loop Boodies Visual Studio 2010 Performance Part 2: Text Editor A scalable reader/writer scheme with optimistic retry Deploying

I agree with the exception of the list being "extensive" <grin>. I consider scripted deployment as a mandatory component of the process.

Without scripted deployment you can't realize the value of the CI artifacts. Scripted deployment identifies platform specific configuration values, allows all team members to frequently deploy and do integration testing, provides repeatability for QA/Operations deployments, provides the basis for automated deployment

Nightly scheduled build that applies database changes to a copy of the production database (or schema) is also an excellent confidence building exercise.

/jhd