I’ll start with a provocative question: 

 

Should the SDK team put samples in the SDK or only make them available online via MSDN?

 

With the follow-up question: 

 

Where do you prefer to get your sample code in the first place?

 

 

First some background by way of example: 

 

The majority of the samples in the .NET SDK are written by the programmer writers who also write the reference topics and conceptual content in the SDK doc set.  The programmer writers often weave these samples in with the various topics they are writing. These samples come into the SDK through its development phase via the documentation.

 

The SDK team also gets a significant number of samples from the technology teams themselves.  This is great, since the folks who created the APIs probably know best how best to use them, right?  We get some really meaty samples and Quick Starts from the technology teams.

 

As we approach major milestones (such as betas), the SDK team is often forced to beg and bribe for more and better sample code from the product teams.  We get the usual response that they are too busy getting Whidbey ready to ship and therefore don’t have time to work on samples and can they give us their samples after (choose one:  beta1, beta2, RTM)? 

 

The problem is that the SDK is a product too and must hit all the same quality criteria and ship milestones as any other part of Whidbey so this procrastination by the product teams usually results in very few samples coming from the people who developed the APIs in the first place.

 

So what’s an SDK team to do?  Lately I’ve been wondering if it would be better to not have any samples in the SDK (except those needed by documentation topics) and ship all samples outside the SDK via online download.  That way we might be able to get more samples from the API teams later in the product cycle. 

 

Is this crazy thinking?  Please let me know!