So to cl;arify I think DayTrader may be mis-characterized as only a benchmark.  I think it is used for a variety of functions of which only one aspect is performance.  the other aspects are testing infrastructure, sample application (which includes deployments plans for various pieces of functionality, etc.).  

With that in mind perhaps we should change it from performance benchmark to something else.  I'm concerned about pulling this at the very end.  I agree that we don't want to put out a sample that can be used inappropriately but in the end we don't really tell anyone what they can do with the software.  I'm concerned about making changes to DayTrader because of commercial positioning.  For this reason I'd rather add the comment about relevance of the runtime mode.  I think simply adding the comment is adequate.

If we're going for best practices I don't think anyone would say the app as it is written today is a best practice but more of a swiss army knife of tools.

It sounds like there are varying viewpoints so perhaps we let the discussion percolate and put it to a vote.

On Oct 3, 2007, at 12:25 PM, Christopher Blythe wrote:

ah.... now you're pointing out the distinctions between primitives and a real application "workload". currently, the web services within daytrader are presented in the context of a "workload" not a primitive.