Pretty much: I'm in favor of the DD being used to provide values to objects that do the work of working with data. I don't know that there is always a large benefit in the deployed application using the metadata live, in most cases. I know for sure it is invaluable during development. And I know for sure that a generator approach is the opposite of what I want -- although if the Waterfall development model worked, and there were to be only one iteration of the database and associated classes based on The Plan, I could see being in favor of the generator method.

The book Boyd, about the USAF Colonel who devised the current strategy of warfare, is at its core about the need to maintain acceleration in a curve, whether one is in a dogfight or writing software in a changing market. The company I work for was recently chosen by a large international as its vendor for our type of software for one of their new initiatives. As the director of the initiative told us, we were chosen not for our existing capabilities alone, but because of our agility which serves their cultural need to make software work to fit their processes, rather than v-v. The competitors, all much larger than us, could not offer them that, because it wasn't part of their culture.

The book was popular at Redmond a few years back, so someone there recognized the need to think differently. I'm sure it had an effect. But big ships take a long time to turn, and in a big organization there is the additional inertia caused by vested interests (my bonus, my promotion, my personal investment of effort) in the current directions.

best,

Hank

>
>I think what Hank means is that more functionality should be moved to an active DD
>- and at least compared to most Dotnet examples I fully agree.
>
>regards
>
>thomas