Thanks Alex,
One thing I'm looking for is validation that there is a need in some
applications to construct an object graph of sorts to describe the
model beyond an array of records to be CRUDed. Surely I'm not alone
here?
If that's the case, is there a best practice for going from value
objects delivered from a service, to first class, intelligent model
citizens in the domain? One thing I've found confusing about the
Cairngorm approach is that models and value objects seem to be
synonymous.
You thoughts?
Cheers,
Lach
On 05/12/2006, at 11:13 PM, Alex Uhlmann wrote:
there many ways to use Cairngorm. I agree with you completly, it's
of vital importance to refactor the right functionality from both,
views and commands into model objects that mean something to your
use case. Then, this functionality can also be easier unit tested. …
However how your model is going to look like depends on your
specific needs. Cairngorm doesn't tell you how to design your
custom model or your custom view. It's the infractucture around that.
My blog entries around that, which Douglas pointed out, have the
indent to show one possible route to go towards this direction.
Nevertheless, it's difficult to show that in examples, since
there's a tradeoff in completeness and complexity vs. easy to
understand examples ....and most importantly...time. ;)
