Hey Daniel,
This design intent part is basically a part just as all the other parts.
You may implement it as a skeleton part as well, but this is not
essential in the concept. The concept is that you seperate the driving
geometry from the details. Driven geometry has to be rock stable. I
think you do not want your driven geometry (the actual design intent) to
go into resolve mode when you make a rather small change in your design
concept. So, the design intent part is a part as all other parts, but
only contains the essential features of your part.
Another consideration. When you succeed to limit the complexity of your
design intent parts, and you have rather complex assemblies, you will be
able to play with your assemblies and consequently regenerate all driven
design intent parts within a reasonable amount of time.
Drafts and rounds are most of the time not essential (if technology
would allow, we would cast without drafts, isn't?), and are heavy to
regenerate. So, keep them out of your design intent parts, but add then
to the third stage, the rough parts. The second stage parts are parts
where you add material allowance to your design intents. The fourth
stage are the finished parts, where you publish/copy and solidify the
rough geometry into, and publish/copy but not solidify the design intent
into. This second publish/copy is ment as reference for the finishing
(because you do not want your finishing to go into resolve when some
silly round fails).
We add the design intent and the finished parts to our assemblies. The
other parts stay out of it. When you implement your design intents as
skeletons, the mass properties of your assemblies will disregard the
design intents, and stay correct. And with simplified reps, you can
easily switch from design intent assemblies to full detailed assemblies.
OK?
Regards, Hugo.
(I think I explained the above earlier, so if you search the exploder,
you should find it in other words ...)