[ODE] some quick notes on 0.6...

[John Miles]

I'm just being a user advocate, really.  I don't have an application for the
convex primitive, so it's not a problem for me one way or the other.  I'm
simply suggesting that a primitive that limited has no business in the main
distribution.  It makes for an incredibly-poor "out of the box experience",
no pun intended, when someone downloads ODE and runs test-boxstack for the
first time.  Few will bother to dig through list archives and manuals after
an experience like that.

Complaining about the lack of callback functions (wow, way to slow down
terrain collisions by 500%) and Y-up/Z-up limitations (wow, a major
inconvenience for at least 0.05% of the users) in an otherwise rock-solid
primitive, while accepting a convex collider that while academically
interesting isn't usable for *any* practical purpose, falls into the "wow,
just wow" category.

Please don't take all this ranting personally: it's strictly an advocacy
exercise.  ODE 0.6 still works great for my own application, and there have
been some very real improvements that are much appreciated.  But the
maintainers need to decide whether they're building a highly-experimental
R&D project or a practical (if limited) tool.  Nothing wrong with either,
but you can't do both.

-- jm


> Had you taken care to read the list entries regarding Convex primitives
> and the note on Changelog (currently only convex-sphere and convex-plane
> work), you would know better to expect
> it to work, or think that its ready for prime time, there is a reason
> its 0.6 and not 1.0.