"boltcapt" <bol### [at] aolcom> wrote:

> I hope this captures what you were looking for. If not, let me know so I can
> refocus.

I think that after going through all of that, you may have added some technical
points and terminology/references of which I was not aware.
Those would be good to have in the documentation.

The larger project, is of course, implementing the fixes discovered for:

The unrendered portions of the revolved spline, due to those portions of the
interpolated Catmull-Rom spline being negative

The first control point not being able to be placed at the same Y value as
another control point (possibly raise a warning and maybe auto-fix)

And the false positive bounding tests that call unnecessary root-solving.

Related to root-solving, there would be the issue of possibly implementing
tighter bounding by the "bounding by proxy Bezier spline method" mentioned in
the thread.

That last could possibly solve some related problems with the sphere_sweep {}
object.

The deep-dive into the how and why of the sor {} and reimplementing it all in
SDL along with various fixes too a lot, and I don't know that currently have
enough mental energy and, more importantly, uninterrupted time to scale the
steep learning curve of learning how to competently code in c++, set up a
machine, and do all of the iterative compiling of edited source and bug-hunting.


So,
The proposal/request would be that you could identify the exact files that need
to be edited, and write the fixes based on my SDL versions, and I could follow
along and try to make sure it all jives with my copious notes,

or

I could edit the source, and you could just do the compiling and supervise the
edits "No, you shouldn't do it that way because" or similar dev code coaching,

or

some mix of the two approaches.

The end goal would be to produce new edited source with excellent commentary and
an eye toward allowing any future changes and extensions without having to
discover fire, invent the wheel, and reverse-engineer the sor {} source code all
over again.

It sounds like a lot, and probably is (isn't everything), but I also think that
since most of the theoretical heavy lifting, and implementation in SDL is
already done, that it might also be a bit easier (in the broad-stroke sense)
than it seems at first glance.

Maybe we can start a new thread and we can discuss the details of what needs
doing and go from there.   I can outline things, you can ask questions, and we
can hammer out a game plan before even touching any code.

Thank you,

Bill "Bald Eagle" Walker

Post a reply to this message