I'm probably going to be one of the most finicky person here about
this kind of stuff, but I actually don't want any of those elements
converted to a point.  Implicit conversion is great when there's a
logical connection between the data (ie curve -> interval is the best
I can come up with at the moment).  With the conversions you listed,
there really isn't a logical, consistent reasoning why it should be
one of those options over another.  The fact that you haven't even
listed what the possibilities could be for mesh, and "etc. etc" for
Boxes leads me to believe that there's just too much gray area to make
a consistent conversion that winds up being usable in a wide variety
of situations.

To me, its better to have that kind of conversion fail, and A)
hopefully the person understands why its failed* and B)hopefully there
are the tools to extract the desired points in an EXPLICIT manner. In
most of those, there is an way to extract the desired points, with
Breps and especially meshes being on the lower end of that scale.  I'd
rather the time be spent on better tools to extract data from those
kinds of data then deal with how they convert to a point...but that's
just me.

I can some this up with using the age old saying, "When you assume,
you make an 'ass' out of 'u' and 'me'."  To me, there's just too much
assumption in those conversions to be worth letting the conversions be
done implicitly.  All that being said, if it had to happen, I'd vote
for Bounding Box center all the way around....it works for everything
and its consistent.

Forever Opinionated :)
Damien

*I am fully aware that more times than not most people will not
understand why it fails


On Jan 23, 1:33 pm, David Rutten <[email protected]> wrote:
> Does it make sense to anyone to have default conversions for:
>
> Curve -> Point        (start point? mid point? end point? centre of
> bounding box?)
> Surface -> Point     (centre of bounding box? centre of UV domain?)
> Brep -> Point         (bounding box centre? Volume/Area centroid?)
> Mesh -> point        (?)
> Box -> Point          (... etc. etc.)
> Twisted Box -> Point
> Generic Geometry -> Point
>
> And, if so, how would you expect it to work?
>
> Thanks,
> David
>
> --
> David Rutten
> [email protected]
> Robert McNeel & Associates

Reply via email to