Follow-up Comment #1, bug #67866 (group groff):
I can't seem to reply via email to Savannah tickets anymore.
Here's the comment I tried twice to submit by that means.
-----------------------
Hi Doug,
At 2025-12-29T16:10:27-0500, Doug McIlroy wrote:
> For splines, pic identifies compass points (.n, .ne, .e, ..., .nw)
> with endpoints (.start or .end). This makes no sense. It would be
> better to declare their use illegal. Alternatively, they might be
> placed on the spline's bounding box.
Thanks for the report.
I have three thoughts on it: one a query for you and other pic experts,
one mathematical, and the last more a musing for myself and other
committers, aspiring or actual.
I am still a rookie with _pic_(1), so I appreciate your patience.
1. I find your proposal of attaching the compass points to the bounding
box of a pic object intriguing. Do we have an exiting syntax for
that that we could generalize to splines and other objects?
For example, when dealing with a circle, primary (orthogonal)
compass points are identical whether the bounding box is used or the
object itself--but the secondary compass points could sensibly
attach to either, and each might be desirable for varying purposes.
Can we already make this choice for circles? How is it expressed?
2. I wonder how the new reference points for vertices and midpoint
should apply to splines. As I (poorly) understand splines, any
vertices apart from `.start` and `.end` are not actually _on_ the
object. However, they _are_ control points determinative of its
shape, and I think it might be useful to be able to make connections
to these control points. (I can easily imagine myself crafting
several example splines with control points thus revealed, to teach
the concept to myself and others.)
Is the midpoint of a spline well-defined? I once surprised a
roomful of undergraduates with the claim that there is no formula
for computing the circumference of an ellipse.[1] If a length is
not well-defined, half of that length also cannot be.
3. If resolving this ticket implies changing the YACC grammar of pic,
we could use that change to exercise/test/prove any fix we devise
for bug #67808.
Regards,
Branden
[1] I _think_ elliptic curve cryptographic algorithms arise from this
property, but I can't swear to it.
> {savane: user = 108747; tracker = bugs; item = 67866}
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?67866>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
