hm, the way it currently works is rather perfectly awkward, if that was intentional i would like to understand the idea of it.
that option unfortunately does also fail on the mac version, there you can pull in whatever direction you want, it will extrude negative… if i understood you correct, which might not be the case because i only slept 3 hours
you did not understand my non native speaker english - or it s not enough sleep, or not enough coffee
trying to define the extrusion direction by the current mouse position seams to be the approach mc neel is targeting, or at least what I understood @pascal is voting for.
so your quote of my post is already a interpretation of what pascal wrote earlier ;-(
i think a pure numeric input independent of the mouse position should be possible. cplane and curve direction can define + / - direction, a (missing) visual indicator could show the +/- direction, similar to _dir command or other UV indicators…
identifying direction by mouse might be an additional option - but should not use the boundingbox-Center as this is counter-intuitive for large curves / complex projects as i show at this post
(with not much resonance)
ic too little coffee then, or just too little sleep…
anyway in the effort of getting things done quickly and with the chance of not having rhino UI understood fully i usually accept awkward unwanted results and manually change them not thinking about it further.
but if one uses such commands a lot repeatedly then such issues naturally become a problem and i am also getting tired reporting quirks and getting very little to no response.
rhino is messy and there is not a very good attitude in trying to clean it all up (or it at least feels like it is not moving anywhere or in turtle speed). it sometimes takes 20 people complaining non stop till a little rock which is in the way gets finally whacked out. ok but all that belongs in the UI overhaul which is now shut down