Hi Michael,

> That would then make blends like this not work:

Ok I see, I was thinking something closer to the end I didn't consider that situation, then you would need a flip button or a click in space to flip the blend.

> Doing too much stuff automatically
> can sometimes cause problems like that.

I see what you mean, you always come up with different scenarios, I guess that's your job :) The trim idea was just a fleeting thought, it sounded good at the time.

Are you actually considering enhancing curve blend or are you just open for ideas?

Cheers

Hi Danny,

> Are you actually considering enhancing curve blend or
> are you just open for ideas?

Well, always open for ideas and if an idea allows for new functionality while not really adding any complexity to the UI (like not requiring extra stages or steps or things like that), then I tend to like to bump the priority of working on it up.

Now that I go back and read Marc's message again, I see he wrote about "in viewport" controls which I think means the same kind of point dragging that you mentioned I just didn't understand it initially.

Having draggable points for relocating the blend seems like it falls in that category (low UI overhead) well - it's similar to how Loft works like you mentioned. Similarity to other commands is good and it could fit in with the current "Adjust blend parameters" stage, so it wouldn't require a new stage to be added to the command, so that's also good.

Right now in curve blend, at that last stage if you click in the viewport on one of the blend inputs it will flip the blend to the opposite side of the curve, maybe that part could be removed if you could locate the blend starting point by dragging it.


The part about throwing trimming in as well seems not quite as promising initially - that seems like it could have some side effects. I suppose the side effects could be mitigated by having a checkbox option to enable the trimming or not, but then that option would be kind of weird to see when blending from the ends since it wouldn't do anything for that case...


- Michael

Hi,

It would be quite awesome to be able to drag points to relocate the blend!
The click to reverse feature sounds great.

For the trim aspect, maybe it could just trim and keep both side of the curve?

Also there could be some visual indicator suggesting there can be an interaction possible there, maybe points could appear at both ends of the blend...

Marc

Hi Marc,

> Also there could be some visual indicator suggesting
> there can be an interaction possible there, maybe points
> could appear at both ends of the blend...

Yup, that's what I was thinking - that would again be similar to how Loft works currently when lofting between 2 circles. Points are displayed at the seam locations where you can drag them to update them.

So that would make sense for Blend adjusting as well.

Still not sure about automatic trimming though.

- Michael

I was thinking about this and it could be cool to have these interactive points in 2d fillet also.

Marc

Hi Marc,

> I was thinking about this and it could be cool to have
> these interactive points in 2d fillet also.

Unfortunately that's probably pretty difficult - curve fillet currently works by calculating the result from 2 curves and a radius value. Trying to drag one point to manipulate the curve fillet would probably involve some kind of reverse calculation from that which would involve writing a completely new solver.

- Michael

Maybe a center radius point that could be draggable?

The vector scribe plugin for illustrator does something like this:
http://www.youtube.com/watch?v=uZHtamBiBHo

Marc

Hi Marc,

> Maybe a center radius point that could be draggable?

Right, but that's the one that requires a different solver.

Instead of solving the problem "find the points on the 2 curves to make a fillet arc of the given radius", it's more like "find the points on the 2 curves to make a fillet arc with the center closest to the given center point, with radius being calculated".

It's a different problem to solve and would require a different solver mechanism to be written to calculate it - that's certainly possible but it's a fair amount of work to do it. So that could be possible in the future.

- Michael

Oh, I think I understand, but then again I have a bad cold. :-)

I guess It's a kind of mechanism that looks simple but takes a lot of work to implement.

Thanks,

Marc

Hi Marc, well it kind of depends on what you mean by "a lot" ... :)

It's probably a couple of days worth of focused work, but that kind of puts it into a category of stuff that becomes more difficult to schedule.

It's something that's simple if you only have a fillet between 2 line segments of course, but with 2 arbitrary wiggly curves rather than just lines it's a much different scenario to deal with.

It's a good idea though and I'd like to work on it at some point, it's just difficult to know when.

- Michael

It looks to me more like a slider for the radius of the fillet.