Intersect Tool Missing! Available as separate plugin.

Dimitri · 04-03-2013, 09:29 PM

Sure, the rmb, lmb, mmb solution too is not bad... it is ok. The reason I would prefer the sub-menu solution is just because I believe that visual commands, which make things self evident, are always more easy to work with. Commands connected to mouse moves may be prone to be forgotten (and in the case of the newcomers there may be a difficulty to find them).

As to the other related matters, however, one very important one is having an option to keep the existing hard edges and make appear as hard ones the newly appearing seam edges (as was the case in many previous ML versions).

The boolean seams do not constitute regular loops, so it is impossible to select them with one move as loops... and, therefore, in the case they would not appear as hard edges by themselves there would be the need for a very tiresome process to select them edge by edge so as to apply to them the 'hardness' property.

ggaliens · (This post was last modified: 04-05-2013, 04:53 PM by ggaliens.)

Been re-working and reworking HARD-EDGES

And I think I have a very very good approach now Dimitri.
Release momentarily.

http://www.youtube.com/watch?v=f0p9Lbatqq8

Dimitri · 04-05-2013, 09:47 AM

Ok Mark... waiting to see and try it... Smile

(btw the youtube link under your message is connecting to a login page, not to a video streaming one)

puzzledpaul · 04-05-2013, 10:46 AM

What exactly are you trying to get the boolean ops to return ... re hard edges ... just the lines of intersection?

(I saw several issues after a quick play)

Re 'testing' ...rather than just making stuff, I'd suggest

single primary / single secondary
multiple primary / single sec
single pri / multi sec
multi pri / multi sec

all above with different combinations of pre - existing hard / soft edges on primary / secondaries .. objects ... and in the case of multi-object scenarios, have some objects with hard edges, others not.

Across all options, of course Smile

Whilst not hard edge related, screen still goes blank with a multi primary / single sec situation for intersect (prims not combined) ... upon execute

pp

ggaliens · (This post was last modified: 04-05-2013, 05:01 PM by ggaliens.)

PP ... if you wanted to set up 5,6,7,8 test case models ...

I would use them in my testing. Otherwise .. I'll check more of your cases cited sometime soon.

What I did last night was to use Three menger cubes level 2 ... and arrange them as haphazardly intersecting. I thought it made for a fairly stressful case. I know there are still several if not a bunch of outliying cases.

Yes ... it really should just HARD-EDGE the intersection lines ... or "NEW EDGES". Old edges that have been cut ... they should not be selected. So the problem is how to GRANULARIZE the points just the right amount so that an edge with only geometry attributes and no NAME ... can still be matched to SIMILAR ENOUGH other edges. How to do that has been a bit puzzling for me for having used too many approaches.

I need to formalize the approach and perhaps even "publish it" for anyone who might want to comment.

Also ... thanks for the more feedback.

puzzledpaul · (This post was last modified: 04-05-2013, 08:02 PM by puzzledpaul.)

Whilst the setup in the vid seems fine, it's having problems with simpler stuff - basic prims - which is what I often use for initial tests.

ie Cube with sphere ... move sphere so's its center is approx over a cube corner.

Subtract sphere from cube ... ALL new edges are hard.

Subtract cube from sphere ... some of the edges that should be hard ... aren't ... and maybe the ones describing the cube's corner shouldn't be hard either ... if it's only lines of intersection you're after.

pp

Edit
Using your setup / objects as in vid
Doing a union produces no hard edges
also

Making the original 12 edges of the cube hard ... then doing a subtract ... as in the vid ... results in ALL edges being hard.

Edit 2 ... and I know this'll sink like a lead balloon Smile

... but I wonder about the (supposed) 'need' for hard edges at all ... yes, I know why they're wanted ... to help offset selection issues associated with the inherant problems of using booleans.

But I also wonder about the circumstances and (selection) techiques being used by ppl using booleans.
Since Shift L returns boundary edges of a face selection and there are various ways of selecting regions of faces ...

Also, I suspect booleans are often used with symmetrical objects ... often non-organic, methinks ... and thus I (select similar) could be used ... I have already used both of these approaches with some of the tests ... albeit on relatively simple objects.

However - not having to figure out how to produce hard edges only on lines of intersection (for all manner of complex circumstances) would free you to sort out other stuff, maybe.

Without knowing how ppl have gone about selecting (manually) hard edges - on realistic (not test) models ... I'm not in a position to say much more ... + have just been told food's ready Smile

ggaliens · (This post was last modified: 04-05-2013, 10:48 PM by ggaliens.)

With two new PP test cased tried in video (just now with release just now).

http://www.youtube.com/watch?v=VocO9qGIo...e=youtu.be

puzzledpaul · 04-05-2013, 11:29 PM

Well, have just d/l the latest build (but not tried) + watched the post 67 video ... and I think we're on slightly different pages Smile

This'll be brief, as am just about to go to bed.

The cube with sphere, roughly centred on a cube corner is correct setup, re what I tried ... and the result is as I got (with previous release ... haven't tried latest) ... but what I'm querying is whether that's what people want ...

ie Do they want ALL the edges that form the concave region to be hard, rather than just the boundary edges - leaving the 'inner' ones soft.

I'd consider the better / more correct result being the one where just the lines of intersection (as with Union result) are hard.

Rather than take any notice of what I think, tho ... ask Dimitri ... as he's the (main) one wanting H edges to be returned.

If the cube with a bit of sphere subtracted was smoothed - in the all Hard Edge state - it wouldn't smoothe correctly ... whereas it would (sorta) with just the boundary edges made hard.

I see the use of hard edges as a temporary / interim measure - to maintain the geometry created by using booleans - whilst a better way (bevelling the seam / join) is implemented ... in D's case, by retopo approaches to re-establish proper eloops / geometry Smile

Having hard edges where they shouldn't be doesn't make sense - to me.

Until this is clarified - ie what ppl consider is the preferred / correct result - I can't see much point in doing any more messing around re this particular issue.

pp

Dimitri · (This post was last modified: 04-06-2013, 10:10 AM by Dimitri.)

'I'd consider the better / more correct result being the one where just the lines of intersection (as with Union result) are hard.'

Exactly, Paul... the hard edges are needed only in the 'seam' (or 'intersect') edges. The other ones are useless and trouble producing... to select them and turning them to soft ones is a needless tiresome process (Mark, I do wonder how did you conceive such a thing)...

'I see the use of hard edges as a temporary / interim measure - to maintain the geometry created by using booleans - whilst a better way (bevelling the seam / join) is implemented ... in D's case, by retopo approaches to re-establish proper eloops / geometry'

Just so... : - )

puzzledpaul · 04-16-2013, 07:59 AM

I'm assuming that once you've sorted out the (main?) issues associated with the Boolean functions you'll let us know here / somewhere ... or will this depend on input from the author of the main (Carve) code?

pp