THE N-GONS PROJECT

Compiling, libraries, modules, coding guidelines and porting

Moderators: jesterKing, stiv

bobthevirus
Posts: 48
Joined: Wed Jun 25, 2003 7:11 am

Post by bobthevirus » Thu Aug 19, 2004 8:19 am

Instead of an ngons on/off selection, how about a totally new meshtype - winged edges, based on the code from wings3d. Then things could be converted to/from the current meshtype as required.

ideasman
Posts: 1000
Joined: Tue Feb 25, 2003 2:37 pm

Post by ideasman » Thu Aug 19, 2004 8:49 am

bobthevirus wrote:Instead of an ngons on/off selection, how about a totally new meshtype - winged edges, based on the code from wings3d. Then things could be converted to/from the current meshtype as required.
Bad plan IMHO-

If youd be better off Intergrating wings as a blender plugin- could be accsessed from the MEsh menu- Mesh >> Wings..... (edit you mesh).. Save from wings and blender will reload the mesh apon wings3d exit.
- Id write this if I used wings.

Intergrating wings or wings stule mesh is totaly non consistant with blender as well as the extra code you'd need (bloat)- As you may have read, I vode for fake NGons- In that they are selected and viewed like NGONS but are realy Tri's- Since NGons useually need to be converted to tris somewhere allong the way.

slikdigit
Posts: 213
Joined: Wed Oct 16, 2002 3:52 am
Location: Northampton, MA (US)

Post by slikdigit » Thu Aug 19, 2004 2:55 pm

"fake" ngons are the way to go, I think. The user would probably not notice any difference. One thing though is that the edge tools (alt-b select, shift-r, ctrl-r ) need to terminate at the n-gon, not follow a "phantom' edge that is invisible to user. one of the benefits of ngons is giving a place to terminate a bunch of edge loops.

ideasman
Posts: 1000
Joined: Tue Feb 25, 2003 2:37 pm

Post by ideasman » Thu Aug 19, 2004 4:11 pm

Fake also means internals dont have t obe rewritten.
- Cam

Youssef|3d
Posts: 21
Joined: Sun Aug 15, 2004 8:02 pm

Post by Youssef|3d » Fri Aug 20, 2004 8:06 am

Instead of an ngons on/off selection, how about a totally new meshtype - winged edges, based on the code from wings3d
I think thats a lot of work to integrate a new mesh type
maybe "FAKE GONS" are a better altaernative
I mean if they have the same characteristics why we need "REAL GONS" :D

I know many tools that dont have "REAL GONS"
3DS MAX or Cinema4D NGONS are just a fake because they just hide the edges
they are not real n-sided polygons but the work is like real NGONS :shock:

I believe this "FAKE GONS" concept will solve blenders NGONS problem :wink:

Youssef|3d
Posts: 21
Joined: Sun Aug 15, 2004 8:02 pm

Post by Youssef|3d » Fri Aug 20, 2004 2:22 pm

well, it looks like there are some smart coders that want to integrate NGONS on blender
they will not be real NGONS. something like 3dsmax or c4d but the same NGONS characteristics. Furtheremore faked NGONS can have holes :shock:

here are some ideas/functions that blender could have :
- the NGONS [ON / OFF] function
- extrude vertex (like wings or c4d)
- extrude edges (like wings or c4d)
- split tool ( a kind of knife; like a multiple face loop cut but it can be modified per edge

more ideas are welcome :D

ideasman
Posts: 1000
Joined: Tue Feb 25, 2003 2:37 pm

Post by ideasman » Fri Aug 27, 2004 9:49 am

Fake NGons realy need to be implimented with the shift to merge Face Select/UV and edit mode. - Ton plans this.

joeri
Posts: 2243
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri » Fri Sep 10, 2004 10:28 am

Aren't all N-GONS fake N-GONS?
The renderer will cut them up into tri anyway I think.

The only advantages I see in N-gons is in displaying objects in wireframe.
Then there is the issue of planar. Do N-gons need to be planar?
Even with quads non-planar faces can be a problem, thus I convert them to triangles.
If the nice object view is the main thing that people want (like the building and the cube), maybe the focus could be on that: display planar faces as 1 'n-gon' face?

bmud
Posts: 169
Joined: Sun Feb 08, 2004 8:13 am
Location: Raleigh, NC, USA
Contact:

Post by bmud » Mon Sep 13, 2004 4:01 pm

Here's a thaought:
I have max and I've used it for a long time.
I love blender and I've loved it for a short time.

1. "Fake" N-Gons WOULD benefit some, and I like the idea.
2. The concept of "fake" suggests to me that the code will simply involve some auto-inteligent dividing of the mesh, based on edges, into triangles -- preferably keeping as many quads as possible.

Input:
Then there is the issue of planar. Do N-gons need to be planar?
I hate 3ds max when I pull a point, and then my n-gon loses it's planar integrity. So I would suggest making an n-gon retain a planar appearance. Then an option in WKEY's menu would allow "break into quads", and perhaps "break into triangles", and maybe (going far here) "break up manually". All of those would then allow bending of the planar n-gon into whatever shape afterwards.

One of the reasons I think non-planar n-gons being allowed is a bad idea is because you could then theoretically have an entire mesh be one n-gon. I think that would create problems in logic, or processor weight.. or.. something. It's an interesting topic. I'll make some diagrams when I get home. =)

theeth
Posts: 1184
Joined: Wed Oct 16, 2002 5:47 am
Location: Montreal
Contact:

Post by theeth » Mon Sep 13, 2004 4:29 pm

bmud wrote:Input:
Then there is the issue of planar. Do N-gons need to be planar?
I hate 3ds max when I pull a point, and then my n-gon loses it's planar integrity. So I would suggest making an n-gon retain a planar appearance.
How would that work?
When you pull a vertex, is it rotating to polygon or stretching and how? Or is it just disabling moving a single vertex?

Forcing the ngon to stay coplanar sounds like a nice idea, but it also implies finding solutions to other problems (and I can't currently think of any nice way to do it).

Maybe you have some ideas on that?

Martin
Life is what happens to you when you're busy making other plans.
- John Lennon

stiv
Posts: 3645
Joined: Tue Aug 05, 2003 7:58 am
Location: 45N 86W

Post by stiv » Mon Sep 13, 2004 5:48 pm

theeth wrote:Forcing the ngon to stay coplanar sounds like a nice idea, but it also implies finding solutions to other problems (and I can't currently think of any nice way to do it).
Perhaps by constraining N ==3 ...

theeth
Posts: 1184
Joined: Wed Oct 16, 2002 5:47 am
Location: Montreal
Contact:

Post by theeth » Mon Sep 13, 2004 7:19 pm

stiv wrote:
theeth wrote:Forcing the ngon to stay coplanar sounds like a nice idea, but it also implies finding solutions to other problems (and I can't currently think of any nice way to do it).
Perhaps by constraining N ==3 ...
Great! Lets do this! :D

Martin
Life is what happens to you when you're busy making other plans.
- John Lennon

jesterKing
Site Admin
Posts: 1848
Joined: Fri Oct 18, 2002 12:48 pm
Location: Finland

Post by jesterKing » Tue Sep 14, 2004 12:01 am

theeth wrote:
stiv wrote:
theeth wrote:Forcing the ngon to stay coplanar sounds like a nice idea, but it also implies finding solutions to other problems (and I can't currently think of any nice way to do it).
Perhaps by constraining N ==3 ...
Great! Lets do this! :D
I think it's already implemented. Dunno who committed that.

/Nathan

:idea:

bmud
Posts: 169
Joined: Sun Feb 08, 2004 8:13 am
Location: Raleigh, NC, USA
Contact:

Post by bmud » Wed Sep 15, 2004 3:26 pm

For those of you familiar with AutoCAD or 3dsmax, n-gons would allow a chamfer tool as well, which is one of my favorite things about 3ds max. crazy to think about how complicated that could potentially get if n-gons were treated differently from a quad.

And on a side note, if faces were selectable with the ability to ignore backfacing, would we even need n-gons? What about if selecting a face could optionally also select those that are co-planar..

But both of those do not allow locking co-planar integrity of polygons, which is what makes n-gons "unrealistic", because even in 3ds max you could draw a square, and pull a point up on the z-axis and then you've got two triangles that look weird.

Perhaps then, n-gons are needed for ease of selecting? Savable selection groups could do similar things. But then you would need to define a lot of groups for a model of a soccer ball for example (football for you that are non-USA).

Blender also has a really cool co-ringular selection tool under KKEY while in Edit Mode, so would n-gons prevent proper ring forms? Currently it uses quads. So would it also make sense to have a co-planar selection tool?

If we define n-gons as strictly co-planar, would that make the welding on edges for two n-gons difficult? If you select one point and pull it, what happens to the others? Would n-gons benefit when I want to use FKEY to fill holes in my models by creating an n-gon face?

I'm all for n-gons, but really think about the big picture, and you'll come up with similar thoughts. There's places for them sometimes, just not all the time. They will also disable the use of other functions that require quads possibley. So this project is really about, as i said before, figuring out how an n-gon should be defined automatically as triangles, whether we want them to be co-planar or not. I think it's more about "freezing". I'll be happy to help once some of you absorb this post.

steve_t
Posts: 62
Joined: Tue Nov 05, 2002 7:10 am

Post by steve_t » Sat Sep 18, 2004 5:58 pm

Another benefit of ngons is the possibility of what I like to call "non destructive mesh editing" . You could delete an edges in the middle of your mesh without getting rid of the faces. You could then strip down an edge line and make a new one without deleting or making faces

I also think having the ability to force ngons to remain coplaner is a good idea.

Another idea is to be able to force a mesh to retain edge length within a certain perameter. you could then drag arround parts of the mesh like proportional mesh editing but keeping the same surface area. This would be good for things like textured cloth, clothes and book pages.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest