Sure - here's a 3 storey hotel built programmatically:

http://www.pictureandword.com/test_platforms/floorplan_viewer/v2/index.html

That's pretty much worst case scenario - I don't actually need the
third storey, I just added it to see at what point the frame rate will
start dropping.  It's just on the edge now with 100k polys.  Any tips
on squeezing a bit more performance from it?  I can't imagine I'll get
much more than that.

The absence of floors will tell you why I'm waiting for a concave
DelaunayMesh or preferably a Swf class.  Guests will immediately fall
into infinity with this kind of setup.

Wag



On Jun 9, 4:16 pm, Fabrice3D <[email protected]> wrote:
> screendump or url of these "seriously complex building layouts" we could 
> see???
>
> Fabrice
>
> On Jun 9, 2011, at 4:46 PM, wagster wrote:
>
>
>
>
>
>
>
> > I know - patience is a virtue and all that :)
>
> > I ended up using Wumedia Swf class for the floor and LinearExtrusion
> > for the walls in 3.6 (everything is drawn in Flash IDE as we need to
> > production-line our building layouts).  This has the advantage that
> > you can draw labels and all sorts of things on the floor - perfect for
> > me.  Think I'll probably do the same when Swf resurfaces.
>
> > BTW - LinearExtrude + Merge allows for some seriously complex building
> > layouts at high framerates!
>
> > On Jun 9, 3:14 pm, Fabrice3D <[email protected]> wrote:
> >> no, that would be one of next classes coming.
> >> this one simply "connects" vectors as Mesh. It doesn't have any awareness 
> >> of a "path" or defined contours/constrains.
>
> >> Regarding architectural applications, one that I've started a while ago 
> >> and hope to finish one of these days, will allow you to define one or more 
> >> "holes", basically closed shapes, rects, circles etc
> >> and eventually another class would use that result to extrude. think here 
> >> the wall with door and window, extruded at wall thickness. But as some 
> >> features required are very near of booleans operations,
> >> so I started do more research in that field first, to see if I could do 
> >> this via a smarter way...
>
> >> I will gradually add these (hopefuly) usefull classes as I go. But I can't 
> >> garanty they will be released in a logical order. By that I mean that like 
> >> in case above of the wall being extruded
> >> the class doing this might be released before the one actually generating 
> >> the base mesh with "holes".
>
> >> There are still many 3.6 classes that need to join the 4.0 packages and 
> >> get enhanced in the process.
> >> As said before, all happends 1 by 1 :))
>
> >> Fabrice
>
> >> On Jun 9, 2011, at 3:48 PM, wagster wrote:
>
> >>> Oh cool - another prezzie from the Fabrice factory!
>
> >>> I put together a Delaunay class myself (for architectural generation)
> >>> a couple of months ago but gave up because I couldn't make it handle
> >>> concave shapes.  Did you get that figured out?  (I know, I know, I
> >>> could just try it out, but I'm at work and I can't wait until I get
> >>> home!)
>
> >>> On Jun 9, 11:40 am, Fabrice3D <[email protected]> wrote:
> >>>> Hi all,
>
> >>>> I've just added 2 new classes to Broomstick
> >>>> and 1 example class.
>
> >>>> - DelaunayMesh:
> >>>> Located in "extrusions" package this class allows you to generate meshes 
> >>>> from (random) vectors. At least 3 are required.
> >>>> This is very usefull in many cases, such as terrain building from 
> >>>> pointclouds data, closing shapes, architectural generation, editing 
> >>>> tools (such as Prefab ;)  ) etc etc
>
> >>>> - Trident
> >>>> Our good old dev friend is back, but this time, located into "debug" 
> >>>> package instead of primitives
> >>>> It works exactly as the previous versions, except you are now also able 
> >>>> to change it's position and it looks a bit more sexy :)
>
> >>>> - Demo class
> >>>> I've added a simple example of a random cloud of vector3d's to mesh 
> >>>> using the Delaunay class, in the code
> >>>> you'll find as well the Trident in action for those new to Away and even 
> >>>> added a small snippet for the new MeshDebugger in case you've missed my 
> >>>> post on this one.
>
> >>>> Cheers,
> >>>> Fabrice

Reply via email to