HI, The error message is this one : Output parameter Index[2] too high or too low for Component Sort List. But like i said once you delete the sort list component and replace it by a new one everything seems to work.
Thanks for the quickk answer that is exactly what i was wondering. The only strange thing (I'm a newbe) is that when you display the points in the rhino viewports they are all located at a corner of the surface and not at each intersection as your sketch. Thank you very much On Apr 3, 2:29 pm, visose <vic...@gmail.com> wrote: > What's the error message? Remember to flatten all the inputs of the > sort component. > > You mean the point components before the surface evaluation? They are > just uv coordinates used to achieve the distances along the center of > the interval (for the panel size) and the center point to place the > panel:http://grasshopper3d.googlegroups.com/web/pointsdiagram.jpg > (sorry for the lame drawing, i used the grasshopper sketch tool, I'm > not sure why i didn't use photoshop since i was posting it as a > screenshot anyway) > > On Apr 3, 9:18 am, Claudio <claudioarch...@gmail.com> wrote: > > > IHis Visose > > > The definition works if you delete the sort list component and replace > > with a fresh one. But the last fonction component which returns panel > > types still red colored. don't know why. Anyway Its is beatiful. One > > question though. Whats the use of the point component after the > > surface evaluation?. > > > Thank you very much. > > > On Apr 2, 11:34 pm, visose <vic...@gmail.com> wrote: > > > > Very strange. It's got to do with the sort list component. Looks like > > > a bug in the new input manager. The thing is i noticed it, fixed it, > > > saved it and uploaded it again but it's still there. To fix it: > > > Go to the sort list component, right click on it, go to input manager, > > > delete the B input, add it again, and link the B output with the 'y' > > > input of the expression component at it's right. If the bug persists > > > maybe delete sort list component and add it again. > > > > On Apr 2, 8:47 pm, aloysius lian <aloysius.l...@gmail.com> wrote: > > > > > Hi Visose, > > > > > When I tried to open the defn, there was an error: > > > > > Output parameter Index[2] too high or too low for Component Sort List. > > > > > Any ideas? > > > > > Al > > > > > On Thu, Apr 2, 2009 at 1:42 PM, Claudio <claudioarch...@gmail.com> > > > > wrote: > > > > > > HI Visose, > > > > > > Thank you very much. I will take a very good look at your definition. > > > > > > On Apr 2, 3:14 pm, visose <vic...@gmail.com> wrote: > > > > > > Here:http://grasshopper3d.googlegroups.com/web/flatpanels.ghx > > > > > > There are probably alternate ways of doing it. I added the rounding > > > > > > down of panel sizes. To use the panel tolerance slider: 0 will round > > > > > > all panels to the nearest unit, 1 will round them to the nearest 0.1 > > > > > > unit, etc. In the previous example, setting this to 1 created 11 > > > > > > different types of panels without any noticeable difference. > > > > > > There's also an "option B" that places the panels instead on the top > > > > > > of the surface, on the medium position of the interval, this is > > > > > > because if you place the panel directly over the surface, in convex > > > > > > or > > > > > > concave intervals the panel will sit on the top/bottom instead of > > > > > > aligning it's sides with the boundary panels. > > > > > > > On Apr 2, 12:03 pm, Claudio <claudioarch...@gmail.com> wrote: > > > > > > > > Hi Visose. > > > > > > > > That sounds great. Can you post the definition plz. I really like > > > > > > > to > > > > > > > see how you do it. > > > > > > > > Thank you very much > > > > > > > > On Apr 2, 12:12 am, visose <vic...@gmail.com> wrote: > > > > > > > > > I tried to reproduce the ceiling in GH. > > > > >http://groups.google.com/group/grasshopper3d/web/fosterceiling.jpg > > > > > > > > It sort of looks like it, if they did it like this (they most > > > > > probably > > > > > > > > didn't), then it's very simple, but most of the panels have a > > > > > > > > different size. > > > > > > > > I just used the divide interval2 component on the surface, > > > > > > > > found the > > > > > > > > centers of each interval, found the width (in the u and v > > > > > > > > directions) > > > > > > > > of each interval to use as the panel size (along the center of > > > > > > > > the > > > > > > > > interval, since the distances will vary) and placed a 'center > > > > > > > > box' on > > > > > > > > the centers normal to the surface using the calculated width as > > > > > > > > x and > > > > > > > > y dimensions. If there is not much curvature and/or the panels > > > > > > > > are > > > > > > > > small, most of the panels will have a similar size, so you could > > > > > round > > > > > > > > the panel widths down so you've got only x number of different > > > > > > > > panels > > > > > > > > adding a little margin between panels. > > > > > > > > > On Apr 1, 6:34 pm, Claudio <claudioarch...@gmail.com> wrote: > > > > > > > > > > I see what you mean I was completly wrong about the > > > > > > > > > teatragon. So > > > > > in > > > > > > > > > this case all the panels used in fosters porject are flat and > > > > > > > > > perpedicular to the curvature. Is that Right? > > > > > > > > > > Thank you again > > > > > > > > > > On Apr 1, 3:41 pm, David Rutten <da...@mcneel.com> wrote: > > > > > > > > > > > Tetragonal doesn't mean a square base. A tetragon is what > > > > > > > > > > you get > > > > > when > > > > > > > > > > you deform a rectangle. A 'regular tetragon' is what you > > > > > > > > > > get when > > > > > you > > > > > > > > > > skew a rectangle.http://www.mathopenref.com/tetragon.html > > > > > > > > > > > The angles and dimensions of each panel may still be > > > > > > > > > > different. > > > > > > > > > > > I think the main reason Foster picked this solution is > > > > > > > > > > because > > > > > all the > > > > > > > > > > panels are flat, which makes them cheaper (actually, I know > > > > > that's why > > > > > > > > > > they did it, because I spoke to them about it). > > > > > > > > > > > -- > > > > > > > > > > David Rutten > > > > > > > > > > da...@mcneel.com > > > > > > > > > > Robert McNeel & Associates > > > > > > > > > > > On Apr 1, 5:59 am, Claudio <claudioarch...@gmail.com> wrote: > > > > > > > > > > > > Hello David. > > > > > > > > > > > > Glad you"ve answered. > > > > > > > > > > > > I dont really know if the panels are really identical. > > > > > > > > > > > The only > > > > > fact > > > > > > > > > > > that makes me think of that is an article in the detail > > > > > magazine (an > > > > > > > > > > > architectural german magazine) they explain that the > > > > > > > > > > > panels are > > > > > > > > > > > Tetragonal, Which means that they have a square base (a ; > > > > > > > > > > > a) > > > > > and a > > > > > > > > > > > different height (c). So iI'm guessing that at least the > > > > > > > > > > > panels > > > > > are > > > > > > > > > > > identical y the x & y axis. Otherwise as architect my > > > > > > > > > > > self, I > > > > > know > > > > > > > > > > > that curved surface are very difficult to populate with a > > > > > single panel > > > > > > > > > > > solution, but is this solution that we all try to acheive. > > > > > > > > > > > > Anyway I'll continue my research. If you have any more > > > > > > > > > > > ideas on > > > > > how to > > > > > > > > > > > acheive this just keep me posted plz. > > > > > > > > > > > > Thank you > > > > > > > > > > > > On Mar 31, 4:03 pm, David Rutten <da...@mcneel.com> wrote: > > > > > > > > > > > > > Hi Claudio, > > > > > > > > > > > > > how do you know the panels in the Foster design are all > > > > > identical? > > > > > > > > > > > > > Panelling with identical panels is a very complicated > > > > > > > > > > > > area of > > > > > > > > > > > > geometry. There are a number of obvious solutions: > > > > > > > > > > > > > - rectangular, triangular and hexagonal planar grids) > > > > > > > > > > > > - icosahedrons and other platonic solids > > > > > > > > > > > > - singly-curved surfaces with rectangular panels (this > > > > > > > > > > > > one > > > > > only works > > > > > > > > > > > > for identical panel shape, not joint-angle) > > > > > > > > > > > > - special cases where the underlying surface is a direct > > > > > emergent > > > > > > > > > > > > result of the panel geometry (for example penrose > > > > > > > > > > > > tilings) > > > > > > > > > > > > > And then there are weird special cases. > > > > > > > > > > > > > I found that many architects/engineers who face a > > > > > facade-panelling > > > > > > > > > > > > problem don't even try and come up with a single panel > > > > > solution. > > > > > > > > > > > > Instead, they try and minimize the number of different > > > > > > > > > > > > panels > > > > > they > > > > > > > > > > > > need to build the whole shape more or less accurately. > > > > > > > > > > > > This > > > > > is always > > > > > > > > > > > > a per-project problem that requires a per-project > > > > > > > > > > > > solution. > > > > > Sometimes > > > > > > > > > > > > they'll even design the panel and the underlying > > > > > > > > > > > > surface at > > > > > the same > > > > > > > > > > > > time, which gives you much more possibilities for > > > > > > > > > > > > solutions. > > > > > > > > > > > > > -- > > > > > > > > > > > > David Rutten > > > > > > > > > > > > da...@mcneel.com > > > > > > > > > > > > Robert McNeel & Associates > > > > > > > > > > > > > On Mar 31, 12:23 pm, Claudio <claudioarch...@gmail.com> > > > > > wrote: > > > > > > > > > > > > > > Hi David, > > > > > > > > > > > > > > Can you please be more specific, In which cases this > > > > > > > > > > > > > can be > > > > > possible? > > > > > > > > > > > > > I'm also iterested in diving a surface with identical > > > > > panels. Please > > > > > > > > > > > > > refer to this article for an exemple : > > > > >http://www.detail.de/rw_5_Archive_En_HoleArtikel_5990_Artikel.htm > > > > > > > > > > > > > And the architects website : > > > > >http://www.fosterandpartners.com/Projects/1276/Default.aspx > > > > > > > > > > > > > > In this case the roof has been divided with a glass > > > > > > > > > > > > > pane > > > > > responding to > > > > > > > > > > > > > a very simple rule called Tetragonal crystal system. > > > > > (http:// > > > > > > > > > > > > > en.wikipedia.org/wiki/Tetragonal_crystal_system). I > > > > > > > > > > > > > guess > > > > > the > > > > > > > > > > > > > curvature of the surface is a critical parameter in > > > > > > > > > > > > > this > > > > > case. > > > > > > > > > > > > > > I'll be more then happy to continue this subject > > > > > > > > > > > > > > Thank you all > > > > > > > > > > > > > > On Mar 30, 4:06 pm, David Rutten <da...@mcneel.com> > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > Hi Steve, > > > > > > > > > > > > > > > creating a filling pattern with identical panels is > > > > > > > > > > > > > > only > > > > > possible in a > > > > > > > > > > > > > > very small subset of cases. > > > > > > > > > > > > > > It's also impossible to create a closed pattern of > > > > > hexagons on a > > > > > > > > > > > > > > freeform surface, unless you allow the hexagons to > > > > > > > > > > > > > > be > > > > > non-symmetrical. > > ... > > read more »
