Benson Shaw

Wowzer! 9 days??? Thats the meaning of patience. Hope its on a separate computer so you can work on other things. I have tried the Split tool which works ok. But not on huge data sets. Needs a separate split for each edge -B

In the View menu you have a clip cube turned on. Un check the Clip Cube option, or adjust the existing clip and your model should reappear! Bit of a mystery, but I think it's solved! -B

Another thing to try is the default shortcut or key command to show/hide the OIP. The Mac OS key shortcut: cmd i (Command key and i key) -B

Another thing to check is your Vectorworks Preferences>Display>Navigation Graphics. Could be the model has surpassed your graphics card ability at current pref setting. See if view changes behave with lower setting. Or maybe something else entirely. Reboot? -B

A little update on the quit error. Several file issues were addressed at about same time. Now the quit error is gone: Update to SP4 (This is most likely source of the "cure") Corrected a small user origin offset in one file. I did not move it on purpose, so not sure how it moved. I do not normally display the rulers. Corrected several broken and out of date DLVP references in a group of files. All happy again. -B

No guarantees, but the fillets work on my test with generic solid. Not as confident about the plain shell. Sticking point seems to be when the shell creates a knife edge or some kind of pooch in or out at the corners. The fillets fail even on single objects. I didn't test, but maybe it would help to fillet the corners way back at the source objects for the extrude. Sorry this is all so unknown. Vwx doesnt have a play book, so making it up as we go. A totally different approach. I will try it if I get a minute: Create a little site model terrain representing the dish surface. Select all the 2d source objects then Send to Surface. They will become 3d polys with MANY vertices, i think they will shell or convert to NURBS surface, then Shell. Hopefully this shell will fillet. Apply the fillet pref for select Tangent, as well as the All edges blah blah -B

Oh, one reason the fillet might fail is that some of the mosaic "tiles" are smaller than the fillet radius. Try to clip out anything that small at the extrude stage. Another thing to try (on a duplicate) is to convert the Shell into a Generic Solid (Modify menu>Convert). It's still a single object with non adjacent components. Fillet>All Edges option. Might also need the Tangent Entities option. Anyway I'm making the fillet work this way. -B

HI, Everett - Sorry. I have no contacts to recommend. Maybe others here will suggest a process or a vendor. But . . . Seems you have this solved except for the fillets. But that should work, too, and without manually clicking every edge. So, some things to verify: 1. Select all the planar mosaic footprint bits and extrude them in one go. Result is a single extrude object. (It can be ungrouped, but don't do that) 2. Pierce the big dish surface shape with the extrude of the mosaic and make the solid intersection. Result should be a single object with lots of un connected curved surfaces - your flakes. OIP shows just the single object. 3. Engage the shell tool, set the thickness, click any one surface of the Solid Intersection. That surface highlights. Click the checkmark. Result - every component is thickened in one go. OIP identifies a single object. Type is Shell. 4. This shell object has many components, your 3d flakes, not touching each other. 5. Engage the Fillet Edge tool. In the prefs, choose Select All Edges option. Click any edge. ALL the edges of all the components in the shell highlight. Click the checkmark (or press Return). May take a while, but all the edges should fillet - tops, bottoms, sides. All in one go. No need to click every edge. 6. Those bottom fillets might be a problem, but make the shell thick enough, and cut off with the dish surface via Solid Section. or just sink the filleted shell into the dish. Add Solids if desired. If this fillet fails, it might be choking the CPU. Try splitting copies of the shell - slightly more than half each side (or quarters), then fillet. Then recombine them via Add Solids. The unwanted fillets at the joints will fill in because the larger-than-half components overlap past the joint. At least this works with my little test. Post more updates. -B

@ebramhall Good work. I'm getting the sense that your final product is a manufactured version of this drawing. The fillets are quite inefficient, and obviously not working well. So, perhaps another software would work better, but I have no experience. Maybe one of the Rhino folks, or other will comment here. If this is representing a mosaic, eg broken pottery or glass, then it might be worth modeling individual bits (with or without fillets) and placing them on, or sinking them into, the surface. Maybe start with several symbols - extrusions of different source shapes - which can be placed/rotated onto the surface, then scaled in OIP. These might benefit from subsequent conversion to generic solid. If you are willing to abandon your current mosaic arrangement, the Create Surface Array may present a fussy way to place multiple filleted shapes en masse. 1st, make symbols of several "tile" pieces with filleted edges. Array one of them onto the surface with lots of repetitions, basically cover the surface. Overlaps are OK. Duplicate and work on the dupe: Ungroup the array. Now you have the individual symbols. Select any one instance. Rotate tool as desired (automatic plane), rescale via OIP as desired, replace with any other symbol tile (orientation/scale is preserved). Maybe create another superimposed array with a different tile symbol. Ungroup. Edit orientation, and replace or delete individual tiles as desired. Good luck! -B

Exactly! I am not sure why this has not been addressed as the software upgrades year after year. Basic arc and circle shapes must indeed be some legacy relic, desperately in need of repair, so that the EAP and other functions with circles remain continuous. If vwx is using or has access to same geometry representation as other, similar software (eg parasolid?), then why cannot we at least have circles that represent as continuous rather than faceted. grrrrrr! Also wonder about that difference between extrude of closed shape with circular holes vs solid subtraction of extruded cylinders from volume. Also I don't like that circles become polys in the edit screen when they are clipped as holes from a 2d shape. Best edit is to delete the polys and replace with new circles (which become polys as soon as exit the edit screen) sigh -B

Circles extruded along path, especially multiple circles, create many surfaces. The circles extrude segmented, creating about 250 facets per circle. Rather invisible in renders unles the view is very close. So, maybe two models on separate layers or in separate files? One for those long views with lighting using extruded triangles or squares or hexagons instead of circles. This will really cut down the rendering overhead. Another model for close views and shop drawings uses the circles. This one might be only a partial sculpture, just enough to illustrate the weaving/overlapping, or other detail. -B

Totally roundabout. I was just playing with a work flow to extract the perimeter: Make the DTM, Edit Crop, Select/Copy the crop (a 2d poly), Exit. Paste in Place. Send to Surface. Result is a Group. Ungroup to reveal a 3d Poly following surface perimeter. The DTM has an algorithm to choose which source points are perimeter and which are interior topo, so all this probably doesn't meet the need. at least I can't make it do just a lasso selection of a curving line. Oh, well. Left to right, or screen upward/downward might also be needed if closed path is not desired (eg edge of roadway or path - it never connects to edge at the other side). But thanks for this script! -B

Not sure if this accomplishes anything desired, but ... The vwx site model (DTM) from 3d loci produces a point to point perimeter, plus the interior topography. If the raw data is somehow delimited, eg if provided with separate layers or classes for roads, beds , creeks, paths, etc, then a separate DTM for each will produce separate perimeters. and contour interval adjustments might be useful in some way for this -B

There is a work flow for correcting Aligned Slab Hardscapes which curve and twist improperly. Basically, use the Edit Surface Modifiers>Grade>Placement mode. Place several connected grade objects just outside of each edge. Cross connecting grades can be added into the network. If the grade objects are snapped together, the node "sticks" together if the value is edited in the Edit Surface Modifiers dialog. Surrounding values for slope, segment length, projected length update automatically. Or, any segment can be disassociated. To me, this is a cumbersome process.One can envision a solution which is tangent to each aligned end and twists evenly to produce a smoothly curving slab. The ungraded aligned wavy, dipping shape is contrary to that vision! But the regraded slopes seem to be accurate for drainage specs, construction, etc. More/shorter Grade Objects increases smoothness, but probably increases file overhead and might decrease file stability? The interpolation option can decrease faceting. These are shown without interpolation. -B

@jeff prince has the best idea. As usual! They need to provide better/usable data. but . . . I had that feeling, too, but just could not stand it. Sooooooo, here is something crazy, probably useless, but fostered my illness to not let vwx fail: The file is full of 2' cubes (actually they are 2d lines and squares on 3d planes), each with a text box. I think the ctr points of the cubes are relevant data points related to their classes (topo, roads, building tops, etc). There are other non terrain data objects which need to be hidden for terrain production. But a terrain can be made. Here is the gist of it: 1. Duplicate file, plan view all objects. Delete the North Arrow, the VP boundaries, any stuff far from the cubes, etc, etc. 2. Custom Select>text. Delete it all. 3. Front view - Delete/hide stuff at z=0. (the relevant geometry is above z=450') 4. Select all the cubes. Modify>Convert to 3d Polys. 5. Make another pass for class hiding, eg the buildings, trees, other stuff that's off the imagined surface, stuff that isn't cubes, etc. 6. Note: I deleted lots of big 3d polys and some other things which instead probably could have been hidden via class vis. sorry. 7. Save View to preserve this class list. And another view with ALL the classes visible. 8. Select what's left, Duplicate in Place. Put it in a new layer : Terrain 9. In the new layer make the Terrain. Select all, AEC or LandMark>Terrain>from Source Data. Set min/max 450' to 550' 10. Terrain is created mainly from the top surfaces of the cubes. It has some blank facets, but not bad. 11. If desired, Drape Surface. Or, make another pass with class vis. Or mess with Mesh Smoothing. 12. Since terrain uses tops of the 2' cubes, the whole thing is (probably?) 1' above proper grades. Maybe. Depends on what those cubes mean. Edit the source data and move all down 12"? Move the terrain down? Or ignore it (what I did). Images: Terrain plus another layer with all classes visible showing cubes for the fences , buildings, trees, etc. 50x50 drape surface of the terrain. OK. crazy, useless, but a fun and interesting terrain never the less. -B sam test 2021-06-01 copy.vwx sam test 2020-06-01 copy.vwx