00:36.18 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
01:17.04 | *** join/#brlcad DTRemenak (n=DTRemena@c-24-23-59-104.hsd1.mn.comcast.net) |
02:32.33 | *** join/#brlcad eorl (n=kvirc@dsl081-116-172.dfw1.dsl.speakeasy.net) |
05:26.14 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
09:57.23 | archivist | hmm interest in dimensions is building up, time for another hard look at the database format |
13:24.30 | brlcad | dimensions can be done regardless of the database format |
13:24.43 | brlcad | the database format can support the stashing of any key/value data |
13:32.27 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
13:37.35 | brlcad | the real "trick" is two-fold |
13:38.20 | brlcad | 1) extracting the dimensions from the geometry (convert to brep and calculate, or sample the implicit directly) |
13:38.54 | brlcad | 2) providing a user interface for specifying what they want dimensioned and how to display those dimensions |
13:39.51 | archivist | dimensions that are set should control the geometry imho |
13:47.35 | *** join/#brlcad DTRemenak|RDP (n=DTRemena@c-24-23-59-104.hsd1.ca.comcast.net) |
13:49.57 | archivist | the end result will a mixture of derived on the 2d and some driving when the items are created |
13:58.08 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
14:05.36 | *** part/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
14:28.48 | brlcad | i presume you mean dimensions that are effectively tied to the geometry directly such that they become constraints of a sort |
14:30.38 | *** join/#brlcad docelic (n=devv@clj34-71.dial-up.arnes.si) [NETSPLIT VICTIM] |
14:31.43 | archivist | yes |
14:32.19 | brlcad | hmm.. from a modeling perspective, that gets very tricky |
14:32.47 | brlcad | if I have arbitrary geometry, even just a simple primitive with a dimensioned length |
14:32.47 | archivist | i realise that |
14:33.28 | brlcad | there can be numerous transforms that will evaluate to an increase in that dimension, e.g. stretching and scaling might both make the dimension increase |
14:34.06 | brlcad | and for many prims that could be a completely non-linear transform |
14:35.01 | brlcad | unless it just happened to be one of a handful of fairly well-behaved dimensions, like the diameter of a cylinder or a sphere |
14:36.03 | brlcad | which are pretty trivial of course.. throw in CSG operations and arbitrary dimensions across several primitives and you suddenly have a fairly unsolvable constraint |
14:36.23 | brlcad | unless you always scaled uniformly or something wierd.. :) |
14:37.40 | archivist | yup I get some wonderfull drags in solid works or a nasty error message |
14:38.24 | archivist | one then does it properly forcing dimensions to be sensible or fix relatiions between lines |
14:38.33 | brlcad | heh |
14:40.36 | archivist | It generally checks before attempting the stupid, later versions are better in that respect |
14:42.16 | brlcad | does solidworks let you dimension across parts? |
14:42.31 | brlcad | and modify that dimension directly |
14:42.45 | archivist | drawing correct sizes at the start is less fraught |
14:43.01 | brlcad | hmm? |
14:43.47 | archivist | you can relate a new part to existing parts in an assembly yes |
14:44.10 | archivist | known as drawing "in place" |
14:44.13 | brlcad | but then is that just a position constraint? |
14:44.57 | archivist | well a mod of the assembly will stretch the part or give an error |
14:45.01 | brlcad | i'm thinking of a case where you set a dimension on the span of say a car engine, one for width, one for height etc |
14:46.14 | brlcad | if I selected the width and "increased" it.. what would that do if there were no constraints set? |
14:47.17 | archivist | holes would not line up graphics will overlap |
14:47.46 | brlcad | so it wouldn't stretch all the geometry, just some of it |
14:48.14 | archivist | if it was drawn in place then some stretching would tacke place |
14:48.47 | archivist | it all depends wether parts are related or not |
14:49.10 | archivist | relations can be added or deleted |
14:49.14 | brlcad | relations are implicit constraints.. this simplist case is fully unconstrained |
14:49.40 | brlcad | (just conceptualizing the generic case) |
14:50.52 | archivist | unconstrained part grow and the graphics merge (an interference check needed) |
14:52.11 | brlcad | and what about the case where all parts are related in an assembly (again, the case of a simple engine for example) |
14:52.37 | brlcad | would it simply stretch the engine? |
14:53.27 | archivist | if it could yes (as long as the related join distance does not have a constraint) |
14:54.25 | brlcad | right, was presuming there aren't internal constraints that might impose additional constraints like having a hole some distance from an edge |
14:55.39 | archivist | hole from edge may be able to drag the edge |
14:57.43 | archivist | I shall have to do some test cases I suppose so we can see whats happening |
17:36.22 | CIA-7 | BRL-CAD: 03brlcad * 10brlcad/src/tclscripts/mged/pkgIndex.tcl: commit the empty index like all the others if only to be consistent |
18:40.41 | archivist | did a test case draw chassis solid, insert chassis into assembly, draw engine in place, stretch chassis and engine stretches to fit, keeping bolt holes registered |
18:41.34 | archivist | stretch was by editing the dimension in the chassis part |
18:57.09 | brlcad | that's sort of what I'd hope expect since your dealing with a lower dimension with dependent constraints |
18:57.25 | brlcad | how does it behave if you create a new dimension for the entire enginge? |
18:57.56 | ``Erik | vodka would be nice right about now |
18:58.21 | ``Erik | the funk soul brother |
19:01.44 | archivist | attmpting to constrain the bolt distance of the engine is not allowed |
19:03.27 | archivist | breaking the relation to the chassis would then allow it |
20:32.58 | CIA-7 | BRL-CAD: 03brlcad * 10brlcad/NEWS: 7.6.8 window of opportunity was lost due to merge complications on the 7.6 branch, so cancel it outright. revert, poof, gone. |
21:21.06 | *** join/#brlcad ChanServ (ChanServ@services.) |
21:21.06 | *** mode/#brlcad [+o ChanServ] by irc.freenode.net |
21:48.24 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
22:02.58 | *** join/#brlcad IriX64 (n=IriX64@toronto-HSE-ppp4305815.sympatico.ca) |
23:15.52 | *** join/#brlcad ``Erik (i=erik@c-69-250-155-85.hsd1.md.comcast.net) [NETSPLIT VICTIM] |