Thread: Re: [brlcad-devel] Should I begin to implement the missing brep conversions now?
Open Source Solid Modeling CAD
Brought to you by:
brlcad
From: p. <284...@qq...> - 2012-05-29 05:05:45
|
Hi, > Given there are at least five output formats of near-term high-interest, possibly more to come, and two of them are BREP-related (how to distinguish unevaluated brep from evaluated brep?), how about just defining one new option? That should be more extensible and simplify the interface. > A single option could control all the types of conversions being performed similar to how the OBJECTS parameter controls which objects are converted. Maybe an EVALUATE option or some similar means for specifying what outputs to include or suppress. Would still need to sort out the naming convention for the various output formats.. mentioned earlier: > unevaluated boundary representation faceted polygonal mesh (nmg+csg) > evaluated boundary representation faceted polygonal mesh (nmg) > unevaluated boundary representation faceted triangle mesh (bot+csg) > evaluated boundary representation faceted triangle mesh (bot) > unevaluated boundary representation trimmed nurbs (brep+csg) > evaluated boundary representation trimmed nurbs (brep) > unevaluated implicit representation gridded volumetric (vol+csg) > evaluated implicit representation gridded volumetric (vol) > Thoughts? My main concern is recognizing the evaluated and unevaluated forms while keeping the form simple. The EVALUATE option sounds like a good idea, simplifying the logic a lot. As far as I'm concerned, the current conversion.sh implements evaluated nmg and bot conversion, right? If we select an unevaluated conversion, should the nmg and bot conversion of evaluated combinations be included? > Unfortunately, "brep" is a bit of a misleading name since most of them are technically boundary representations. Could use 'nrb' and benefit from them all being 3-characters long, but then the names don't match the actual object output format being generated. I'd like to keep the name "brep" instead of changing it to "nrb". Because we commonly use the name "brep" for boundary representation trimmed nurbs, such as rt_brep_internal, the rt_*_brep() functions, the brep command in MGED, and so on. So I think it isn't quite misleading, as nmg and bot both have their own common used names. Cheers! Wu |
From: Christopher S. M. <br...@ma...> - 2012-05-29 15:00:18
|
On May 29, 2012, at 1:05 AM, phoenix wrote: > > unevaluated boundary representation faceted polygonal mesh (nmg+csg) > > evaluated boundary representation faceted polygonal mesh (nmg) > > unevaluated boundary representation faceted triangle mesh (bot+csg) > > evaluated boundary representation faceted triangle mesh (bot) > > unevaluated boundary representation trimmed nurbs (brep+csg) > > evaluated boundary representation trimmed nurbs (brep) > > unevaluated implicit representation gridded volumetric (vol+csg) > > evaluated implicit representation gridded volumetric (vol) > > > Thoughts? My main concern is recognizing the evaluated and unevaluated forms while keeping the form simple. > > The EVALUATE option sounds like a good idea, simplifying the logic a lot. As far as I'm concerned, the current conversion.sh implements evaluated nmg and bot conversion, right? If we select an unevaluated conversion, should the nmg and bot conversion of evaluated combinations be included? Hm, I wasn't thinking about the implication of EVALUATE to mean whether the CSG recipe is evaluated, but merely to indicate what output formats to "evaluate" the input geometry to. You are correct, though, that it presently performs evaluated nmg and bot. You'd be adding unevaluated brep nurbs and Anurag will probably end up adding evaluated volumetric. > I'd like to keep the name "brep" instead of changing it to "nrb". Because we commonly use the name "brep" for boundary representation trimmed nurbs, such as rt_brep_internal, the rt_*_brep() functions, the brep command in MGED, and so on. So I think it isn't quite misleading, as nmg and bot both have their own common used names. Agreed, "redefining" our usage of the term brep to imply anything other than nurbs isn't within scope. Deviating without changing all other instances would just introduce inconsistency and probably be confusing. Just a shame they aren't all 3-letters. :) So then the main question remaining is how to distinguish the different forms, particularly brep vs brep+csg but also as it'll apply to the other formats? Space separated using the names above would be: EVALUATE="bot nmg brep+csg" Cheers! Sean p.s. Your replies keep coming back as html-formatted and the quoted lines get rendered literally instead of as quoted text. Possible to change that to plain-text or change a setting that preserves the quoted context? |