Take #5. Still debugging with sf.net staff. New datacenter so maybe this one will finally make it through..
On Sunday, August 24, 2008, at 02:54AM, "Christopher Sean Morrison" <[EMAIL PROTECTED]> wrote: > >Feel free to ignore this message if it makes it through. I'm >debugging a mailing list issue with the Sourceforge staff. The >message apparently matches some "pharma ban" regular expression >blacklist spam filter on sf.net mail system (before it even gets to >mailman). > > > > ---------- Forwarded message ---------- > > From: Dawn Thomas <[EMAIL PROTECTED]> > > Date: 2008/6/30 > > Subject: [brlcad-devel] gsoc: Libpc > > To: John Anderson <[EMAIL PROTECTED]> > > Cc: [EMAIL PROTECTED] > > > > > > Dear John, > > As you would have seen in the repository, I have integrated Boost > > Graphics Library ( :) and corrected that crappy build break ) . I >have > > implemented also the basic binary constraint solving framework. The > > current solver doesnt work since there is no backtracking. I am a > > few days away from completing a fully functional binary constraint > > solver. By functional I imply a solver which returns all solutions of > > the constraint with proper representation of ranges of variables in > > case of multiple solutions. > > > > Regarding BGL, as of now i am using only their data structure and > > access mechanisms and none of their algorithmic support. I have also > > enabled postscript support via graphviz /dot so as to check the >graphs > > generated < http://parametrics.files.wordpress.com/2008/06/x.png>I > > think when we actually deal with non binary constraints ( next on > > agenda after I finish the work on binary constraint solver and the > > remaining part of the C code ( dbio, params, pc_set ) and in > > particular hypergraphs/bipartite graphs, such template algorithms and > > in particular visitor<> based extensions of bgl template algorithms > > would be important. > > > > I am also writing most of the parts as template classes so that later > > we could use it for generic constraint solving as well if required. > > Though it is really far away i guess. Geometry constraint solving and > > representation is a tough job in itself. > >> > >> If the constraints are stored as objects in the database file, > >> they will need names (as I see you already know). Do you intend > >> for the constraint object names to be exposed to the users? If > >> not, a method could be developed to provides unique names that are > >> not likely to collide with user created names. I guess this could > >> be addressed in a UI developed later. > >> > > My intention was not to expose the constraint object names to the > > users. I mean a user does not stand to gain much from that > > information. We could write a random name generator. I am still > > thinking about how to represent the constraints / list of constraints > > in the UI. > > Should it be that the user provides a command such as > > > > #listconstraints sph1 > > would provide a list of constraints > > eg. > > > > 1. tangency with box2 > > 2. radius = 30 > > 3. centre = apex of cone3 > > In such a situation to support editing we can straightaway give users > > access to the constraint object names or we can provide a > > functionality such as > > > > #constredit sph1%1 > > or some similar character for % implying edit tangency with box2 > > > > I somehow find both the methods rather unsatisfactory. > >> > >> I suspect the solving of the constraints will need to be > >> triggered in a few places. Obviously, some command in MGED to run > >> the constraint solver. Also in rt, to get the constraints > >> evaluated prior to raytracing. I suspect this should be done in > >> "rt_gettree_leaf" (tree.c in librt) before the prep routine of the > >> primitive is called (or perhaps from the prep routine itself). > >> > > I was thinking of calling it from the prep routine itself. I am not > > sure if some of the specific checking needs using the constraint > > solver functionality. Regarding interaction with mged, my idea was > > that we invoke the constraint solver by some particular command which > > generates the solution if it exists . If multiple solutions exist, it > > informs the user of the same asking him to support particular set of > > values within the domain solution ( or select an optimal solution if > > the user chooses so) This would be followed by solver/mged calling >the > > geometry update functionality which will update the geometry whose > > parameters have changed as well as other geometry associated with > > these. > > > > Sincerely yours, > > > > > > > > > > > > -- > > Dawn Thomas > > IIT Kharagpur > > > > " the strength to change what i can, the inability to accept what i > > can't, and the incapacity to tell the difference" > > > ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ BRL-CAD Developer mailing list brlcad-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/brlcad-devel
