SOLVESPACE -- parametric 2d/3d CAD
Examples
Tutorials
Features
Download
Reference
Technology
Library
Forum
Contact
USER FORUM

(you are viewing a thread; or go back to list of threads)

solvespace ideas (by Daniel Jones)
Hey I don't really get into forums. Well I guess I won't reply anyway, but heres my ideas on the bug, well I am not a programmer, but I did a CAD diploma of two years recently and heres the bugs I have considered in many 3D programs, even the fancy ones at 10 G's like inventor.
Umm yeh So if you create a rectangle sketch, there are two possible solutions, one being the hourglass of line crossover, and one of the rectangle. Ie. negative and positive solutions of the numberline, and I was thinking that a vector mate, and it would possibly lock the actual lines endpoints somehow,for temporary vector mating vs the global 3d vector, and in the programming it can be vs the central xyz global origin vectors, like they autocompare, to double solutions ability. then theres def no bugs, Like yeh it would not be like a point mate, but the lines endpoints would auto be able to be like vecotr mated to six degrees of freedom comparison, and this is always autocompared to the origin vectors. I reckon that would give the 9 degrees required for the two factors of things being out in the air as such, and you know it might not be automatic, but lines could autolock. A lock point is still not comparable for the fail factor of line crossover negative and postive solutions, maybe manual vector mates, but I really think it has to do with the lines, like theres a real bug and for out in the air stuff there needs to be this 9 degrees of freedom comparison equations. Maybe simple vector mates would work, but yeh lines, they can go back and forth so it has to be like that. But It could be automated I suppose. I think that al those crashes even in advanced cad programs like inventor have happenned from this problem. Yeh I don't know how you would program that to be automated stability of sketch lines anywhere in 3d, but it should be along the lines of a double comparison vector thing. And you would be surprised. It does happen in advanced programs, and if you could sort that one thing out, crashes could be automated to be impossible, if theres triple comparities.
also, another side idea is that for extrusions, the parts when mated could compare extrusions. You could switch off an on say an over difference extrusion to put a hole in another part, I reckon that would so advance any cad program. Or even if its not switchable, thats debatable that it would be, maybe you can set difference extrusions to correct level that it would not put a hole in the next assembled part, and you then have to do it properly, and that would also be awesome anyway. But yeh, I reckon you could charge 50 bucks for that, as an addon. And the vector stuff of the 9 degrees of freedom line non negative positive dipolars, I dunno if these guys in big co.s like inventors autodesk, like the programmers bs, to retain work stability, but its not at all going to affect them if the solvespace program does that.Poss its a rolling error not going back to basics,for line end point vector controls, or on the other hand they just have gone off on a tangent and lost various cleanliness factors, or haven't invented above basic degrees of freedom control. But yeh this stuff could autotraingulate, and not just a point, like a lock point, but to control the negative positive solutions of an actual line in 3d space.
I def know what I am talking about. I did two years of cadding for a diploma. its not about foolish constraint use. The systems are inherently unstable, or they don't have like double or triple cores well effectively, anyway, or degrees of freedom control which are triple cored, vs the global for floating cadding in space, where equations become unstable.
But yeh even after that, it might be real small code and it can be raspberry pi stuff, I think thats where you are aiming.
Maybe its foolishness of not thinking of lines as scalar and can become negative also for a solution. And the old ideas of a lock point mate are just rubbish, you really need to lock a dipolar line, and this can autocompare to the central origin.
It's def something like that. I am not a programmer just a visual cad observer. I reckon you could 9 deg of freedom purify it, where it yeh has double or triple comparities vs the origin, so can always calculate clean as a whistle.
Yours Sincerely
Daniel Jones
Wed Aug 26 2015, 05:20:15
sorry (by Daniel Jones)
sorry for that email being a little repetitive and unclear, yeh basically to lock a lines scalar from the end point vectors of both ends it can give a triangulation in 3d space vs the origin that controls flaws. like 3d degrees here on that end point of a scalar line, then the other end also the same, and they autocompare to keep the line stable, for neg positive, and yeh its like a triple cor out in 3d space, vs the origin. and that would make the day. Poss it could auto 6 deg of freedom for line endpoints then these autoremove as further lines and constraints are added, but the basic initial stability, is inbuilt, and def think its a programmers flaw not a constrint makers flaw, and the code rools to reatin this flaw, even to huge assemblies in inventor etc. and in the most advanced programs, and crashes occur cause no double 3d vectors of the first lines endpoints made on a sketch vs the origin, Yeh i reckon it amps onto crashes in majorly big programs like inventor.
Another thing is tangents. I reckon the chord shit can stuff things up for mates, if you could somehow triangulate two end points of the chords of a circle, vs the origin for auto triangulation, that would be a flaw solver too for crashes.
The point is yeh these lines, in 3d space. and the chord of a circle is on too, so they need to auto triangulate. And never the coding is just for lowest solution like scalary, but more like the lowest solution for triangular solution, and if that doesn't work, there is the manual vector mate to keep the circuits happy,and you could do that as well, if it gets a bit psycho.
Good Luck. I hope you appreciate these ideas. They are from a guy that does know somethings wrong in these programs, and that poss its the line triangulation, like a scalar line, and most programmers might be thinking of points as comparities whereas instead you can can have a line comparity, meaning 6 deg, to control crash flaws.
I mean like yeh they all respect past programming and its get cut and pasted, but this is like the line flaw gone, vs the point for programmers flaws, like a line is considered the device, rather than a point being considered the device, like of basic triangulation. and yeh it must be cut and paste code back from the days of autocad and crap, but they missed something I guess. Like line triangulation of something.
Its what I am thinking. Takes extra effort though.
Yours Sincerely
Daniel Jones
Wed Aug 26 2015, 06:02:22
(no subject) (by EvilSpirit)
If you have written so much, i suppose, you can provide illustrations of your thoughts, so we will get ability to understand you much clearly.
Thu Aug 27 2015, 07:08:01
Post a reply to this comment:
Your Name:
Your Email:
Subject:
(no HTML tags; use plain text, and hit Enter for a line break)
Attached file (if you want, 5 MB max):
© 2008-2018 SolveSpace contributors. Most recent update Nov 22 2018.