parallel lines
BlitzMax Forums/BlitzMax Programming/parallel lines
| ||
| If a vector cross product zero , the vectors are parallel right? But what's a good value for messuring that the product is zero? If (cx>0.00001) And (cy>0.00001) And (cz>0.00001) ? Short and long vectors will not have the same 'parallellity tolerance' :) Should I calculate the magnitude first and do something like: If (cx>0.00001*mag) And (cy>0.00001*mag) And (cz>0.00001*mag) ? But the magnitude would be close to or zero also, so I'd have to check that against another threshold first. What threshold? |
| ||
| you'd need abs(cx) wouldnt you? to capture small -tive values |
| ||
| I haven't really tested this extensively, but if I were trying to find parallel lines I'd do something like this: Local ax# = 1.0, bx# = -2.0 Local ay# = 2.0, by# = -4.0 Local az# = 3.0, bz# = 6.0 Local ratiox# = ( ax / bx ) Local ratioy# = ( ay / by ) Local ratioz# = ( az / bz ) Local ratioxy# = ( ratioy / ratiox ) Local ratioxz# = ( ratioz / ratiox ) If ( 0.9999 < ratioxy ) If ( 1.0001 > ratioxy ) If ( 0.9999 < ratioxz ) If ( 1.0001 > ratioxz ) Print "parallel" EndIf EndIf EndIf EndIf |
| ||
| thx, but I'm looking for cross product test. I got another suggestion on IRC on normalising the cross vectors first. That seems like a good idea, to get an even compare. Meanwhile I coded myself out of needing to test for parallel lines :) |
| ||
| A ratio test is faster and simpler. A cross product test has a cool, clever sounding name. Both do the same job. Take your pick. |
   |