Changes between Initial Version and Version 1 of Ticket #4629, comment 11


Ignore:
Timestamp:
2019-05-08T10:42:55Z (6 years ago)
Author:
Francesco Casella

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #4629, comment 11

    initial v1  
    11OK, now I get it. In equation
    2 {{{ zeros(3) = frame_a.t + frame_b.t + cross(r, frame_b.f)}}}, it turns out {{{r = {1, 0, 0}}}, and that is marked with {{{Evaluate = true}}}, so the second element of the 3D array is {{{1*frame_b.f[3]}}} which gets simplified to {{{frame_b.f[3]}}}. Of course this is ok from a numerical point of view, but not from a dimensional point of view.
     2{{{ zeros(3) = frame_a.t + frame_b.t + cross(r, frame_b.f)}}}, it turns out {{{r = {1, 0, 0} }}}, and {{{r}}} is marked as structural, so the second element of the 3D array that you reported is evaluated to {{{1*frame_b.f[3]}}}, which then gets simplified to {{{frame_b.f[3]}}}. Of course this is ok from a numerical point of view, but not from a dimensional point of view.
    33
    44In order to keep things simple, I guess we should carry out unit checking before any constant-evaluation and simplifications are carried out by the NF.