Currently, the secondary dimension is truncated to be able to comply with not exceeding the value permitted by the primary dimension. The problem is that when the dimensions are displayed in reverse ("secondary [primary]"), the values are inherently incorrect.
To further elaborate, we model in SI Units as we work internationally. Bur our parts are fabricated in the US, so we draft with the primary dimensions displayed as Imperial.
For this example, drawing dimensions are shown as .XX +/-.xx [.Y +/-.y] where X is in inches and Y is in millimeters. A modeled dimension of 100 +/- 10 mm currently displays as:
3.9 +/-.3 [100 +/-10]
This is terribly incorrect, as the dimension above would require the displayed secondary dimension tolerance to be truncated to a max of 9 mm to fit within the bounds:
* max bound of 109.42 mm = 3.954 + .354 in (this assumes 5 is alway rounded down)
* min bound of at 89.05 mm = 3.860 - .354 in
But we want this tolerance to be 10, so the dimension should really display as: 3.9 +/-.4 [100 +/-10]
* max bound of 111.96 mm = 3.954 + .454 in (this assumes 5 is alway rounded down)
* min bound of at 86.51 mm = 3.860 - .454 in
NOTE: a tolerance of 11 mm is also acceptable in this range, but the primary displayed as secondary, should continue to be unaltered from it's rounded form.
As another example, a dimension of 110 +/- 11 mm should display as:
4.3 +/-.4 [110 +/-11]
* max bound of 122.12 mm = 4.354 + .454 in (this assumes 5 is alway rounded down)
* min bound of at 96.67 mm = 4.260 - .454 in
Ideally, an immediate fix - for all builds - would be to round the secondary dimension when it displays as primary dimension (versus truncating it) - this would make it only slightly incorrect, versus how terribly incorrect it is now.
A future build would adjust the secondary dimension displayed as a primary dimension/tolerance to createbounds around the rounded value of the true primary dimensiondisplayed as secondary (essentially reversed from what is currently done to match the reverse display).