On 06-Sep-2014 16:49, alvinpenner wrote:
However, that's not the point. The point is that it is clear that the original intent of the unit change code was that the unit change conversions would be done directly on the actual coordinates of the objects and not be preserved in a group transform.
Yes, but is that a rational thing to do?
We already know it breaks clipping, it probably breaks masking, and who knows what else it breaks. Moreover it is exceedingly complex. In theory the simplest way to rescale a drawing is to just put a single group with a transform on top and put everything within it. If that level already exists, then just change the existing transform.
Is that not much (MUCH!) simpler than adjusting every value in the drawing?
And what is the advantage of changing all the underlying values? These are often already scaled by other transforms. So those numerical values may or may not correspond to the units.
Also, in terms of numerical stability. Mucking around with the units is going to introduce rounding errors into every single value in the document, and these are going to run off in random directions. Do it the other way, at the top level, and repeatedly rescaling the document might introduce rounding errors in the top transform, but it won't do anything to the values underneath, all of which will stay "sharp", as it were.
Regards,
David Mathog mathog@...1176... Manager, Sequence Analysis Facility, Biology Division, Caltech