✓ Accepted answer
Personally, I use offset rather than scale. This means that I have pixel-perfect interfaces, but I have scalability issues. As such, I will need to do scaling myself in code.
Of course, it is easier to use Scale (and remember RelativeXX and RelativeYY!), but images can appear stretched (curse the inability to disable filtering) and text will not scale uniformly.
But there is no reason not to use both. A combination of Scale and Offset is probably the best approach, with code to augment the transitions (such as choosing the correct image from an atlas based on available space), but will be the most complex and take the most amount of time and resources to get right.
1 comment
Thanks for the explanation!