Simply scaling would change the weight relationship and in most situations would make things look wrong. Witness the disdain for faux small caps and faux fractions. A more sophisticated rule could probably be devised (probably relying upon hints), but would require a sophisticated implementation built into layout and rasterizing engines and would be highly unlikely to gain any traction I suspect.

BTW, parentheses and braces are generally scaled down (in size, but not weight) for all-caps settings, not up.

Thank you

I'm not certain of the coding, but couldn't you use contextual alternates for substitution with alternate glyphs?

That would still require creating new glyphs instead of applying an action/effect to existing glyphs.

It does solve the problem in the absence of other options. Or, am I missing something. Even if there is an operation for scale, existing software has to support it.

That would still require creating new glyphs instead of applying an action/effect to existing glyphs.

Yes, it entails applying an action/effect to existing glyphs on the design end, where you have greater opportunity to evaluate and refine the results on a situational basis, rather than at the rasterization end where you will have little control and where one size rarely fits all.

For alternative approaches, you can explore parameterized design and the Metafont format developed by Knuth for use in conjunction with TeX.