>>>>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> I have thought about the proper handling of super- and
Andre> subscripts and came more or less to the conlusion that some
Andre> kind of base has to be part of the inset.
Andre> Three reasons for that:
Andre> 1. I am not aware of too many situations where there is a
Andre> *script on its own in a formula (one might argue with
Andre> "prepended" scripts, but that's solvable).
I think some people do that in physics to have *scripts n the left of
a symbol.
Andre> 2. In all most of the cases there is some kind of 'semantical
Andre> tie' between base and *script. So having the base as part of
Andre> the inset, anything related to semantics could be easier (think
Andre> of export to Maple/*)
Yes, but this could be done via a distinct macro, as we discussed for
integrals. I do not think that the maple export feature should force
everybody to provide semantics. OTOH, one could argue that LyX is
supposed to encourage semantics vs WYSIWYG.
Andre> 3. The implementation of drawing will be cleaner (no need to
Andre> look outside the inset to determine in which height the
Andre> superscript to draw)
Remember that this is what TeX does, so you would be closer to its
semantics.
Andre> A reason against that:
Andre> - The uservisible behaviour will change (i.e. when moving in
Andre> 'ab^c' it will take three right steps from the left- to the
Andre> rightmost position: One to go behind the a, one to enter the
Andre> inset (optically the same position), one to leave the inset.
I think people will be pissed by this change, unless you can find a
clever way to hide it (which seems difficult).
JMarc
PS: do you have plans to support the \big* family of delimiter modifiers?
