On Wed, Dec 03, 2003 at 04:01:15PM +0000, John Levon wrote: > On Wed, Dec 03, 2003 at 04:48:25PM +0100, Andre Poenitz wrote: > > > > > Given range style LCS > > > > > > > > aaasa dfas dfa ds fasd va sdcv asdv casd > > > > > > > > |----- a ------| > > > > | > > > > |----- b -----| > > > > > > > > what should be output to LaTeX/XML/whatever? > > > > > > That was always my question. > > > > And that's _the_ argument against 'styles as ranges'. > > That's it ? > > Doesn't seem a very strong argument to me. > > "Because a user may for some reason try to apply overlapping > non-recursive styles,
You miss the point. I am arguing that users rarely need overlapping styles and consequently all-boxes is sufficient. You are arguing that you want to select half of one range and half of an other. Ranges might be sufficient, too, but as they ultimately have to go to boxes at some point internally anyway (as e.g. the backends don't handle improper nesting) there is no compelling reason to use them. Andre'
