On Sat, Sep 09, 2006 at 10:32:25AM +0200, Juergen Spitzmueller wrote: > Helge Hafting wrote: > > Now, if we had a way of starting the selection > > before the numeral/bullet/section-number but after > > the previous paragraph. I.e. an extra cursor position. > > Then the user gets a better way of deciding whether > > an enumeration goes inside the box or if we > > just have a boxed item. > > Yes, agreed. > > > Perhaps this is the real problem, we're lacking a > > cursor position? > > Probably. > > > Your patch is OK for boxes, as all the weird box cases (try > > boxing two items that are neither first nor last) are > > impossible to get sane in latex anyway. So for boxes, the > > user may ask for something stupid - and gets something silly. OK. > > > > Branches are different though. It makes sense to > > make a few items conditional, but this does not work. > > But it is not possible currently, as you have worked out. So I'd propose to > get this patch in, which obviously does the right thing, and try to enhance > the branch handling later. One thing we have to do, obviously, is to make > branches context sensible. > > Agreed?
I agree. Your patch makes improvement, it is a step in the right direction. Perfection must wait till branches truly can make anything conditional. Then, we should aim for: "Adding an enabled branch around any selection should not affect output at all, and disabling the branch should affect output the same way as a cut would". I don't think this will be possible in absolutely all cases though. Removing a list item that has an embedded sublist will change the environment depth of the embedded sublist. Doing that with branches forces the question: should the environment depth (and on-screen enumeration/bullet symbols) really change as the user turns a branch on and off? I guess this scenario will be an error in most cases, and jsut confuse. On the other hand, it seems correct to have section numbering change when the user disables a section. Helge Hafting
