Wed 13 Dec 2006 09:56:00 AM UTC, comment #6:
Indenting on screen, but not on print might have been the least bad solution, had it been possible.
I don't know a miracle solution to that problem, else I would have suggested it in the original report. That being said, one can imagine:
- There is actually place for two circle symbols in the margin, the rest is taken up by blank space in the box that turns green when you point your mouse on it. Just having the circle symbols take their "natural" space would allow nesting at level two instead of no nesting at all like now. Mock-up created with gimp attached.
- The two two-state circles may "merge" into one four-state widget (a thing that cycles between four states instead of two when double-clicked on). It could look like two partially overlapping circles. I can imagine this would be quite difficult to program, and would probably be more confusing than anything else for the user.
- The aforementioned unified widget could also give a contextual menu with the four choices when clicked.
- A fold-like construct that would have three arguments instead of two and behave like two nested folds (cycling between the four states).
The application is that I want to present the user four pieces of information a, b, c and d, where:
- a is grouped with b and c with d.
- Hiding the {c, d} group is one operation.
- Hiding b is one operation.
- Hiding d is one operation.
- The show/hide widgets are present on screen, but not in print.
- No effect on indentation a,b,c or d.
- Hidden parts don't take any vertical space.
- At least these states must be reachable:
* Only a shown
* Only a,b shown
* Only a,c shown
* Only a,b,c shown
* All of a,b,c,d shown
In particular, it would be acceptable if d could not be shown unless b and/or c is/are also shown. The best I could find up to now (except for this bug) is:
(unfolded-env
(unfolded-env a b)
(unfolded-env c d))
The next best thing:
(unfolded-env
(document
a
(detailed "" b))
(unfolded-env c d))
but then b always takes a line of vertical space.
Yes, I'm talking about the same version of TeXmacs all through this bug report.
(file #11513)
|