David Pokorny wrote:
The point is that a user does not think of a formula as chunks of anything - of course they are implemented as nested insets, but they all look like characters on the screen (not images or tables, unless there really is a table involved) and I expect that pressing backspace = delete one character.
It would be interesting to do a poll on this. I think of insets as individual entities, so when I backspace over an inset (any kind of inset -- math, figure, note, ...) I expect the entire inset to be vaporized. The same goes for tables, although technically I don't suppose they're insets. I can see your point of view, but I'm pretty sure we'll get a split on this.
Just for grins, I cranked up Word Perfect (which I use less and less often, thanks to LyX) and Word (which I use only when a large-caliber weapon is held to my head) (and only after verifying it's loaded). In WP 9, backspacing over a formula deletes it; backspacing over the boundary of a table enters the last cell of the table, at the end of the content, but does not delete anything. In Word 2000, backspacing over a formula highlights the entire formula; backspacing a second time deletes the entire formula. You never enter the formula. (At least this is true if the formula is produced by either the WP or OO equation editors. I don't seem to have the MS equation editor.) Backspacing while the cursor is adjacent to the last cell of a table pops up an error message saying the operation is illegal at row's end.
So, depending on what you used before LyX (and perhaps still use for some things), expectations for the role of the backspace key could be all over the map. I happen to like LyX's current approach (clobber whatever entity precedes the cursor).
/Paul
