On Mon, Sep 03, 2007 at 09:46:11PM -0400, David Pokorny wrote: > Thanks for the lively discussion :) > > This is actually a somewhat common problem for me, and I understand > the value of asking that new users acclimate to a new interface, but > in this particular instance, I have to disagree: > > First, the current behavior is a small (but important) violation of > the principle of least astonishment: in all other office applications, > typing a backspace either has a small effect or it deletes the current > selection. I have also never observed the behavior of "typing a > backspace selects something" in any other office application.
If the chunk in front of the cursor is big you have the the options of - doing nothing - selecting the big chunk - immediately delete the big chunk - doing something even weirder. None of which fits in your describtion of 'Typing a backspace either has a small effect or it deletes the current selection. I have also never observed the behavior of "typing a backspace selects something" in any other office application'. > Second, the canonical workaround (typing a left arrow key) is a burden > since it means the user has to move his/her right hand to the arrow > keys. But only because nobody came up with vi bindings so far.... [semi-joking] > This constitutes an interruption to the flow of typing if the > user finishes a formula, types a few words outside a formula, then > decides to rework the formula. Personally this is a very common > activity, but I would hazard a guess that any user who is, say, typing > up marginally legible class notes, would be in this position > moderately frequently. The issue is not the time saved (which is > trivial) but rather the necessary interruption of flow. > > So the argument is that the following change to the behavior of the > backspace key is both easier to learn for new users and more useful > for all users: > > * Both inside and outside math, typing a backspace when the cursor is > adjacent to the right hand side of an inset moves the cursor to the > last position of the left-adjacent inset; typing a backspace has an > effect identical to typing the left arrow key. Nah. This would make it impossible to delete insets by using backspace. Conceptually an inset is, well, an inset. I.e. one entity equal to a single character. If I put the cursor at the end of a document and put the cat on the backspace key I expect the document completely deleted after a while. This would not be possible with your solution as it would leave empty insets behind. > To the positive argument that users get used to the existing behavior > in practice, I would agree that it is possible for users to get used > to this behavior, but I would add the normative argument that users > should not have to adapt to accommodate a behavior that is (even > marginally) astonishing. Well, so far I've not seen the request that backspace should move left instead of deleting to the left, so I take this as indication that during the last decade we either had no users or the user did not use backspace or they were sort-of-happy with backspace deleting things. Andre' PS: Please do not top-post with full quote attached to the mail. I know that I am fading out of this world but I still have this habit of vaguely rmemebering the stuff I wrote yesterday.
