https://bugs.documentfoundation.org/show_bug.cgi?id=154075
--- Comment #12 from ady <[email protected]> --- (In reply to Heiko Tietze from comment #11) > (In reply to ady from comment #10) > > ...it is very rare for me to use relative _worksheet_ references. > > That's true. My proposed first toggle on/off should ignore what is set for > the sheet. When the reference has no worksheet name, that is how it works. When the reference includes a worksheet name, the worksheet name is also considered within the F4 cycle. Are you proposing that the F4 cycle should modify the Row and Column types only, while leaving the worksheet name to be modified "manually" (i.e. F4 would not act on the reference type of the worksheet)? FWIW, I would vote against, if anyone asks. > > > No user needs to use eight F4s on the same reference... > > If the sequence is always the same with ABCDE, you may start at A going to B > and vice versa from B to A. One is easy to achieve the other not so. If you > always start from A even when the actual state is B, this is less of a > problem, of course. Changing from C to D would be a challenge then, however. All my proposal was/is doing is changing the order of worksheet references' types as seen in comment 8. I am not proposing other changes in behavior, which you seem to be proposing(?). In the current F4s cycle, if the formula is (already saved, we are no longer in edit mode): =$Sheet2.A$1 ...which currently corresponds to what I called "step 6", and then I get into edit mode (press F2 while the cell with the formula is in focus) and then press F4 once, the formula changes to: =$Sheet2.$A1 ...which corresponds to the current "step 7". Even when I am pressing F4 just once since starting edit mode, it doesn't mean I'm going back to "step 1" with just that one F4. In my proposal, this is all exactly the same, except I called them "step 2" (instead of the current "step 6") and "step 3" (current "step 7") respectively. If the formula were to include several references, each one has its own "step" and they can be changed by positioning the cursor on any part of a reference or by selecting it, while the other references in the same formula are not modified by pressing F4. This is how it currently works, and my proposal changes nothing about it. This is all being described from the POV of users and how the behavior looks (and will still look). If the source code to keep this same behavior is more complicated than it is now, that is beyond my skills and knowledge. > > My point is that you cannot click through eight options in a reasonable > time. I don't know what that means. What reasonable time are we talking about? What clicks? I'm sorry, but I don't understand. Should I link to some youtube videos showing how this works for experienced Excel users, where the F4 method is _much_ more frequently used? LO Calc takes advantage of F4s in only a small part of the UX situations in comparison to Excel, even when Calc has relative references for worksheet and Excel doesn't. The whole point of using F4 is that it reduces both, clicks and typing "$" in specific positions. In Excel it has additional advantages too. > time. We may a) reduce the number of options or b) make it easier to access > by changing the sort order according the probability of needs. The cycle must be the always same (as per memory muscle). The proposal is limited, as already described. The "first" F4 moves the reference type to the "next" step. Starting from step 5, the first F4 modifies the reference to step 6, and so on. Starting from step 0 (zero) is just the "default" case when no edition was ever performed to the reference type. Moreover, the "default" type is not the same when using the Formula Wizard, but when later-on we use F4 (because sadly Calc's FW doesn't accept F4s), the F4s cycle is the same as described. It doesn't "start over" from step 0 (zero). No probabilities. No randomness, no assumptions, no different behavior depending on context. There is no starting over. The cycle is always the same: from "step 'n'" to "step 'n+1'". With eight F4s, we are back to whichever the starting step was before pressing the first F4. > > I have some doubts in these "smart" functions, on the other hand. It takes a > lot of assumptions and might be annoying for users. I can understand if you > reject my comments. Justin's idea is an improvement by itself. The reason for me to express the proposal from the POV of a common user is because Justin's expressions are as a developer, and I cannot be sure I interpret them as they were intended. Considering the amount of phrases I didn't understand already, I don't want to opine on any other proposal regarding F4 cycles unless I completely understand its consequences. I'm truly sorry I had to make this post so lengthy, but ATM I have the impression we are talking about different things and I needed to be explicitly clear about my proposal. If Justin presented a different idea than mine, then indeed I/we have a communication problem. Justin initiated this ticket as a response to my suggestion, which is summed-up in comment 8 here. If we are/were discussing other proposals, then I have no idea of what that is/was supposed to be. -- You are receiving this mail because: You are the assignee for the bug.
