Description: UI: Hard to tell which side of a shape being used for as reference for rotation Steps to Reproduce: 1. Open the attached file 2. Change the rotation of the vertical rectangle to horizontal Actual Results: * The rectangle initially drawn horizontal has 0 degree rotation * The rectangle drawn also has 0 degree rotation. Converting to horizontal means a 90 degree (or -90 degree) rotation. Where the negative or positive rotation doesn't matter, until you you add text to a shape :-(. Which you can fix by rotation the shape (or more by rotating the text itself). This rather common experience, IMHO. Lets assume some - like me - re-using a shape (by copy paste) initially drawn say vertically but used horizontally. The horizontal shape will have 90 degree angle. Really counter intuitive Opposite happens to: copy/paste of horizontal shape rotation to vertical. So horizontal shape getting angle of 90, which feels natural (to me) Expected Results: No clue A) Use one angle as basepoint (say horizontal). So vertically drawn object is has automatically a 90 degree angle. No clue if this being workable B) Some visual indicator on the shape itself marking where the top side of the shape is; improving the UI feedback C) Both D) Something else FWIW: I didn't do an extensive comparison with other apps (yet). Quick MS Word appears to do A & B Reproducible: Always User Profile Reset: No Additional Info: Version: 24.8.0.0.alpha1+ (X86_64) / LibreOffice Community Build ID: e3bd3c7e3178dc091fd002628f052666b4db3be6 CPU threads: 4; OS: Windows 8.1 X86_64 (6.3 build 9600); UI render: Skia/Raster; VCL: win Locale: nl-NL (nl_NL); UI: en-US Calc: CL threaded
Created attachment 194566 [details] Sample
Sorry, can't you just switch width and height to get the desired result instead of rotating?
If the properties pane of the side bar is open and you click on a shape, you can see the current angle there. Do you think of something similar to the rotation-handle in MS Office? There is only one handle and it is on the originally top of the shape, so that you can identify the originally top when you click on the shape. Or do you want to have an additional info in the status bar? Both would be an enhancement request.
(In reply to Regina Henschel from comment #3) > Do you think of something similar to the rotation-handle in MS Office? There > is only one handle and it is on the originally top of the shape, so that you > can identify the originally top when you click on the shape. Yes. That would surely improve the end-user experience. I have changed the status to enhancement request
Sounds to me like an artificial use case. In what situation is it crucial to know where the original top-left point is located?
(In reply to Heiko Tietze from comment #5) > Sounds to me like an artificial use case. In what situation is it crucial to > know where the original top-left point is located? What exactly do you perceive as artificial use case? It really depends how draw an object (especially rectangle), where the top of a shape is. Rotating an object which is drawn horizontally feels 'natural'. However if the shape is drawn vertically everything becomes upside down (metaphorically). A shape having a 90 degree rotation resulting in horizontal orientation (as long the same doesn't contain text) is rather counterintuitive without UI feedback. The problem above doesn't occur when drawing a line. The drawn object has a rotation set automatically. So if you draw line horizontal 0 degrees. Draw a line vertically 270 degrees. So it might be that that's the bug. In which case the rotation handle being a nice additional feature (comment 4), but not relevant for the problem experienced here.
(In reply to Telesto from comment #6) > What exactly do you perceive as artificial use case? The rectangle has axial symmetry. Whether you create a rectangle with 1x5cm or 5x1cm + 90°- both look the same and I see no use case to make the difference more clear. Did you ever run into any trouble or are there any questions on ask.libreoffice, for example.
Created attachment 194623 [details] Text in shape on optical 5x2 rectangle The difference becomes visible when you add text to the shape.
(In reply to Regina Henschel from comment #8) > The difference becomes visible when you add text to the shape. And with text you see what's going on. Why add some indicator, and in what situation?
(In reply to Heiko Tietze from comment #9) > (In reply to Regina Henschel from comment #8) > > The difference becomes visible when you add text to the shape. > And with text you see what's going on. Why add some indicator, and in what > situation? If I understand Telesto correct, then he copies a shape that has no text, inserts it and wants to add text. And only then he detects that the shape is not suitable. He wants to notice it earlier.
Created attachment 194626 [details] Screencast (In reply to Regina Henschel from comment #10) > If I understand Telesto correct, then he copies a shape that has no text, > inserts it and wants to add text. And only then he detects that the shape is > not suitable. He wants to notice it earlier. This is indeed the one part of my experience The other half is about seeing horizontal oriented rectangle shape, but having 90° rotation; it for me counter intuitive. And doesn't make any sense until you add text. Even if you add text, you ask yourself, what's wrong here... why is the text placed top/down.. There is no UI feedback An a drawing easily becomes chaos: And if you have 10 rectangles at various degrees: 270 degrees some at 180 degree some at 0 degree some at 90 degree but only shown horizontal or vertical... what is going on? And you want rotate a few of those.. the rotation itself becomes a challenge and distracts from what you're actually are drawing Turn the shape upside down by 270° degree rotation is odd. A shape with 180° being upside down feels natural. If i asked you to draw a rectangle with a 270° rotation without further information.. what of a result would I get? Vertical or horizontal oriented shape? Or reply: I can't because lack of baseline rotation. You don't think of a shape with a 270° rotation to be upside down containing vertical text --- The whole rotation angle stuff would be less noticeable WYSIWYG with a rotation handle on the shape/ image itself, instead of fiddling with dialog control.
Created attachment 194635 [details] Screenshot (In reply to Telesto from comment #11) > The whole rotation angle stuff would be less noticeable WYSIWYG with a > rotation handle on the shape/ image itself, instead of fiddling with dialog > control. Have to back paddle on this part of my comment: There is (are) rotation handles when slowly double clicking a shape; however the accessibility is a different topic
My take: WF/NAB
(In reply to Telesto from comment #0) > UI: Hard to tell which side of a shape being used for as reference for > rotation Are you sure that is exactly the problem? See below > This rather common experience, IMHO. Lets assume some - like me - re-using a > shape (by copy paste) initially drawn say vertically but used horizontally. > The horizontal shape will have 90 degree angle. Really counter intuitive In itself, this is not counter-intuitive to me, nor does it matter. The problem, I would say, is with the ffect you mentioned earlier: > the negative or positive rotation doesn't matter, until you you add > text to a shape So, I would say that the meaningful issue you're pointing out is that, on one hand, the shapes looks the same, but on the other hand, they have significantly different behaviors. What other behavior distinguishes the shapes other than the text block? Because, for the text block, one could argue that once you "enter" the block, you see a rectangular frame for the text block itself, that lets you know what you can expect when typing. Please explain why that is not good enough. > Opposite happens to: copy/paste of horizontal shape rotation to vertical. > So horizontal shape getting angle of 90, which feels natural (to me) I didn't understand this sentence. Nor your definition of a "horizontal shape". Do you mean a 2D shape for which the page-horizontal extent is larger than the page-vertical extent? > A) Use one angle as basepoint (say horizontal). So vertically drawn object > is has automatically a 90 degree angle. No clue if this being workable Don't quite get this either. You seem to have defined a "vertical shape" and a "horizontal shape", but what does it mean to be "vertically drawn"? > B) Some visual indicator on the shape itself marking where the top side of > the shape is; improving the UI feedback How would this be useful other than for knowing how the text area behaves? (In reply to Heiko Tietze from comment #5) > Sounds to me like an artificial use case. Disagree that drawing rectangles of various shapes and rotating them is artificial. I mean, the reproducer is a tiny document so it's always going to feel a bit artificial.
(In reply to Eyal Rozenberg from comment #14) > (In reply to Telesto from comment #0) > > UI: Hard to tell which side of a shape being used for as reference for > > rotation > > Are you sure that is exactly the problem? See below You're right, this isn't the core problem. FWIW: when reporting an it's always flipping between reporting a 'solution' with a description of the problem. Or more a description of experience without clear cut answer to the solution (and sometimes even real clue why the experience being off) > > > This rather common experience, IMHO. Lets assume some - like me - re-using a > > shape (by copy paste) initially drawn say vertically but used horizontally. > > The horizontal shape will have 90 degree angle. Really counter intuitive > In itself, this is not counter-intuitive to me, nor does it matter. At the same time: Yes, you right and No, it does matter. Yes, you are right: It doesn't matter as long I can archive my goal. It would have gone unnoticed if I used the WYSWING rotation mode .uno:ClickChangeRotation. However I struggled to realize it exists (Draw/Impress have it; Writer does not) and properly accessing it (see bug 161500). So I did look at sidebar -> rotation. With the experience that a visually similar oriented shape (rectangle) can have a 0 degree rotation or 90 degree (compare green rectangle and red rectangle in attachment 194623 [details]). Also if you rotate the green rectangle 90 degree (right). Having 2 shapes with same rotation degree set delivering a different result > problem, I would say, is with the ffect you mentioned earlier: > > > the negative or positive rotation doesn't matter, until you you add > > text to a shape > > So, I would say that the meaningful issue you're pointing out is that, on > one hand, the shapes looks the same, but on the other hand, they have > significantly different behaviors. Yes > What other behavior distinguishes the shapes other than the text block? The rotation seen in the sidebar/dialogs. > Because, for the text block, one could argue that once you "enter" the > block, you see a rectangular frame for the text block itself, that lets you > know what you can expect when typing. Please explain why that is not good > enough. > > Opposite happens to: copy/paste of horizontal shape rotation to vertical. > > So horizontal shape getting angle of 90, which feels natural (to me) > > I didn't understand this sentence. Nor your definition of a "horizontal > shape". Do you mean a 2D shape for which the page-horizontal extent is > larger than the page-vertical extent? > > > A) Use one angle as basepoint (say horizontal). So vertically drawn object > > is has automatically a 90 degree angle. No clue if this being workable > > Don't quite get this either. You seem to have defined a "vertical shape" and > a "horizontal shape", but what does it mean to be "vertically drawn"? I sometimes have hard time expressing myself; sorry. I hope the illustration given based on attachment 194623 [details] helps > > > B) Some visual indicator on the shape itself marking where the top side of > > the shape is; improving the UI feedback > > How would this be useful other than for knowing how the text area behaves? Well it's suboptimal solution, IMHO. The core issue: the experience that a visually similar oriented shape (rectangle) can have a 0 degree angle or 90 degree angle. And follow-up on that the adjusting the angle becomes math. Rotate the green shape (with text) (attachment 194623 [details]) 45 degree. Now rotate the red shape (with text), to match the green shape. I have to really think/or use the try and error mode to conclude:315 / 135 degree. Normally I mindlessly adjust to the same value (45 degree). Also notice that the label text is hard to read. So you want to adjust that (already an additional step I didn't intend to do). To conclude: there is no way to adjust the label orientation (as far I can tell). So no need to redo you work by re-draw the rectangle, I suppose. A (simple) rotating action which I normally perceive as something done mindlessly becomes mind numbing, time consuming frustrating activity. The application working against me (in my perception). And no it's hard to draw a rectangle with a 90 degree rotation
IMHO what could be workable would be an adjustment to our rotation symbol. It should remain at the center of rotation, but could simply have a leg/pointer extending in the vertical direction from origin of the object when created. As the object is rotated, the rotation symbol would rotate to match.
skipping all discussed details: me too often has to find out by doing how rotation goes. So if a reasonable improvement is possible: +1
We discussed the topic in the design meeting. Some kind of indicator might come handy sometimes. It could be a symbol such as an angle or some icon. Ideally shown when the object is selected. Alternatively a different color/size of the top-left edge indicator could work too. Or we change the rotation mode completely by removing the slow double click and show a rotation interaction on the top edge similarly how MSO does.