GNOME Bugzilla – Bug 86888
Set workspaces with (length,width) instead of (total,num rows)
Last modified: 2020-11-06 20:22:44 UTC
Right now, the workspace preferences allows you to change the number of workspaces and the number of rows in which they are displayed. Although this reflects the desktop's underlying implementation (window manager workspaces), I think that this interface doesn't reflect the user's perspective. Workspaces are shown to the user as a 2-dimensional grid, so I think that the interface should let you specify the number of columns and rows directly, rather than indirectly. If I want a 2x4 desktop, it is mentally cumbersome to make 8 workspaces in 2 rows, for instance, it makes more sense to specify 2 rows and 4 columns. Furthermore, the current way of doing things leads to (workspace #, row) combinations that don't make sense. For instance, it doesn't make sense to have 2 rows of 7 workspaces -- the bottom-right corner will be blank. Allowing the user to specify rows and columns directly avoids this problem.
Agreed that this is terribly broken. Might even be a dup, though I can't find it offhand. [If it is a dup, it doesn't have the usability keyword and it should.]
Updating keywords. Mark? Mark? :)
Agreed that it does seem a bit weird. But one thing to consider is that probably the fundamental thing you want to configure is the number of workspaces. e.g. I've twelve workspaces .. thats the number I want .. and what I want to do is play around with having them on one row, two ros or three rows. Maybe that's more common than "I have four workspaces and I want to play around with having one, two or thre rows of four workspaces"
I think that's a subjective issue. I don't find it particularly compelling to choose the number of workspaces a priori (before choosing the layout of the workspaces).
Ok, so, usability people: do you have some wise thoughts about this?
Michael's right, I think this is one of those situations where it's almost impossible to guess how users are thinking, we really do need to ask a good few and see what they say. I think we'll probably find a few different types of people, but I don't know in what proportions: - people who add worskpaces as and when they run out of space on the default workspace layout; - people who just start minimising or closing windows as and when they run out of space on the default layout; - people who've used the same workspace layout for years, and set up any new desktop the same way; - people who'll only use one workspace however many you give them :)
Lowering priority and serverity. I'm tempted to mark this as WONTFIX. Usability people: if you think this should happen (or we don't have enough informations yet to close the bug), now is the time to speak up :-)
I think there's definitely still room for improvement here, although I agree it's a fairly low priority. Personally I'd like to see something a bit more interactive in the dialog, akin to the representation on the panel.
I don't think that any of the four usage patterns that Calum proposed are better with the current model of (num workspaces,num rows) than the (x,y) model I'm arguing for. Even in the second case, where you want to add new workspaces as you run out, the task is accomplished just as easily if you add the workspaces in terms of (length,width) than if you add them in terms of (num workspaces,num rows). In general, I don't think that an analysis of usage patterns can provide an argument for either model of interaction, because the methods of accomplishing tasks in each model is so similar-- you're just increasing or decrease a couple of sliders in either case. For example: 1a) To add a row in the current model: increase workspaces by n, then increase rows by 1 1b) To add a row in the x,y model: increase y by 1 2a) To add a column in the current model: increase workspaces by n 2b) To add a column in the x,y model: increase x by 1 (Note that the x,y model never requires more actions than the current model, and sometimes requires less. This is assuming you don't want a workspace grid that does not fill the entire grid, such as 3 workspaces in two rows, which doesn't make sense, even though the current model allows it.) In general though, there really isn't that much difference in the *actions* people take to accomplish tasks in either model--the reason for using the x,y model is just because of consistency. The pager is displayed as an x,y grid. The workspaces are laid out as a giant x,y grid. Everything in the *interface* is based in x,y coordinates, it's only the implementation that uses a workspaces/rows coordinate system... and only the workspace control panel that exposes this implementation.
Wait, I guess there IS a situation in which the status quo (num,rows) model requires fewer actions: switching from an (n,m) grid to a (m,n) grid, or other cases like what Mark was suggesting. But man, if people are really thinking about rearranging their workspace grid in terms of conservation of workspaces, they're freaking geniuses. I didn't even think about this technique at first, and I'm a programmer. I mean, kids only develop the cognitive faculties for reasoning about conservation of volumetric solids and liquids at like age 8, right? Maybe it's 5. Before that age, you can give kids a short fat glass and a tall skinny glass, show them that if you fill the tall glass with liquid and then empty it into the short glass, the liquid exactly fills up both glasses, and they will still think that the tall glass holds more water. Reasoning in terms of conservation of ____ is a late-developing cognitive ability that requires some advanced abstract reasoning. This is a pretty well-known psychological issue. The (x,y) model is direct-- you want to increase width, you just increase the width slider... you don't need to figure out how many rows and columns you need to divide by to get that amount. Anyway, I'm a fan of direct manipulation. Go Shneiderman.
I suggest changing to wontfix or wishlist. I don't get why choosing workspaces and rows doesn't work well enough.
We need to change the representation because, as Michael explained in comment #9, the whole workspace concept is based on a grid. The pager displays it as a grid, the switcher displays it as a grid, and they are switched using the up/down/left/right actions. For now, in the pager preference, if you ask to only display the current workspace you _can't_ select the amount of rows, which is purely wrong because it still matters. I suggest to change the preference dialog box to: * Selecting wether all the workspaces or just the current workspace is displayed * Specifiying the workspace grid size N x M * Choosing their name (left as today, linearized)
Created attachment 110283 [details] OS X Leopard's Spaces config dialog FWIW, I've attached a screenshot of the Spaces configuration dialog in OS X Leopard. I don't think it's particularly elegant, but it might be useful to get the discussion going again...
bugzilla.gnome.org is being replaced by gitlab.gnome.org. We are closing all old bug reports in Bugzilla which have not seen updates for many years. If you can still reproduce this issue in a currently supported version of GNOME (currently that would be 3.38), then please feel free to report it at https://gitlab.gnome.org/GNOME/gnome-panel/-/issues/ Thank you for reporting this issue and we are sorry it could not be fixed.