Re: [Gimp-developer] maximize work area
I can presume that he asked about a checkbox in Preferences - to choose between a big tab with a thumbnail of a current canvas inside AND a much smaller tab with a small text containing the filename ( this way he can gain another 30 or 40px of vertical space) ..this could be somehow relevant for small 16:9 lanscape orientation screens. 2014-06-28 16:43 GMT+03:00 Alexandre Prokoudine alexandre.prokoud...@gmail.com: Ж-д 28 июня 2014 г. 17:26 пользователь Johann jo.y.v...@gmail.com написал: And have a checkmark entry in the application toolbar menus too, to complete things. What checkbox exactly? Please be more specific. Alex ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list
Re: [Gimp-developer] Blend Tool UI
mitch: The alt shortcut has always worked in my prototype. (You sound like you might not have been aware of this in your email.) I'm also not sure that the handle size should should necessarily be consistent with other tools. Gradients have only two endpoints, while paths or free selections can get significantly more complicated, which I think justifies having larger handles. Ofnuts: If a gradient is modified in any way, the preview will update properly. Simon: Spherical and Dimpled are actually portions of sinusoids, not quarter circles, but that might be what you meant. So, the transformations functions look like this: http://www.wolframalpha.com/input/?i=plot+1.0+-+sin%28x%29%2C+plot+cos%28x%29%2C++from+x+%3D+0+to+pi%2F2 I'm actually in favor of just removing the Spherical and Dimpled versions of the shaped gradients. There's not too much to differentiate between the three right now, and the transformation functions seem a little silly to me. I'd like to add in a euclidean-based shaped gradient though, and give the user a choice between euclidean and manhattan distance metrics. The manhattan ones aren't very useful alone. On Wed, Jun 25, 2014 at 6:53 AM, peter sikking pe...@mmiworks.net wrote: Maybe drawing circles most of the time, and then hiding the circles and switching to crosshairs while dragging the points, would be good? the alignment of the endpoint with the underlying content needs to also be reviewed when not moving the point. Fair enough. - make the control points big (30–50px diameter) and essentially transparent; clearly mark the centre where the co-ordinate is. My first naive implementation looked like this: http://i.imgur.com/ynQTFHi.png Those handles tended to grab your attention away from the image too much. I went back to my original ones briefly, and I have to admit that they are definitely more difficult to grab than the larger ones. So, I tinkered with them and came up with this: http://i.imgur.com/RK0odEE.png The user still has a 40 px grab target (the entire circle), but I now hide the circle whenever the mouse isn't within 60px of the center of the handle. So, they look quite minimal (just the cross) most of the time, and the grabbable area is still large. I'm generally satisfied with them. - you can drop the line because the gradient is there to represent itself; saves on clutter. I did remove the line, but I think I might like to add it back in if nobody's strongly opposed. Here's why: * I feel like the handles are minimal enough, even with the line present. * I'd like to implement mitch's idea of being able to drag the line to move both points at once. * apparently some people would miss the line (see Jason's post) * I'm planning on allowing the user to disable the preview, (which is something that basically all of gimp's tools have, just in case the user is working on a huge image where the preview would be painfully slow to update), so we can't always rely on the user seeing the on screen preview. At the very least, we should show the line whenever the preview is hidden. OK, I should have consulted the manual and now I have done it. I am now completely convinced that the shape burst belongs in the gradient tool. there is nothing contradictory about that. the gradient tool applies a gradient fill (as everything: modified by the selection) and Shape is a fill ‘mode’. and when talking interactive: next move would be to control these funky fill shapes on the canvas with a handle or two. Okay. I think I have a decent idea of how to control the shapebursts with handles. I'm going to focus on getting the other gradients working first though. Ofnuts wrote: good plan. combine it with updating the colours of the endpoints to make it truly adjustable to get it _right_ This would be updating the gradient while using it, but a gradient can be much more complicated than one color at each end. And what do you do with gradient colors that are not explicit but bound to FG/BG colors? well, I imagined straightaway that the endpoints would be highlightable and then modifiable through the regular color selection dialog(s). this point - that (adjusted) color a more complex gradient is defined by ‘waypoints’ on the canvas, while working interactive, the position where these waypoints fall on the gradient can be shown. then each of them can be highlighted and color-adjusted. when the points are there anyway on the canvas, users can move them, in 2 dimensions. and gee, once you got that. users can build a complex gradient out of nothing by: - starting with a begin and end color; - set up a multi-point path (click, click, click, double-click or return; to begin with: a straight-segment path; intermediate colours get interpolated, based on distance along the path) - update the points’ position and colours; - commit (another double-click or return). if this overloads Michael: you can introduce this
Re: [Gimp-developer] Blend Tool UI
On 29/06/14 21:08, Michael Henning wrote: mitch: The alt shortcut has always worked in my prototype. (You sound like you might not have been aware of this in your email.) I'm also not sure that the handle size should should necessarily be consistent with other tools. Gradients have only two endpoints, while paths or free selections can get significantly more complicated, which I think justifies having larger handles. Ofnuts: If a gradient is modified in any way, the preview will update properly. Simon: Spherical and Dimpled are actually portions of sinusoids, not quarter circles, but that might be what you meant. So, the transformations functions look like this: http://www.wolframalpha.com/input/?i=plot+1.0+-+sin%28x%29%2C+plot+cos%28x%29%2C++from+x+%3D+0+to+pi%2F2 I'm actually in favor of just removing the Spherical and Dimpled versions of the shaped gradients. There's not too much to differentiate between the three right now, and the transformation functions seem a little silly to me. I'd like to add in a euclidean-based shaped gradient though, and give the user a choice between euclidean and manhattan distance metrics. The manhattan ones aren't very useful alone. On Wed, Jun 25, 2014 at 6:53 AM, peter sikking pe...@mmiworks.net wrote: Maybe drawing circles most of the time, and then hiding the circles and switching to crosshairs while dragging the points, would be good? the alignment of the endpoint with the underlying content needs to also be reviewed when not moving the point. Fair enough. - make the control points big (30–50px diameter) and essentially transparent; clearly mark the centre where the co-ordinate is. My first naive implementation looked like this: http://i.imgur.com/ynQTFHi.png Those handles tended to grab your attention away from the image too much. I went back to my original ones briefly, and I have to admit that they are definitely more difficult to grab than the larger ones. So, I tinkered with them and came up with this: http://i.imgur.com/RK0odEE.png The user still has a 40 px grab target (the entire circle), but I now hide the circle whenever the mouse isn't within 60px of the center of the handle. So, they look quite minimal (just the cross) most of the time, and the grabbable area is still large. I'm generally satisfied with them. - you can drop the line because the gradient is there to represent itself; saves on clutter. I did remove the line, but I think I might like to add it back in if nobody's strongly opposed. Here's why: * I feel like the handles are minimal enough, even with the line present. * I'd like to implement mitch's idea of being able to drag the line to move both points at once. * apparently some people would miss the line (see Jason's post) * I'm planning on allowing the user to disable the preview, (which is something that basically all of gimp's tools have, just in case the user is working on a huge image where the preview would be painfully slow to update), so we can't always rely on the user seeing the on screen preview. At the very least, we should show the line whenever the preview is hidden. OK, I should have consulted the manual and now I have done it. I am now completely convinced that the shape burst belongs in the gradient tool. there is nothing contradictory about that. the gradient tool applies a gradient fill (as everything: modified by the selection) and Shape is a fill ‘mode’. and when talking interactive: next move would be to control these funky fill shapes on the canvas with a handle or two. Okay. I think I have a decent idea of how to control the shapebursts with handles. I'm going to focus on getting the other gradients working first though. Ofnuts wrote: good plan. combine it with updating the colours of the endpoints to make it truly adjustable to get it _right_ This would be updating the gradient while using it, but a gradient can be much more complicated than one color at each end. And what do you do with gradient colors that are not explicit but bound to FG/BG colors? well, I imagined straightaway that the endpoints would be highlightable and then modifiable through the regular color selection dialog(s). this point - that (adjusted) color a more complex gradient is defined by ‘waypoints’ on the canvas, while working interactive, the position where these waypoints fall on the gradient can be shown. then each of them can be highlighted and color-adjusted. when the points are there anyway on the canvas, users can move them, in 2 dimensions. and gee, once you got that. users can build a complex gradient out of nothing by: - starting with a begin and end color; - set up a multi-point path (click, click, click, double-click or return; to begin with: a straight-segment path; intermediate colours get interpolated, based on distance along the path) - update the points’ position and colours; - commit (another double-click or return). if this overloads Michael: you can introduce
Re: [Gimp-developer] maximize work area
Even if it's not a setting as such, it would be easy (from a UX perspective, probably less so from a coding perspective) to just have the tab section be collapsible. As an example of how something like that might look, consider the sidebars in darktable. They're generally useful, but you can collapse each one if you want more space, then you can bring it back when you need that functionality back. Better utility than a checkbox hidden in Prefs, but still gives you the option to make it go away. I've had the same desire when I'm working on a dark image, and I want to make everything bright go away. If I have a light image open, then it becomes impossible to make it go away. --xsdg On 06/29/2014 11:47 AM, SorinN wrote: I can presume that he asked about a checkbox in Preferences - to choose between a big tab with a thumbnail of a current canvas inside AND a much smaller tab with a small text containing the filename ( this way he can gain another 30 or 40px of vertical space) ..this could be somehow relevant for small 16:9 lanscape orientation screens. 2014-06-28 16:43 GMT+03:00 Alexandre Prokoudine alexandre.prokoud...@gmail.com: Ж-д 28 июня 2014 г. 17:26 пользователь Johann jo.y.v...@gmail.com написал: And have a checkmark entry in the application toolbar menus too, to complete things. What checkbox exactly? Please be more specific. Alex ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list
Re: [Gimp-developer] Blend Tool UI
I like the idea of shaped gradients being handled in the Blend Tool. If the length of the stroke determines the maximum distance from the selection boundary, it would provide a quick means of stroking the selection or creating shadows and highlights with interactive control of the thickness. With regard to moving the blend stroke, I think that ALT+drag should be required, to be consistent with the behavior of the Vectors Tool. For further consistency with the Vectors Tool, CTRL-drag on the blend stroke should add a new stop and then repositioning all the stops while dragging. With no modifier key pressed while dragging along the blend stroke, no new stop should be created and the existing stops should repositioned based on the drag. Any mouse drags that are not on the blend stroke or its endpoints should result in the existing blend being committed and a new one commenced. However, a mouse click by itself that is not on the blend stroke or one of its endpoints should result in just the blend being committed. ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list
Re: [Gimp-developer] Wayland wacom support: binding tablets to the canvas
On Fri, Jun 27, 2014 at 10:47:12AM +0200, Michael Natterer wrote: On Fri, 2014-06-27 at 11:54 +1000, Peter Hutterer wrote: [Please CC me on the reply, I'm not on the list] Hey guys, Hey Peter, As you may be aware we're currently working on getting wacom graphics tablet support into Wayland. Lyude (on CC) has sent out an inital draft for the protocol, archived here: http://lists.freedesktop.org/archives/wayland-devel/2014-June/015583.html Dmitry from the Krita project has also triggered a discussion here. http://lists.freedesktop.org/archives/wayland-devel/2014-June/015605.html This is all still in flux, so don't expect the current draft to be the final project, but now's the time to wish for what you need. All that aside, here's a specific question: GIMP has/had a feature where the surface of the tablet could be mapped to the canvas on the screen. Is this a feature that is currently used or in demand, or is this something we don't need to worry about? If we do need to consider it, what are the requirements here? Many people are not even aware of this feature, because this window mode has been quite broken for a while in GTK+ 2.x, however in my opinion it is very useful. GTK+ used to handle that by translating the coordinates, I don't know this works with Wayland, does Wayland have to care about that at all? Does it have XInput-like semantics of master/slave/floating? Or is all of that handled in the GDK backend? tbh, I'm not sure yet. The feature doesnt' exist as such yet, we're discussing whether it's needed and if so how to integrate it. There are at least three pieces in the stack before GIMP sees anythying and they all need to work together. The current idea so far is to have a bind request to explicitly connect the tablet to a given client. How this will be exposed to clients is unclear at this point, we're just in the drafting phase. it'd be nice to know how GIMP would like to handle this bit. Is it something that you'd enable generally for the tablet, or on request (e.g. when the focus takes effect), etc.? Cheers, Peter ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list
Re: [Gimp-developer] Blend Tool UI
On 29.6.2014 at 11:19 PM Ofnuts wrote: I'm not convinced that shape bursts still need a complicated UI. +1 Once you have a nice clean black-to-white linear fill you can always use the Levels or Curves (or a gradient map) on it to achieve whatever you want. How would you bend and distort a straight linear fill (i.e. from left to right) to make it fit into an arbitrary shape with the Levels or Curves tool? They are Color tools and all they can do are tonal corrections. In our case they can only shift the gradient a bit more to the left or right or make it wider or slimer. I think in the context of gradient fill these tools a) don't meet the needs (see above), b) they add an extra step. Modifying a linear, circular etc. gradient with handles, but other gradient shapes with the Color tools is simply confusing. Extra steps also impede the users in working fastly. Therefore Color tools can be taken as optional step for fine tuning a gradient, but never be part of the default filling workflow. I also agree that too many options are more confusing than helpful, but let's not drop useful functionality. Kind regards, Sven ___ gimp-developer-list mailing list List address:gimp-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list