MJW

OptiPNG might help. For JPEG, you can specify increased compression when you save the image.

Making an image of the word "GARAGE" distorted by a transparent button is possible with PDN, for instance, by using the technique I used with the transparent heart. Turning that into a button in a program is outside the realm of PDN.

You'll have to install another plugin, but if you want to draw a sphere, you should try MKT's Shape3D. It's a very useful plugin most everyone should have, so it's worth downloading.

A previous thread on the layer-deactvated-when-made-invisible topic gives a least a ray of hope from Rick Brewster. I would certainly edit the wrong layer less often if making a layer invisible didn't deactivate it.

I think you need to describe it in terms of what kind of image you want to produce, not what you want a button to do. Paint.Net can't make your buttons do anything, it can only help you produce images. If you can provide examples of the sort images you want, that would be helpful. I'll mention a Water Drop Tutorial, which might or might not be related to your magnifying glass idea. A comment I made containing an image of a glass heart in front of an image on cloth may also be related (or not).

Congratulations to @Pixey, @Woodsy, and @lynxster4 and @welshblue! Some really fine entries. Pixey, I loved that rectangular light reflection. If I hadn't run short of time, I would have tried to steal the idea. Thanks to @toe_head2001 for hosting Yes, Pixey, I produced the top using the Texture Merger, along with the (still beta) Texture View Skewer to change the viewpoint. (That's a plugin I need to update and release.) I noticed some very impressive tops among the other entries. I have no idea how they were produced. Without the Texture Merger method, I would have been stuck.

It could be made more efficient by determining the centering values only once. bool firstTime = true; int x1, x2, y1, y2; void Render(Surface dst, Surface src, Rectangle rect) { if (firstTime) { // find centering variables, x1, x2, y1, y2. . . . firstTime = false; } . . . } If the centering variables were possibly modified while being computed (set to one value then another), I'd rewrite it or store them in local variables until the end, at which point I'd assign them to the "global" variables. E.g.: if (firstTime) { int locX1; locX1 = sLeft; if (locX1 > 100) locX1 = 100; x1 = locX1; firstTime = false; } I don't think that's absolutely necessary, but it's the safe way. The reason is, it would still work if you take off the firstTime test. Otherwise with multiple processors, a thread could access a variable during the time a different thread had set it to the non-final value. The basic rule is that a "global" (class-level, really) variable must be the same for every thread, and can never be assigned anything except its final value. BTW, what's with the dst[0, 0] = dst[0, 0]? I don't see what it's for, but in any case, it seems wrong to assign to a pixel outside the ROI. I think it could conceivably cause a problem, also (assuming it's not optimized out). A thread could load the 0,0 pixel just before the thread assigned to that ROI updates it, then store back the old value over the updated value.

I think you should start a new thread. This one is from 2005. Let it rest in peace. The rules:

The only real PDN developer is Rick Brewster. I doubt he's working on it. If you expect it soon, I doubt your expectations will be met.

I suggest in your opening comment you state a bit more directly what you want. I think what you want is a tool option to automatically complete the tool action without a Finish.

You can certainly start a new thread suggesting that feature be added to the tools settings. I suspect quite a few users would like it. Whether it would be implemented is another matter. You can use the Enter key to Finish an action instead of clicking, if that helps any.

Glad it works.

In that case, I expect the values I gave should work. Each control is independent: it only cares about the values that enable it, so there's really no additional complexity in enabling two slider controls with one list control. If you were trying to enable a single control based on the values of two controls, that would be more complex -- and as far as I know, impossible.

I'm not clear what each option is supposed to do. For instance, what the difference is between Option2 and Option5, both of which are labeled "Slider X, Y". it's certainly possible to enable any combination of the sliders with the list box. Assuming the labels specify which controls to enable: object[] condX = new object[] {Amount9Options.Amount9Option2, Amount9Options.Amount9Option3, Amount9Options.Amount9Option5, Amount9Options.Amount9Option6, Amount9Options.Amount9Option8}; object[] condY = new object[] {Amount9Options.Amount9Option2, Amount9Options.Amount9Option4, Amount9Options.Amount9Option5, Amount9Options.Amount9Option7, Amount9Options.Amount9Option8}; You probably have something more complex in mind.

When using the Paintbucket to do a fill, once the little square and arrows are active, I think the keyboard arrows should move its location, and not the location of the Paintbucket icon.

To make the box disappear, hit Finish in the toolbar, just above the canvas. It's at the right side, by the green check mark. The way it works is, clicking the Paintbucket causes the little square, along with the Move arrows to appear. You can then use the mouse to move the box around, using the move arrows. The box specifies the location to where the fill will start. The fill completes when you either do something like start another fill, or click Finish. There's no way to make it do something different, and you get used to it (at least I did).

I can't say I fully understand what you want to do, but perhaps Paste From Clipboard would help.

I believe the proper format is: object[] conditions = new object[] { MyDropDownOptions.MyDropDownOption2, MyDropDownOptions.MyDropDownOption3 }; propRules.Add(new ReadOnlyBoundToValueRule<object, StaticListChoiceProperty>(PropertyNames.MySlider, PropertyNames.MyDropDown, conditions, false)); EDIT: A more complete explanation is here.

I've used Paint.Net for quite a few printed greeting cards. Seems to work fine for me.

Welshblue, f you have more than one layer, the Image setting will return the pixel color of all the layers blended together, while the Layer setting will give you the pixel color in the currently active layer. For a simple case. If you have two layers, the top opaque and white and the bottom opaque and black, and the bottom layer is active, Image will return white, and Layer will return black. If the top layer had 50% opacity, Image would return gray.

Another, perhaps even better, solution is to: Set the Primary Color to a transparent color. Select the Line/Curve Tool. In the Tool Bar (just above the image area) set the Blend Mode to Overwrite. (For the Blend Mode, look for the flask icon in the Tool Bar.) Now whatever you draw will be erased from the image. This, of course, also works with all the other drawing tools, such as Shapes.

I don't think you really want to precisely select an area -- that's just a means to an end; what you really want to do is precisely erase an area. Probably the best way to do that is to use BoltBait's Paste Alpha plugin. Draw the curve on a separate transparent layer. Copy the the curve layer to the clipboard (no need to select the layer -- just use the Ctrl+C keyboard shortcut). Make the separate layer invisible, so it won't distract you. Make the layer you want to modify the active layer. Run Object>Paste Alpha. Set the Alpha Source to Clipboard alpha. Check Invert calculation. Set Paste method to Minimum of alpha and clipboard alpha or (Multiply alpha and clipboard alpha). This generally works far better than using the Magic Wand and Deleting.

I say, put them out of their misery.

I like the can-of-soda theme fine, but I think the rules regarding the artwork on the can should probably be spelled out. Is it to be original, a copy of an actual soft drink (like Coke or Pepsi), or either?