On Sat, Nov 12, 2011 at 12:46 PM, Tavmjong Bah <tavmjong@...8...> wrote:

On Sat, 2011-11-12 at 19:57 +0100, Johan Engelen wrote:

...

Hi all,

I don't know what to work on anymore... Please help me out by test driving the Powerstroke LPE and tell me what you dislike and what you'd like to be added.

Looks good. One request and one comment:

Request: To be able to control the angle of the butt end like you would have with a calligraphy stroke. You could use the placement of the node closest to the end to control the angle.

Comment: 'Smoothness' maybe not the best term. Try setting it to 1.0 and the line is anything but smooth.

Johan,

It appears Illustrator got a new tool in CS5, the Width Tool which is pretty much like Powerstroke. Their UI for it appears pretty well thought out and it sounds like it has handy modifier functionality. It may be a model that we want to look into following or basing some features on. Here is a video demonstrating it's use:

http://www.youtube.com/watch?v=lHhk4ku0c74#t=55s&hd=1

Cheers, Josh

On 11-1-2012 15:08, Defiant00 wrote:

Hello,

Having just tried both the nightlies and the latest version of Synfig I wanted to come give some feedback.

I believe one of the main uses for Powerstroke (at least how I'm hoping to use it) is to be able to modify stroke width while still maintaining the ability to fill the original shape; whereas the current implementation applies both stroke and fill properties to the powerstroke as if it is a shape (which I realize it is).

Would it be possible to either modify this behavior or provide a toggle? For me the ideal behavior would be what Synfig does; which is the equivalent of the powerstroke fill being set to the path's stroke color and then the original path being filled with the original fill.

I realize I can do this manually by duplicating the path then applying the powerstroke to the topmost path, but then the shapes are not linked, which isn't ideal for future edits.

I'm afraid this cannot be done with LPEs. The result you want to obtain (variable stroke on a filled object) cannot be represented in SVG as a single path. Therefore, it cannot be done with an LPE. There is however a trick that we can play..... I could make an LPE that simply outputs the _original_ d from another path. Effectively cloning the _original_ path data...

Ciao, Johan

On 11-1-2012 21:22, Defiant00 wrote:

Johan Engelen-4 wrote

...

I'm afraid this cannot be done with LPEs. The result you want to obtain (variable stroke on a filled object) cannot be represented in SVG as a single path. Therefore, it cannot be done with an LPE. There is however a trick that we can play..... I could make an LPE that simply outputs the _original_ d from another path. Effectively cloning the _original_ path data...

Ciao, Johan

Sorry for my lack of knowledge about how LPEs work; would the steps to use such a 'trick' just involve adding an extra LPE to the effect list chain?

If so, would this allow the original and powerstroke fills to be different, or would they share fill and stroke settings? If the settings are shared then that unfortunately wouldn't be of much help.

You would have to create a second path and apply a "Clone original path" LPE. The second path acts as a dummy, because the actual path is not used at all. It simply copies the original-d path data from another path. See it as cloning original path data, instead of final path data (after LPE) that a normal clone does. The first and second path can have completely different styles, and are only linked by their path data.

Ciao, Johan

On 11-1-2012 22:32, Defiant00 wrote:

Johan Engelen-4 wrote

...

You would have to create a second path and apply a "Clone original path" LPE. The second path acts as a dummy, because the actual path is not used at all. It simply copies the original-d path data from another path. See it as cloning original path data, instead of final path data (after LPE) that a normal clone does. The first and second path can have completely different styles, and are only linked by their path data.

Ciao, Johan

Ah I see, that would be an excellent LPE to have both for this and other places I and others might want to completely restyle a clone. Would it be safe to assume that it would just take the ID of the path to clone so that it would update whenever the original does?

I don't know if there's a formal procedure to request such a feature be added, but it'd definitely be greatly appreciated.

No need, you got me interested, so I just made and committed it. :-) Have a look at the new 'Clone original' LPE. I have attached a file that shows how you can use it. Interestingly, this also allows for 'paths' where the stroke is only applied outside the path data, by putting the filled original-pathdata-clone on top :)

Cheers, Johan

On Thu, 2012-01-12 at 22:09 +0100, Johan Engelen wrote:

On 11-1-2012 22:32, Defiant00 wrote:

...

Johan Engelen-4 wrote

...

You would have to create a second path and apply a "Clone original path" LPE. The second path acts as a dummy, because the actual path is not used at all. It simply copies the original-d path data from another path. See it as cloning original path data, instead of final path data (after LPE) that a normal clone does. The first and second path can have completely different styles, and are only linked by their path data.

Ciao, Johan

Ah I see, that would be an excellent LPE to have both for this and other places I and others might want to completely restyle a clone. Would it be safe to assume that it would just take the ID of the path to clone so that it would update whenever the original does?

I don't know if there's a formal procedure to request such a feature be added, but it'd definitely be greatly appreciated.

No need, you got me interested, so I just made and committed it. :-) Have a look at the new 'Clone original' LPE. I have attached a file that shows how you can use it. Interestingly, this also allows for 'paths' where the stroke is only applied outside the path data, by putting the filled original-pathdata-clone on top :)

Cheers, Johan

Nice addition! But it seems a little bit awkward to have to create a new path that then never gets used. It seems that it could be done by adding a "Fill path" effect that just fills the original path. If the effect comes first in the Effect list then it would be painted under the result of the other effects. If it comes last in the list it would be painted on top (giving you the stroke applied to outside of path result).

Tav

---- Defiant00 <defiant00@...400...> wrote:

Tavmjong Bah wrote

...

Nice addition! But it seems a little bit awkward to have to create a new path that then never gets used. It seems that it could be done by adding a "Fill path" effect that just fills the original path. If the effect comes first in the Effect list then it would be painted under the result of the other effects. If it comes last in the list it would be painted on top (giving you the stroke applied to outside of path result).

		Tav

My understanding is that having a single stroke and fill for the path is what would cause this not to work.

Ideally, the 'fill' object would have no stroke and use the path's fill. The powerstroke object would also have no stroke, but use the path's stroke for its fill. From my understanding both objects would use the same stroke and fill, so there would be no way to specify a fill that didn't also then affect the powerstroke.

To answer both of you: an LPE can output only a single path. It is a mathematical procedure on path data, for one path only. Because the desired result cannot be represented by a single path, it is impossible to get that result by applying any kind of LPE on a single path. You have to create multiple paths to create the desired result. Indeed it is awkward to create a dummy path. This is because we have no LPE-generator objects in place that effectively generate path data out of nothing (so without needing path data to begin with). I could create a menu item under "Clone" that makes an 'original path data clone', that would create a dummy path, apply the Clone original LPE and link it to the selected path.

Ciao, Johan

On 13-1-2012 17:01, Defiant00 wrote:

Johan Engelen-4 wrote

...

To answer both of you: an LPE can output only a single path. It is a mathematical procedure on path data, for one path only. Because the desired result cannot be represented by a single path, it is impossible to get that result by applying any kind of LPE on a single path. You have to create multiple paths to create the desired result. Indeed it is awkward to create a dummy path. This is because we have no LPE-generator objects in place that effectively generate path data out of nothing (so without needing path data to begin with). I could create a menu item under "Clone" that makes an 'original path data clone', that would create a dummy path, apply the Clone original LPE and link it to the selected path.

Ciao, Johan

Yeah that's what I thought. A menu item would certainly be useful, both for productivity reasons and exposing the feature to new users.

I've added a menu item for it. Edit > Clone > Clone original path (LPE). I would be very happy if someone can invent a better UI name and description for it (also on the release notes wiki page).

Thanks, Johan

On 17-1-2012 0:22, Defiant00 wrote:

Alright, just got to try out the latest nightly build and the new LPE is excellent, works exactly like I'd hoped.

If I could make a possible UI suggestion, it might be nice to have this menu structure added to the right-click context menu for paths:

Variable Widths -> Create Variable Width Stroke Create Filled Variable Width Stroke

With the first setting up the Powerstroke LPE and the second doing both that and creating the 2nd fill path with the Clone original path LPE.

I refrain from commenting on this. I hope someone with a better feel for our UI can take over here. It is easy to implement, but I won't do it until I get green light.

This way the new features would both get more attention and be significantly quicker to get to. Also, if possible it would probably be better for "Create Filled Variable Width Stroke" to create the fill underneath the Powerstroke, since that would likely be the primary use case.

With that said, even without any further improvements this is already a great usability enhancement for me, and any further tweaks would just be icing on the cake.

Thanks :)

-Johan

Hi, Thanks for your feedback! The response may be a bit slow, so be patient :)

On 24-1-2012 14:57, Defiant00 wrote:

Hello again, here's some further feedback:

1. A way to directly set width values for Powerstroke points (or at least a

command that sets the value to zero without having to directly edit the XML) would be useful.

Yep I know. Thinking about adding an input field to the parameter list in the LPE dialog. Or a popup input field when you double click on the node...

2. I noticed the 'Triangle in' and 'Triangle out' shapes now use

Powerstroke. Is there any chance we could get an extra shape option added to that drop down that produces both a Powerstroke and the clone original LPE at the same time.

Yes :-)

3. Adding and removing Powerstroke control points should not require

switching to another control and back to have them show up properly. 4) The 'Smoothness' control cannot be dragged (at least on Windows 7). Instead, it changes values by 0.01-0.03 (appears to be random) and then loses focus.

Both are known issues, but low on the priority list at the moment... 4) bugs other LPEs too. 3) may be easily fixed, but I do not know how yet. (Inkscape should do something like the switch back-and-forth automatically for you).

Ciao, Johan