Thanks for those hints. It makes more sense to me now with a few starters. Nice 2D editor. As well as the other features mentioned, I particularly like the two "handles" controlling the Beziers being independent at the point where they are created. I don't like the other systems where the Bezier point has controls on either side of the point connected like symmetrical dumbbells. Plus the <scroll mouse>+CTRL provides a basic Zoom feature as well. Rob On 09/10/16 02:36, Carsten Arnholm wrote: > On 08. okt. 2016 15:49, Jerry Davis wrote: >> Carsten: Nice work documenting this. >> >> I had no idea about the nice curving options. >> >> You seem to know quite a bit about this tool: >> >> Do you know what the button on the right does (brings up the url and >> file option)? And how to use it? >> >> How do you add points that you want to be in between two other points >> instead of adding the point at the end? > > Hi Jerry, > > Well, I had not seen the tool before, I am just able to guess :-) Now > I guessed what the file option is all about: You are supposed to pick > up an image file to use as background for manual sketching. > > At first I tried an OpenSCAD source file ... I thought perhaps it > could read a polygon as well as write one, but then it complained that > the file I gave it wasn't an image. That gave me a hint it wanted an > image :-) > > So I tried the Ford logo image file (.png) and it worked > https://gyazo.com/094c5662fa9aea39373a0b9b811d9c10 > > The window that says "Paths" allows you to define several individual > paths (a.k.a. polygons) in the same sketch, click [New] to start a new > path. To do the Ford logo example manually it means 4 different > polygons/paths. > > - outer contour of the letter "F" > - inner contour of the small loop in the letter "F" > - outer contour of the letters "ord" > - inner contour of the letter "d" > > So in theory you could sketch the logo by tracing the image manually, > but it gets complicated/tedious and inaccurate. It is much better to > use an automatic tool when possible. For simpler stuff I guess it > could work, or if the image is not directly suitable for automatic > analysis (some are not). > > Apparently, there is no way to use the GUI to add points in between to > other points, or at least I could not figure it out. You can always > insert the extra point by typing in the correct place in the tiny > "OpenSCAD code" window, then the curve will be updated correctly. > However, this to some degree defeats the purpose of the tool, because > you are then back to editing the source directly, just in a much > smaller window than usual.... > > I also tried pasting polygon code from OpenSCAD into the code window, > but that failed when the coordinates were outside the window range, > the polygon builder does not fit the coordinates automatically, and it > does not understand scale() or other transformations. It just not > usable for too large or too small coordinates. It also turns out that > the definition of "too small" and "too large" is unclear, I opened the > polygon builder in Firefox on both Windows7 and Kubuntu 15.10, for > some reason the x-range is [-72,72] on Kubuntu and [-88,88] on Win7, > so when I copied a polygon from polygon builder from Win7 to Kubuntu, > parts of the polygon were hidden under the toolbox and I could not > reach it. > > It is a nice idea, but with clear limitations. > > Carsten Arnholm > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > -- *Rob Ward* Lake Tyers Beach, 3909 Lake Tyers Beach Website <http://www.laketyersbeach.net.au> XP to XUbuntu <http://www.laketyersbeach.net.au/XP2XU.html>

On 09. okt. 2016 02:53, MichaelAtOz wrote: > There is a resize tool on the bottom right of the text box. It's a big > clunky, you need to drag slowly. > You can also drag the toolbox elsewhere if you need to. Ah, yes you are right on both accounts. It makes it a bit more useful. Odd resize feature, but with a bit of care it works. Thanks. Carsten Arnholm

On 06. okt. 2016 18:58, Torsten Paul wrote: > We still need to work out some of the details, but > https://github.com/openscad/openscad/pull/1110 should > be able to handle at least the simpler cases using > the potrace library. I have been playing a bit more with generating 3d models from images or other data sources. For the case of an image, instead of using potrace, the idea was to automatically generate the modeller code to create a 'language version' of a profile. For example, using this image as input http://arnholm.org/tmp/profiles/guy.png OpenSCAD code was generated, the image profile is now a module: http://arnholm.org/tmp/profiles/guy.scad Same thing in AngelScript code, the image profile is now a function http://arnholm.org/tmp/profiles/guy.as Both produce very similar looking results, using quite different internal libraries. Regardless of the preferred modeller, with such an approach one can build a library of profiles expressed directly in the modeller language, not having to rely on external files. I also find it interesting to compare the syntax. In Angelscript, the shapes are real language variables that can be modified. Assembling the final model is a simple linear process where the profile object is modified several times. In OpenSCAD it has to be expressed as a recursion, not quite as easy IMHO. But in both cases the result is the same. Regards Carsten Arnholm