USER FORUM
(you are viewing a thread; or go back to list of threads)
Experimental SolveSpace Web Edition (by whitequark)
I've ported SolveSpace to the web platform.
Try this: https://files.whitequark.org/SolveSpace/solvespace.html
It doesn't work very well yet, as it's just an experiment. Many features are missing. Still, it's usable as a demo.
Try this: https://files.whitequark.org/SolveSpace/solvespace.html
It doesn't work very well yet, as it's just an experiment. Many features are missing. Still, it's usable as a demo.
Thanks (by fxzaragoza)
it works perfectly in Core i7 Geforce G7 520MX WIN10 Firefox 61
Thank you
fxzaragoza
www.fxzgz.tk
Thank you
fxzaragoza
www.fxzgz.tk
(no subject) (by EvilSpirit)
Works good for me!
(no subject) (by Bep van Malde)
I cant save, open,import nor export.
Drawing etc seems to work.
Drawing etc seems to work.
(no subject) (by Jonathan Westhues)
That's pretty cool. To my knowledge, that makes SolveSpace the most powerful CAD program running entirely within the client browser (since OnShape etc. do lots of stuff on the server). I'm not sure what that's good for yet, but it must be something...
(no subject) (by Eric Buijs)
For an experiment it looks very promising. Are there any plans to develop it further?
(no subject) (by Azial)
Super awesome! You use Web Assembly, right?
It crashes on Safari shortly after a few seconds after loading:
Error: Out of bounds Trunc operation (evaluating 'Module["asm"]["dynCall_vi"].apply(null, arguments)')
Works fine on Chrome.
It crashes on Safari shortly after a few seconds after loading:
Error: Out of bounds Trunc operation (evaluating 'Module["asm"]["dynCall_vi"].apply(null, arguments)')
Works fine on Chrome.
(no subject) (by Azial)
Its also super fast at rendering and makes correct use of high DPI (testing on 5k iMac). Both is not the case for the Desktop client. Maybe worth to make an Electron Application out of it?
Crashed (by Vikenti)
Crashed in Maxthon.
It will be nice to have a stand-alone version (and more usable for me), too.
It will be nice to have a stand-alone version (and more usable for me), too.
Android version (by Vikenti)
I agree with Azial - stand-alone version could be an application for Android platfom (will be very usable on android tablet). Windows or Android app is more convenient way to use than internet solution.
(no subject) (by EvilSpirit)
Anyone can give us some advices how we should rework UI and controls for getting useable touchscreen application? The main problem is not to make binaries for Android, the main problem is what we should completely change UI.
(no subject) (by Azial)
Hm, maybe something like that:
For the controls:
Map pinch gesture to zoom (mouse wheel).
Map 2 fingers pan to right-click drag.
Map 3 fingers to rotate view.
Alternatively, introduce a hovering cube or two (for drag / rotate) like in inventor as the representation for the view. The user can than tap and drag on the cube and based on the delta position from the tab, the view is panned / rotated.
For the operations:
Map 1 finger to normal left click / drag.
The user can just tap the UI button like you would with the mouse when not using the shortcuts. Of course the "hidden" constraints (M, Z, ...) need an button then as well.
For the menu:
For a mobile application I would rather try to avoid a menu at all and introduce more buttons for the operations. But in general:
Make the typical "..." button. When clicked/taped, a menu page spawns.
For the controls:
Map pinch gesture to zoom (mouse wheel).
Map 2 fingers pan to right-click drag.
Map 3 fingers to rotate view.
Alternatively, introduce a hovering cube or two (for drag / rotate) like in inventor as the representation for the view. The user can than tap and drag on the cube and based on the delta position from the tab, the view is panned / rotated.
For the operations:
Map 1 finger to normal left click / drag.
The user can just tap the UI button like you would with the mouse when not using the shortcuts. Of course the "hidden" constraints (M, Z, ...) need an button then as well.
For the menu:
For a mobile application I would rather try to avoid a menu at all and introduce more buttons for the operations. But in general:
Make the typical "..." button. When clicked/taped, a menu page spawns.
Mobile UI (by Vikenti)
I think we can use ideas from such applications like Caustic (music-creation tool for Android\PC) and from MediaBang Paint (Manga tool for Android\PC).
Two main concepts:
- switchable views (Caustic). The program has a little button in the corner which allows to switch between different views (instruments, effects, some complicated properties, etc). All menus organized as a fixed grid.
- little buttons on the side of the current view. Each button function depends on current view (contextual).
The workflow in these programs is very fluent and simple.
For a SolveSpace it could be a plane screen with few buttons on one side. The first button always shows us view-switching menu - properties, constrains, layers etc.
After switching view - button functions switches and we can work with them. We don't need too many buttons in one view, but we can have multiple views.
Another idea is: when we select button (for example, linear dimension) it toggles and we can define multiple dimensions until we toggle this button again.
Useful thing is snapping. But i prefer behavior like in audio daw - you can press "quantize" and it snaps all points on the specified rules (for example, span to the grid).
Two main concepts:
- switchable views (Caustic). The program has a little button in the corner which allows to switch between different views (instruments, effects, some complicated properties, etc). All menus organized as a fixed grid.
- little buttons on the side of the current view. Each button function depends on current view (contextual).
The workflow in these programs is very fluent and simple.
For a SolveSpace it could be a plane screen with few buttons on one side. The first button always shows us view-switching menu - properties, constrains, layers etc.
After switching view - button functions switches and we can work with them. We don't need too many buttons in one view, but we can have multiple views.
Another idea is: when we select button (for example, linear dimension) it toggles and we can define multiple dimensions until we toggle this button again.
Useful thing is snapping. But i prefer behavior like in audio daw - you can press "quantize" and it snaps all points on the specified rules (for example, span to the grid).
Addition (by Vikenti)
Zoom\scrolling\turning can be simple implemented with few buttons that toggles specified action.
It may be useful because when you create cad model you don't need all of movements simultaneousely. Common practice is to draw in 2D and then move\rotate in 3D. So we don't need all zoom\scroll\rotate instruments simultaneousely! That is why we can use switch buttons to select current action and don't use 2\3 fingers - that could be too complicated when you work fast.
It could be convenient to have 2 or 3 levels of zoom which you can define and switch fast (whole view and some detailed views).
If you want, i can draw my ideas of solvespace mobile GUI for better understanding (each view and links between views). Feel free to email me.
It may be useful because when you create cad model you don't need all of movements simultaneousely. Common practice is to draw in 2D and then move\rotate in 3D. So we don't need all zoom\scroll\rotate instruments simultaneousely! That is why we can use switch buttons to select current action and don't use 2\3 fingers - that could be too complicated when you work fast.
It could be convenient to have 2 or 3 levels of zoom which you can define and switch fast (whole view and some detailed views).
If you want, i can draw my ideas of solvespace mobile GUI for better understanding (each view and links between views). Feel free to email me.
(no subject) (by James)
Is there any further update on this? This seems to be quite promising. Since cloud-based service is more and more popular, this maybe very useful.
(no subject) (by Paul)
There has not been any progress on it in 2 years. Whitequark has resigned from Solvespace development.
From a developer point of view it might simplify the ability to run on different OSes.
From a users point of view I don't think it makes sense. People are getting very frustrated with cloud-based software and SaaS pricing, particularly in the CAD space. "Owning" your own copy of software and storing your work locally are very important features.
We can accept changes to the web version. A few developers think it's interesting, but nobody seems to want to work on it.
From a developer point of view it might simplify the ability to run on different OSes.
From a users point of view I don't think it makes sense. People are getting very frustrated with cloud-based software and SaaS pricing, particularly in the CAD space. "Owning" your own copy of software and storing your work locally are very important features.
We can accept changes to the web version. A few developers think it's interesting, but nobody seems to want to work on it.
Interesting (by Thomas Herrmann)
This is prety neat. No Platform problems and installations required. FIles could be saved nicely on an nextcloud.
But I wouldn´t see this as priority. I prefer local programs too.
But I wouldn´t see this as priority. I prefer local programs too.
(no subject) (by Dave)
Mind blowing! Nice one.
Agree with Thomas though that personally I hope the local version gets plenty of attention. :-)
Agree with Thomas though that personally I hope the local version gets plenty of attention. :-)
looking for a developer (by R, Demirbas)
Hi,
I am looking for a developer/partner for my startup who is able to integrate Solve space 2D drafting tools to the company web site. for more info Search what cut on linkedin.
Thanks
I am looking for a developer/partner for my startup who is able to integrate Solve space 2D drafting tools to the company web site. for more info Search what cut on linkedin.
Thanks
(no subject) (by Paul)
@Demirbas You can email Jonathan directly to discuss commercial licensing of the constraint solver. To get going faster you might ask customers to submit designs as DXF, SVG, or whatever and recommend using existing free tools to create them: Solvespace, LibreCAD, etc...
How to build for web (by Trevor D'Arcy-Evans)
Are there any instructions on how to build the web version of SolveSpace?
(no subject) (by Paul)
@Trevor no instructions that I'm aware of. It was very experimental and incomplete (no file I/O) but you could draw and extrude.
(no subject) (by ruevs)
Here are the instructions https://github.com/solvespace/...ee/emscripten#building-for-web
(no subject) (by ruevs)
Ignore flatc it is not needed any more.
(no subject) (by Trevor DArcy-Evans)
@ruevs @Paul
Many thanks for the quick reply. I'm now getting the following error:
[ 18%] Generating pnglibconf.c
options.awk: bad line (10): com
CMake Error at scripts/gensrc.cmake:68 (message):
Failed to generate pnglibconf.tf5
make[2]: *** [extlib/libpng/CMakeFiles/genfiles.dir/build.make:85: extlib/libpng/pnglibconf.c] Error 1
make[1]: *** [CMakeFiles/Makefile2:811: extlib/libpng/CMakeFiles/genfiles.dir/all] Error 2
make: *** [Makefile:152: all] Error 2
Platform: Linux Mint 20.3
Many thanks for the quick reply. I'm now getting the following error:
[ 18%] Generating pnglibconf.c
options.awk: bad line (10): com
CMake Error at scripts/gensrc.cmake:68 (message):
Failed to generate pnglibconf.tf5
make[2]: *** [extlib/libpng/CMakeFiles/genfiles.dir/build.make:85: extlib/libpng/pnglibconf.c] Error 1
make[1]: *** [CMakeFiles/Makefile2:811: extlib/libpng/CMakeFiles/genfiles.dir/all] Error 2
make: *** [Makefile:152: all] Error 2
Platform: Linux Mint 20.3
(no subject) (by ruevs)
Don't know. But this recently came up and is also relatred to libpng and awk.
https://github.com/solvespace/solvespace/issues/1228
https://github.com/solvespace/solvespace/issues/1228
(no subject) (by ruevs)
And more updates on the Web version https://github.com/solvespace/solvespace/pull/1275
And the previous work from 2020 for reference:
https://github.com/vespakoen/solvespace/tree/emscriptenlib
https://github.com/vespakoen/solvespace/tree/vespakoen-emscripten
And the previous work from 2020 for reference:
https://github.com/vespakoen/solvespace/tree/emscriptenlib
https://github.com/vespakoen/solvespace/tree/vespakoen-emscripten
(no subject) (by ruevs)
Here is the actual current Web edition from the above pull request to try:
https://verylowfreq.github.io/...ace-on-browser/solvespace.html
https://verylowfreq.github.io/experimental-solvespace-on-browser/
https://verylowfreq.github.io/...ace-on-browser/solvespace.html
https://verylowfreq.github.io/experimental-solvespace-on-browser/
Post a reply to this comment: