Open-sourcing Refraction image editor


Maybe if the project is sucessful in getting refraction open sourced the name can change back to Inferno since it would no longer be a commerical product?


Since the pledge collected 575 € only, we should consider this project as dead… at least we tried :wink:

But we should think about how to continue from here

  • Ask Frans von Nispen if they are willing to port it to refraction if there are a number of preorders
  • Try to use this money to improoof Artpaint or Becasso
  • Try to use this money to encurage stippi to realease wonderbrush NG
  • ....?


I’ve forgot to post here the last email I’ve received from Frans:

'There is still a demo package on the website which we have put back online a few months ago.

I no longer have an archive of older and other versions available and as we do not have a working installed BeOS system with our tools installed, doing a new build is not possible, or at least will take op quite some time to set it all up. But there is a BeOS 5 version which we released, so maybe you can track back a version of someone who still has it. I can still generate license keys and can help you get the versions running.’

Considering above, I don’t think he will put resources on further development of Refraction. I also agree that this project is not realistically possible ATM (at least till Haiku beta is released and more new users are brought by it). From my POV it would be better to collect the money for the bounty to fix Haiku beta-related bugs which aren’t related to package management (Waddlesplash should close those I suppose).


I agree with a few who have commmented here, a QT 5 or a improved qt 4.8 port would help haiku significantly more then a outdate editing application.


I agree with a few who have commmented here, a QT 5 or a improved qt 4.8 port would help haiku significantly more then a outdate editing application.


i bet it’d be quicker and cheaper to write an entry-level image manipulator from scratch than it would be to update refraction to a point that it makes sense to use over any browser-based or qt app. for anything meant for professional work, there’s no native system anywhere, under any os, that’s a widely used industry standard so we wouldn’t be any different or lesser for not having a native pro app (except, no way are corel or adobe looking at us). there’re lots of good, portable tools we’d do well to avail ourselves of.


and now I don’t want Pixel32 if that’s what it was :slight_smile:


…and what about improving GrafX2 ?


[]…and what about improving GrafX2 ?[/quote]

Heh – it would be fun to have an app in the DeluxePaint family available! 256-colour mode is a bit limiting, but I’d guess its features depend on it(?)

I don’t see any current port to Haiku, though.


It works on Haiku just fine and there is an haikuports package available.

The 256 color limitation is from the days this was a DOS program. It could be removed, but that would be a lot of work, probably.


and would kill the fun!


I am probably very biased. But I just wanted to mention that Refraction existed before I wrote WonderBrush. And ArtPaint was open sourced before that too. I’ve had very good reasons why I still wrote WonderBrush. No other BeOS painting tool (including Refraction) has a brush system with subpixel-precise positioning, and no other tool (maybe except for ePicture) actually treats transparency correctly in color blending calculations. That last bit alone means that every filter in Refraction (or ArtPaint or Becasso) does not work correctly on layers that have any transparency. If you have an old copy of Refraction, just try it out: Fill your image with light-gray. Then make a transparent layer on top of that, draw some white shape in that layer, then apply a Gaussian Blur with big radius on that layer. Ask yourself why the blur darkens the the white shape… it comes from mixing the “black” pixels outside of the white shape when these pixels are actually completely transparent and should have no color at all.

I should really make WonderBrush NG open source. But is there anybody who would actually work on it?


This is indeed an important consideration. I see equivalently “bozo” mistakes in audio code all the time. Sometimes you can see or hear the results of such goofs out in the real world. You can hear the mistakes in the old Haiku system mixer for example.

While I agree it’s right to be cautious about the potential for “code dumps” to just be abandoned unloved, I would always say that it’s better for the code to be out there than not.

Definitely don’t wait until you lose interest altogether, that’s my recommendation. When you experience a brief burst of enthusiasm to work on it, use that enthusiasm to get a working codebase into Github or similar. Worst case, it’s the most thorough backup you’ll ever make in your life.


Anything new on the possibility of open sourcing Refraction (formerly known as Inferno)?


Maybe instead we should ask Frans how many licenses $575 would buy us?


Open sourcing Wonderbrush to a place like HaikuArchives would allow multiple people to have the ability to improve the application. As for people abandoning the app, I’d doubt that, especially since your app is bundled with the release version of Haiku.