What would /g/ look for in a new open source image editor that a group of friends and I (might) develop?
What features and methods of doing things are programs such as GIMP, Photoshop, Inkscape, or Paint.NET lacking?
It's ridiculous that Photoshop is the industry standard for image editing when it's not just proprietary -- you have subscribe to it in order to run it on your own computer!
Currently, in all well-known image editors, each tool is its own thing with its own settings, and each tool is used individually, one after the other. I think that an image editing program would be more powerful if you had smaller, more abstract tools that combine into larger tools. For example, instead of using a brush tool, you would use a tool that lets you drag your mouse around to make a line (with no width), and then use another tool on the selection to turn it into a line with a specific thickness, and then use the fill tool, erase tool, or whatever. It would run together, in a way, like a programming language. I guess it would work on a hybrid of raster and vector images.
Anyways, what functionalities would you want to see in such an editor to make image editing more powerful?
gimpscripts.org top downloads might give you some ideas.
>>61178202
An intuitive interface.
OBJECT BASED EDITING
Forget about layers or any exclusively layer based approach, like Photoshop or GIMP has. It's unintuitive and harkens back to a dead medium that is film photography. A digital layer should be nothing more than a group of objects.
Properly combine both vector and bitmap workflows. A single underdeveloped or non existent toolset forces users to shift between applications that often are not even compatible. This is an issue with all popular programs these days. Take inspiration from Fireworks in this regard.
High bit depth support. 8bits per channel is erroneous. Gradients turn out bad, other vector applications out there don't even support it. This could be a competitive feature there.
Also remember to develop it in performance oriented languages, like C or at least C++.
>>61178202
> For example, instead of using a brush tool, you would use a tool that lets you drag your mouse around to make a line (with no width), and then use another tool on the selection to turn it into a line with a specific thickness, and then use the fill tool, erase tool, or whatever.
All your mentioned programs already do that via path creation/vector editing, but I get what you're trying to say.
It'd be much easier to use a program which didn't clutter everything up with ten options for every tool, instead relying on a mixture of those tools to achieve the same sort of effect, but it means you'd employ a much different workflow than the standard. Your program would not only have to be innovative but also teach people how to use it more so than other image editing programs.
Take a look at Paint star, a somewhat halfway approach between the basic Microsoft Paint and Abode Photoshop. It might give you a better perspective.
The original is by wangzhenzhou hosted on a google site.
>>61178595
second that, never though about it before but it makes sense.
Although in the place of layers I think you would need some kind of grouping directory, groups and subgroups, and tools for easily moving objects from one group to another.
You could treat every tool as a plug-in, chaining them together like you say. Then have a toolbox which the user can configure to imitate whatever software they used before.
I'm not sure if i represent your end user, but I prefer my software lightweight but expendable. Fast core application which the user can load plug-ins onto to suit their needs.
I assume it will be proprietary, but make it easy for a community to form by making customizable, allow skins, plug-ins etc.
+ on linux, windows and macOS. cross-compatibility could be a feature. And after all 2017 is the year of the linux desktop
>>61178775
I should've been more specific. I meant typical raster monsters. Ones that need a freaking set of 6 tools to do any modification and a butt load of filters for further editing. This leads to unintuitive experience, slow workflows and obscure editing. GIMP is really guilty of this. Photoshop has some features to "objectify" layers but it can't escape the basic nature of the beast.
Vectors however fundamentally give most of the editing potential to the cursor rather than the tools so the experience there is much better.
>>61178202
GIMP is powerful as fuck. It beats Photoshop in some ways. I doubt you can make something as good as gimp/photoshop.
>>61178913
Not OP but there could be better applications. However I agree that it takes quite some a lot of skill to develop a big application.