Make it crossplatform and you have a BIG win.
QT is good.
Standing on the shoulders of giants certainly makes it easier to see.
Part of the reason Krita development has accelerated massively in the last couple of years is that its team decided on a clear mission statement at some point, one very different from GIMP: Krita is a painting app first. The main use cases are painting, illustration and texture-making. That it's also quite useful for cases GIMP traditionally aims at (photo editing, web graphics) are down to toolset overlap between the use cases and a robust, modern tech foundation (colorspace-independent implementation strategies, etc.), but Krita isn't actually trying to compete with GIMP and ventures into territory GIMP doesn't really serve.
One difference might be that Krita existed as a project to write a raster graphics editor for years before deciding to double-down on painting, and did quite a bit of broad foundational work in that space, like aiming to operate colorspace-independently and implementing various general raster editing tools (e.g. a sophisticated layer and adjustment layer system). Krita retains those traits today, which I feel makes it broader in actual scope than MyPaint, which I think directly started out as a personal project to write a stylus-driven painting app and then grew out from there.
Less of the glacial Debian-style politics than have encumbered GIMP never also helped Krita.
No adjustment layers, and with that no non-destructive editing. No actions that somebody who doesn't know a scripting language can write. (Guess what! I CAN write Python but sometimes I just want to say "oh hey I need to do this 200 more times, I'll just hit record, click the buttons I want, stop recording, and make that action Cmd-F2"). Annoying save dialog. No "Just copy the damn Photoshop keyboard shortcuts" option out of the box. Deafness on the part of the GIMP community that their product is still horrible. The development effort that could have gone in to making a decent photo editor went in to people trying to salvage the disaster that is Gimp. Mediocre Mac support.
I love open source, and use Ubuntu primarily (the Mac references above are due to using a Mac, and therefore Gimp on a Mac from time to time, at work). The user experience is actually BETTER running Photoshop on Windows 7 inside Virtualbox than it is running Gimp. That's just pathetic.
I don't like GIMP specifically because of its user interface.
Instead of following the image editor standard of having a selection edit tool and an "edit the pixels selected tool," you have separate tools for scaling a region, moving a region, and so on. A fairly intuitive operation in Photoshop or Paint.NET (drag the corners of a bounding box, switch to the other mode, drag the corners to scale or move the box to move, etc) requires switching between several tools in a way that I have never been able to understand.
Suffers from what I call "emacs syndrome" -- wasting valuable hotkeys on things that no one ever uses. See E, one of the easiest keys to hit with your left hand, which is bound to... select ellipse. Clearly we select ellipses more than we erase, which is bound to Shift-E...
I've used it for a few personal projects out of curiosity, and I still don't understand how you're supposed to do basic things like make a color palette with it. Googling "how to do X with GIMP" always makes me laugh: How do I place two images side by side? "Make a big image, paste the two images into different layers, position them yourself, and crop" is the only answer. Clearly that's what people were looking for.
I don't understand how it's possible to create such a stupid program. I've barely used it, but every time I do, I encounter new sources of frustration. I have literally never had a positive moment with this program. It's like the developers have never actually used it to draw something or edit a photo.
Open source devs sometimes get flack for copying popular closed source programs. I wish the GIMP devs had just blatantly copied Photoshop every step of the way. They're just wasting their time now. No one cares about technical improvements to an editor that you can't even use.
That seems like an unfortunately common trend. If you ever try any of the hex editors on Linux (ghex, Bless, etc), it's very evident that they were written by people who have no experience actually using hex editors. It's as if they saw a missing tool on Linux, and decided to make it, despite having no personal experience or desire for said tool. It's just a checkbox to tick off.
Screenshots don't look promising, but it's all about the available settings. Ghex and Bless would base column count on the window size, and wouldn't remember your window geometry between runs. So every single run I had to resize the window to hold 16 columns so that my addresses on the left were predictable.
Ghex would re-display its 800x480 number converter every time it was opened, with no way to leave it off for good. I've used those things exactly zero times in the 15 years I've been reverse engineering things. I don't know why every hex editor loves them so much. How often are people really editing files full of nothing but IEEE754 float values?
Both lacked side-by-side editing, which thankfully it looks like wxHexEditor has. As long as it has a fixed-column count setting and strong "compare files" features, I think this will do nicely, thanks.
There's no good model to the UI (and the problem isn't even a difficult one) because it's trying to straddle one person's personal ideas with a common and clear mental model of how certain actions should be grouped.
--Working with transparent layers and channels (compared with PS) is clunky.
--Too much duplication of buttons and window tabs; there are like four ways of exiting the program from the same place (window toolbar, tool pallette toolbar, WM exit, right-click tool panel, etc.)
--Weird script/filter difference (why are there both? why do some destroy undo history and others not?)
--No great selection tools (color matching, fuzzy matching, etc. compared to PS).
--Really really obnoxious "Save as..." which directs you to export for using non-xcf files (and it'll even chide you about that, when it's obvious what you want to do).
--Thousands of other miscellaneous gripes and grumbles.
Anyway the new pedantic approach to saving/exporting seems to be pretty unfriendly - like you say it chides you, you go to save and GIMP says "n-uh, you can only save in XCF; what you want is to 'export' to a new format". Argh!
Whilst it's correct that you're exporting, there's no need for the duplication between save/export other than to indicate to the user that some formats will lose information.
But, to have that dialog box somehow actually say "Aha, we knew you were looking for those other formats, go here instead!" pretty much shows that the developers know what the user wants to do and are maliciously putting a roadblock in their way. It's smug and overbearing and quite annoying.
There's no visual difference between "nothing selected" and "everything selected" - both show a marquee border. There's not even akeyboard shortcut to drop a selection (Select -> None) afaict.
When you paste something, it appears in a separate "layer" - even if you're in quick mask mode. Want to move that thing you just pasted? Nope! You have to select the layer move tool first. Oh ho! But now your focus is in the wrong window. Not that you can see that.
It's unutterably painful.
The complaints about lack of features (adjustment layers in PhotoShop are great, and I really miss the Extract filter) are valid, but I used PhotoShop 2.5 and even though the UI was too complicated even back then, it was at least one order of magnitude easier to use than GIMP is 20(!) years later.
Apple + shift + A
Also, it suffers from being "not Photoshop" -- to the extent that there've been various(?) projects in changing the UI to mimic Photoshop (eg: gimpshop).
Personally I've grown more accustomed to Gimp (and the UI has improved) -- but I still find Photoshop to be a better experience. But the difference is more subtle than it used to be.
And that says a lot. Photoshop has a pretty clumsy UI. But it's serviceable and it's well known, so it's passable. GIMP, somehow, manages to be even worse. That's the downside of having a graphics tool designed by systems engineers.
Photoshop is the standard in the arts industry. Photoshop is also darn good at what it does, certainly no competitors even in the proprietary space. There were at one time: Corel Photopaint , and Paint Shop Pro and so on. Photoshop won that war for a reason.
GIMP needs to be a FOSS Photoshop first. And if there are improvements they can include them. Like dumping GTK and using Qt, for instance.
That's a very uncertain territory there. What sort of legal action could Adobe start if gimp started mimmicking their flagship product UI? And who would pay the legal fees to support Gimp? Would you participate?
> Like dumping GTK and using Qt, for instance.
You're saying that as if it were just a finger snap away. GTK and QT are quite different in spirit and implementation. Besides, before becoming the Gnome Toolkit, GTK was really the Gimp Toolkit. It seems very unlikely that they would dump it in favor of QT, which is used by a major competiting desktop environment.
Unlikely they'd start any: https://en.wikipedia.org/wiki/Lotus_Dev._Corp._v._Borland_In....
That said, Krita looks pretty good. I'm mostly using Gimp for drawing and I'd be interested in trying an alternative.
Leonardo is not finished yet, but it is possible to sign up on out beta list.
The trick Leonardo use to be this fast is that your screen only has a fixed amount of pixels and it is enough to update does pixels super fast for the application to feel fast. Then you have a lot more time to update the rest of the pixels.
This makes Leonardo kind of "resolution independent" and is always super fast whatever the canvas size and brush diameter. On top of this we also have a streaming mechanism of tiles between RAM and file so that you never run out of memory.
That aside, it looks like an awesome product, good work so far. I like the focus on super fast and smooth operation; that's why I use Sublime Text :)
Leonardo use as little OS specific stuff as possible to make porting easy so going from Windows to Mac (and maybe Linux in the future) shouldn't be that hard.
Thanks. Sublime Text has been one of the many inspirations for Leonardo! :)
You keep using that word. I do not think it means what you think it means.
But if I set the background color to 0% opacity (aka transparent) the drawing becomes ultra-slow like 1 pixel per second :/
And PS also has a huge set of commercially created plugins, filters, etc.
It's a pity, though. I really wish some FOSS project would make an end to PS!
Then start using them if you aren't already. A community doesn't come out of thin air.
The strength of Photoshop is in its automation, ability to integrate with all kinds of things, ability to open a hundred different formats by default, plugins, 16-bit-per-channel support, color management, work with huge files, etc etc etc.
This looks like a competitor for ArtRage or Corel Paint.
You're entirely correct about Krita's focus on painting (a focus it has greatly benefited from; see also my other comment), but it turns out the kinds of foundations required by a good painting application offer utility in many use cases.
> The strength of Photoshop is in its automation, ability to integrate with all kinds of things, ability to open a hundred different formats by default, plugins, ....
This, as far as I can see, sums up Gimp very well. With its open scripting and plugin system (Python) which not only allows anyone to write simple automation scripts, but allows people like me to find importers, exporters, filetype support, scripts and plugins for free, for about anyhting thinkable.
How does Photoshop compare to the Gimp in these areas? I am aware of the differences in usability, obviously. But purely on technical grounds: is the automatability and extensability of the Gimp comparable to that of Photoshop?