I moved over to Krita from Clip Studio when I decided to go pure-Linux on all of my computers, not because I thought the software was particularly competitive or good, but because I thought it was the most likely candidate to become competitive and good.
Krita still has a lot of rough edges, but I see it on a similar trajectory as Blender. It's very rapidly improving, not just in the sense of adding new features, but in becoming a more stable, reliable product.
I was using Blender a lot earlier in its development, and I remembered the comparisons people would make to Maya, and it's just... honestly, it's a little surreal to see it now well on its way to becoming something of an industry standard. Blender used to be really annoying to use, and only niche hobbyists were using it. All things considered, its rise in popularity/quality happened fast.
So similarly, I'm very bullish on Krita, and it's very encouraging to see members of that community looking at Blender and taking lessons from its development. Krita is one of the more promising and encouraging OS software projects that I follow.
Most of the migrations I've seen have been from PS/SAI to CSP. There are a ton of artists creating custom brushes for CSP now, and you can't import them into Krita. If you're in a particular artist community/discord and want help, no one is going to know what you're talking about. That's classic lock-in.
CSP and Procreate are the hot tools on the Twitter/Instagram scene right now and I don't see the growth slowing. Krita ATM is unfortunately perceived as an outlier tool, like Medibang or FireAlpaca.
It probably isn't an issue to you personally, but those are the factors I see stymying adoption (and to be honest, I have literally never met another artist using Krita)
That being said (and keep in mind I am only one fallible data-point) the community situation is roughly equivalent to what I remember Blender being like when I first started using it (maybe in 2007-2008).
- Very few professionals were using Blender, if any.
- Most tools were being written for Maya. Every time someone would link a cool tech product with crowd simulations or plant growth, it would always be a Maya plugin.
- Any college/high-school modeling/animation courses you took would be using Maya.
- That meant that if someone wanted to learn professional animation/modeling, they'd learn Maya's paradigms, and then Blender would seem extra weird. This was back when you just had to get used to the weird 3d cursor.
- Compatibility between the two programs was awful. I remember wrestling with the options to get 3D models to export in a format that wouldn't be completely messed up when imported in Unity. I don't remember if I ever got it working.
There were people who used Blender because it was free and swore that it was just as powerful once you got used to it, but they were largely techy people. And even if they were right, none of the digital artists or studios who's work I really respected cared at all. If you wanted to get an actual job with animation, you used Maya, period. Only hobbyists could afford to spend the time learning a separate program.
So it doesn't bother me too much to see Krita in the same position. I care more about the trajectory/velocity of development, and who I see using the software. I see a lot of programmers using software like Gimp, but I see a larger focus from Krita on actual artists, and a more pragmatic prioritization of features artists use.
If the core community stays really friendly for artists, and they keep on releasing at their current pace, then I suspect adoption will eventually come. Or at least, I think it's a decent enough bet that I'm willing to use the software in its current state and regularly throw money at its development.
Considering the tools required for digital art (a Wacom, iPad Pro, or even a cheaper Huion etc) the art program is probably the cheapest purchase you're going to make, and they're all perpetual licenses (sans Adobe CC)
I got Affinity Designer just as soon as I found out about it after many failed attempts to understand Inkscape. There's just no comparison. Inkscape needs to have that moment of focus that got Blender over its UX wall.
I'm sure they're in better financial shape four years later, but that's a lot of money to invest without assurance at least 10,000 people ($500,000/$50 sale price) will buy it.
You just convinced me to try out Krita again. I've had a lot of problems getting away from Clip Studio, each time it felt the options weren't ripe enough yet. I remember trying Krita a very long time ago and it was barebones, but sounds like it's had a lot of good development time.
whether the app is "free" enough for their liking (MIT vs GPL-based license, etc),
whether it picked the "right" frontend in the never-ending GTK vs QT battle,
whether it is written in an "acceptable" language (Electron being Satan Incarnate, of course),
whether it has been properly packaged as a DEB/RPM/Snap/Flatpak/AUR/AppImage, etc.
Fail any one of these things, and the app is instantly "unusable" and we should all use ncurses-based obscure-thing-I-found-over-the-weekend-in-some-dude's-PPA.
Meanwhile, an increasing array of "free as in freedom" apps that are widely available and cross-platform (e.g. Blender, VLC, Audacity, Calibre, etc.) serve the larger community by just letting people get shit done, and build up name-recognition as a result.
I know it is tangential to this point, but a lot of developers, including long time Free Software developers, don't get licensing at all. While packaging software (FOSS) for a distribution, I've had to poke multiple people to fix licensing problems that invariably occurred (most common: incompatible licensing of parts of a project, perhaps imported from somewhere else).
> whether it is written in an "acceptable" language (Electron being Satan Incarnate, of course)
There are legitimate technical reasons for not liking Electron, not necessarily due to "disproportionate amount of time bickering" spent by "Linux Desktop users".
Completely agree; it's a complex part of producing software that people (myself being guilty of this) incorrectly assume they can ignore.
> There are legitimate technical reasons for not liking Electron, not necessarily due to "disproportionate amount of time bickering" spent by "Linux Desktop users".
There are legitimate technical reasons for not choosing electron, but I can't think of any for not liking it. It's a perfectly fine and useful way to make a cross-platform desktop application. If it doesn't suit your project's purposes because of performance, size, or current acumen, that's fine.
But I think OP is referring to "Electron is awful", "I hate Electron", or "Never use Electron" levels of ire that crop up occasionally, which should be exclusively reserved for something that no one should be working in.
My main problems with Electron are:
- The size of the runtime
- The fact that every application has to bundle it and it means that multiple Electron applications on the same system do consume a lot of resources if not handled well
- The fact that unless you (as a developer) play due diligence, you might distribute your app with a version of the runtime which may contain vulnerabilities (one of the reasons Linux distributions don't like library bundling that much).
I can't deny it's probably easier for many (not for me, but although I write Python code all day, I'm by no means a "developer"), but it can be potentially wasteful, to say the least, and require far more resources than what you'd actually need with a another toolkit.
That said I for sure won't point a gun at someone who wants to use it. ;)
I distinctly remember a post on here about three or four years ago where a blind person detailed losing their job because Electron does not integrate with desktop accessibility functions.
Not sure if you're referring to a choice between packaging formats, or not packaging at all. If the latter, well, it's pretty tiring looking through a list of filenames which contains the platform and arch, downloading the correct one, then going to the project's wik and copy/pasting commands to create users, create systemd files etc. (Looking at you, Go devs). Oh, you want to upgrade it along with the rest of your software on your system? Fuck you.
I get it, packaging isn't painless. But if you've worked hard writing a piece of software, it seems bizarre not to spend some time on packaging and getting distros to accept it.
We like packaging, because it's awesome. A large collection of extremely trustworthy software, all a single, simple command away.
My personal experience is that it means, at the minimum, having a VM for that distro hot and ready to go, reading a bunch of documentation about how that distro does packaging, consulting an oracle to tell you which versions of that distro you should target, learning how this particular distro wraps your particular build system, translating all the dependencies, etc.
So it doesn’t seem like just “some time” to me. If you pass this off to a maintainer for that distro, they’re usually fairly experienced with that distro’s ecosystems and the tools.
"Blender’s development fund currently brings in about 1,200,000 euros a year, which funds 20 full-time developers. That’s not the only source of funding. Blender has about 172 developers in the past year, and 550 over its entire existence, and 64 in the past month, same as LibreOffice. Looking at the last number, it means that there are anyway more volunteer committers in the Blender community than paid developers. Funded development hasn’t eaten the community.
Let’s hazard a guess: Blender has four times the installed base of AutoDesk Maya. This is pretty rough, of course, so ingest with salt.
These are important non-software aspects of Blender which are critical for acceptance by new users, which is critical to maintain a healthy user base, which should in turn yield a well-community-supported codebase...
In other words, this makes Blender a safe choice in corporate environments where a choice may exist between it and commercial offerings, such as Autodesk's Maya...
Wow not paying them a lot are they.
If you go for Eastern Europe, Hungary/Poland/Ukraine, that's enough to afford the best people.
Even 3000€ or even 2500€ take home pay per month would let me live a very good life (for me) in a sunny, safe, beautiful place in Europe. All the while I'm working for a free software project, one which I'm a fan of (not working to make my boss rich, not inventing better ways to make people click on ads). What else could I possibly ask for.
I think early FOSS programmers didn’t value marketing enough, which produced excellent software with difficult adoption. Now, for any software to succeed, a nice UI and good marketing material is the baseline. FOSS that is born since 10 years was much nicer (from Angular to Kibana).
It is even visible in security breach: At one point after Heartbleed, there was a discussion on whether security bugs would now require a nice single-use website with graphic artifacts, in order to get heard by developers.
What I like in Blender, except its invaluable community, is its UX, they follow a good logic in interfaces. If someone is familiar with 3Ds Max or Maya, in less than an hour can understand the whole software and how to work with it even without any help or tutorial
However, I'm not in the VFX industry and I'm an architect and Building Information Modeling/Management - BIM technologist, so recently we've started to improve Blender for Digital Built Environment industry, an open-source plugin (interfaces) called BlenderBIM 
So hope we help Blender and Blender community to speed up its success
Btw did you know that the icons in blenderbim.org homepage are icons from the Breeze icons theme developed by the KDE community (the community behind Krita)?
The gimp fork Glimpse is planning an MSI installer, so that'll be installed soon
Is it really easier to install compared to an exe installer (say nsis / inno-based)? How so?
Much easier to add an MSI to GPO, and then just as easy to remove it again and it's uninstalled.
Not sure how writing scripts to do all that myself for every piece of software is better.
MSIX is just a wrapper around an msi, exe, etc installer anyway.
In conclusion, just make an MSI file, they're much better, look at all these great features https://en.wikipedia.org/wiki/Windows_Installer
The real budget is likely much higher than the article suggest, especially now that Ubuntu fully switched to Gnome.
Disclaimer: I work for Red Hat but not on Gnome