I work with many people who use Linux. Most of them are serious anti-capitalists and FOSS hardhats who hate even small software developers, let alone Adobe. I don’t agree with their position but as long as desktop Linux has less than 3% marketshare and half the people using it wouldn’t want to buy software (let alone €300 software), spending any time developing a Linux version of commercial software is time wasted.
Add to that the argument above about the 23 different distributions on four major families (Fedora/CentOS/Redhat, Debian, Ubuntu, Gentoo). Not possible. It’s not clear that it isn’t Microsoft and the NSA conspiring to keep Linux fractured and off the desktop. What is clear is that whatever is making Linux non-viable for commercial software is working.
I didn’t say that all Linux users would pay for this software. Only that I haven’t seen evidence of any market research by the software developers. My reasons for using Linux are non political; for using Windows as well: only to use DxO and Luminar.
I would still like to hear from DxO what research they have conducted.
I’m sorry to learn about your hearing issues, Mike. If you go to the list of commercial software on Linux, you’ll learn the list isn’t long and support for most of it is “discontinued”.
If you are deeply interested in the question of why commercial software is rarely distributed on Linux, you might ask Corel about their sales on Linux as Corel have actually explored the Linux commercial market for graphics software. Corel offered three graphics programs on Linux: AfterShot Pro, CorelDraw and PhotoPaint. Curiously enough, the last two have been discontinued.
Hopefully DxO will first add hardware acceleration for its existing platforms, Mac and Linux, rather than adding a platform for which the native video drivers are very sketchy and not generally suited for hardware acceleration.
Hi,
Today I did a comparison of 5-6 different products on external Windows machine using one underexposed RAW photo and DxO PhotoLab 2 was the best of all. Better than Lightroom. The problem is that I’m a Linux user and can not use it. I can disagree of the popular opinion that Linux users do not pay for Software, we don’t pay because there is no good paid software for our beloved platform. Everybody says there is really small percentage of Linux users just 1.5% but nobody thinks that 1.5% of 1 billion desktop users is 15 millions.
Linux users hate to pay for software and all the other software on the platform (DarkTable, RawTherapee) are free and built as native Linux applications. You can buy an old Mac Silver Tower for a few hundred euros these days, update it for free to 10.11.x with no problems and 10.14 with a bit of research and the right graphics card. Bingo, you have a dedicated multimedia machine with full access to everything that runs on Mac. I’d still avoid Adobe if I were you. It’s almost impossible to get their spyware out of your OS, even a trial.
And how many of those 15 million are interested in expensive high end post-processing software, and are willing to pay for it? The audience on Linux is a small fraction of those people using a Mac, and barely a blip compared to Windows users. The potential number of new customers is probably not worth the significant cost and effort it would take to port a version for them, which at the same time would take resources away from the versions that are profitable. That is also the main reason there is little paid software available on Linux from most large and small software developers. There is simply not enough profit potential.
At this time, there are close to 1.5 billion computers in the world, old and new, that run various versions of Windows. Most of them run Windows 1O or Windows 7. If you were going to invest your time, money and resources developing software for a computer platform, would you choose the one with over a billion potential customers or 15 million potential customers?
I’m not aware of internal detail how the software is build, but on first look DxO PhotoLab is build on .net framework and Lua scripts. With upcoming .net 5 it should provide one framework for all platforms, it is still not clear if it will be possible to run WPF under Linux but if it will be so, why not to invest several dev sprints to make one product that runs on multiple platforms: Windows, Linux, Mac OS, iOS, Android and etc.
Linux is not a single platform. Should DxO offer Ubuntu, Debian, Slackware, CentOS/Redhat, Fedora or all five? Linux sadly is a rabbit hole down which few commercial developers come out alive. Show me three successful Linux commercial graphics or video application deployments and I might change my mind.
DXO should support a distribution which Microsoft .net framework supports, nobody wants to support all 1000 distributions. So lets talk about video and graphics:
Here are 10 products from which only Blender is free, but has plugins that cost thousands. I hope the list is enough. Adobe as one of main big producers for Image/Video editing software, does not support Linux in any its programs, but most of them run just fine under Wine.
As a software developer, I can tell you that there is no such thing as a Linux version. Which version of Linux do you want to run on - Red Hat, Ubuntu, Debian, SUSE… ?
It’s very much the same for Android development, which I won’t touch due to the enormous variety of different screen formats and sub-versions that different devices run. Then there’s the fact that Android users are very reluctant to update their OS, unlike Apple users who are very quick to update to the latest version of iOS.
I did once think of porting one of my iOS apps to Android, took a couple of weeks investigating it and then gave up in horror at the amount of work involved compared with the iOS version.
Similarly, before I moved to Mac and iOS development, I tried to port a Windows app to Linux - once again, after a couple of weeks, it became apparent the effort required and the lack of compatibility between versions would mean an enormous amount of work, and for very little reward, as most Linux users are just not interested in paying for anything. On the other hand, I need to make an income for luxuries like housing, food, electricity, etc.
Of this list, most of it is 3D and rendering software which I deliberately left off my list. Let’s go through what’s left one at a time.
Lightworks
The resurrection of Lightworks is a commercial failure and a weird sort of hobby project which hasn’t really managed to crossover. How do I know? I thought seriously about buying and using it but it’s just too unreliable and cranky on Mac. Very limited audience, unfortunately.
DaVinci Resolve
Amazing product, fantastic that they are on Linux. DaVinci Color Correction Suite was originally coded for Linux and it’s wonderful that Blackmagic Design has kept that legacy alive. I’d love to know how the Linux version is doing commercially. Occasionally think about setting it up. I do own a license to Studio which I use on Mac.
Resolve is a point in favour of a Linux version.
Corel AfterShot Pro
Works inconsistently on all platforms. Started as the amazing Bibble with probably a Linux first codebase. Corel have botched the development and marketing of what should be one of the best and fastest RAW developers out there (and hence most popular). I’ve owned and paid for a couple of versions of AfterShot Pro post-Aperture before finding Iridient Developer and then DxO PhotoLab. Yes AfterShot Pro was fast but noise reduction was awful, images were soft. AfterShot Pro whatever platforms it is on is no indicator of wisdom or capability.
Pixeluvo
I don’t know anything about Pixeluvo, either its quality or its commercial success, yet.
Overall, if one goes through the list of commercial software on Linux in the audio/video/graphics section, most of it has status dropped. It’s not a coincidence.
DxO PhotoLab is not a dominant market leader who can force open Linux markets. DaVinci Resolve’s situation gives some hope that there is potential on Linux. Heck, I was thinking yesterday before reading your post, that if DxO PhotoLab worked at 100% on Linux, I could consider migrating my primary systems to Linux. A good photo editor is probably my number one missing piece of software on Linux and I’ve been completely fed up with Apple since they launched the app store (privacy reasons) and forced upgraded everyone to the absolutely awful 10.7.
So while in theory I’m on your side – I’d like to see PhotoLab on Linux – I’m very sceptical about the commercial wisdom of DxO investing in such a port.
I don’t say DxO to invest huge amount to write native Linux application with Qt (for example). But if Microsoft introduce WPF/WebForms under Linux with .NET 5 next year. So what should be done so PhotoLab to run on linux is:
Compile the UI agains .NET 5.
Lua scripts should work
Currently PhotoLab engine (OpenGL / Vulcan layer) works on MacOS so it is not tight to Windows and DirectX, which means that can be recompiled under Linux.
One problem can be graphic drivers, but this can be fix and to support only NVidia and ATI proprietary drivers.
If all this is true then the port should not take more than 1 month for 4 developers. Let say one developer cost 60 000 euro per year. A team of 4 devs for 1 month work it is 20 000 euro investment. The essential edition costs 107.5 euro without VAT, so we need to find 186 Linux users who want to buy a Linux version.
I don’t think you understand the real costs of doing business, including the cost of resources which goes far beyond the individual salaries.
First, how many developers do you think DXO has? Its a small group to start with, and they may not have Linux expertise. And, dedicating a few resources, even for a month or two, would have a major impact on Mac and PC support and development. The testing alone could take far longer. You don’t just port an application and pop it into production.
And then there is the ongoing support for upgrades, enhancements and bug fixes to a third platform. They would require permanent extra staff and adds significant complications and potential issues every time there is a new roll out to three different platforms. Suggesting that all you need to find is 186 Linux users willing to buy PhotoLab is simply not the case. When you have a software development business you have to look at all costs versus the benefits.
Excuse me while I clean my keyboard of coffee I just spat over it reading this !
Are you a software developer ?
Have you ever been involved in porting major software projects to another platform, especially Linux ?
Have you any idea just how much work is involved in porting a UI from one platform to another ?
The DxO UI is not multiplatform; the Mac version is created using Interface Builder (part of the Xcode IDE) and is not portable in any way to any other platform. One cannot “simply recompile” NIB files or storyboards that rely on Apple’s AppKit frameworks for their rendering.
As to your ridiculous assertion that it would only take 4 developers a month to do the work… which is why I now have to hang my keyboard out to dry
For example, I ported a Windows app, for pilots of private jets, to run on an iPad - it took me 2 years of 14 hour days, 7 days a week. 4 developer months my a**e
In terms of the work involved, Joanna is right on the money. In terms of the hidden business costs, Mark is exactly right.
If you can bring a €100 deposit for 5000 Linux users to the table, you might get DxO’s attention. Kickstarter, Indiegogo and GoFundMe require no up front investment. Go to it.
I won’t hold my breath. 100 paying Linux users are more like reality. Honestly DarkTable and RawTherapee are very powerful applications. Most Linux users won’t struggle with their interface conventions (unlike Mac users like myself). On top of that, off your own bat, you’ve named AfterShot Pro which is adequate pro software, despite its weaknesses (if there was no other photo processing software on earth, I could get the job done with AfterShot Pro) and the dark horse Pixeluvo.
It would make much more sense for a company like Adobe who is rich as Croesus after fifty years of Photoshop (the current leadership are accountants and MBA’s who are raping the company and its customers, just trying to milk the well dry, by squeezing every cent out of existing customers at the expense past goodwill: current American business practice is very short term) to try to extend out to Linux.
As Adobe is part of the NSA/CIA cabal at the center of American software (Microsoft is a founding member and Apple was forced to join in 2012 after the death of Steve Jobs), Adobe will not do anything to support Linux. Linux is a threat to the American digital spying and hacking as it can be made relatively hack proof. There’s been a lot of time and money spent, buying companies and compromising open source companies and mixing licenses to cripple Linux.
Why on earth would DxO want to step into these troubled waters? DxO has enough challenges in front of them without this very, very silly idea of extending support to a fractioned platform with at least six major distributions and which is fundamentally opposed to the idea of commercial software.
Hi Joanna,
You are right about the enormous work that will be needed to be ported if they start now. I don’t think that you follow my logic.
Windows version is build on WPF. Microsoft has some plans to bring it to all platforms in .NET 5. At the end nobody know if it will happen or not, but it is quite possible to introduce a cross platform XAML based desktop UI framework. Even now there is an Avalonia project (https://github.com/AvaloniaUI/Avalonia) which is cross platform and it is possible to migrate WPF projects to it. My idea is if someday there is an easy way to migrate WPF to Linux/Mac then DxO to port their product.
I hope that the graphic engine which runs on Windows and Mac is the same and is cross-platform build around OpenGL/Vulcan APIs, so to need only recompilation.
Lua scripts should not be a problem.
So to summarize if everything is a cross platform how long it will take? For example how long it will take to migrate an Electron or Qt Application to other platforms?
Many years ago, I used to be involved at an international level, developing software and mentoring other developers, using something called Delphi, published by Borland. In its day, it was truly ground-breaking and certainly knocked spots off Microsoft’s Visual Basic. A few years down the line, Borland decided to launch a Linux version they called Kylix. It cost a small fortune by Linux standards and, guess what? Sales were disappointing, to say the least, and it died a lingering but inevitable death.
Anders Hejlsberg was responsible for a lot of the design of Delphi and he went on to design the .NET framework for Microsoft. I was involved in writing some pretty massive frameworks and apps in C# and got to play with the Mono frameworks, that allowed us to build .NET apps on Macs. The problem with that was that you had to deploy the Mono runtime with the app and versioning could sometimes cause problems when Apple updated OS X.
There have also been several attempts at cross-platform UI development but most of them involved designing screens in code rather than an interactive visual designer. Borland’s attempt at such a framework suffers from something known as “uncanny valley” syndrome, where the UI looked almost like Mac, almost like Windows, but never something that felt truly native.
Qt is interesting, if it didn’t involve having to code in C++ when you’ve been writing in C#, Objective-C and Swift for years. But, once again, I’ve seen Qt apps that just don’t quite feel right on either Mac or Windows. Can you imagine what it’s like trying to create native look and feel that the different flavours of Linux would require?
A very wise man once said that cross-platform “business logic” was an amazing idea - cross-platform UI, more like a nightmare. From what I can see by tearing apart DxO, they are using truly native UIs, at least for macOS, which means we are getting the best possible user experience. But that comes at a cost of having to support and maintain two separate UIs, one for Mac and one for Windows. As others have said, I doubt if the possible market for a paid Linux version would really warrant the gargantuan cost and effort it would take to create a third UI development line.
If you are a commercial software developer, then you should know that testing a commercial application, especially a newly ported one that has not been in the marketplace on that platform, requires very vigorous testing that can take longer than the actual development, and must include a period of beta testing and subsequent updates, once the in-house testing has been completed. There is a cost to that for the business beyond the cost of the initial analysis and development.
Once implemented, there will be an ongoing significant additional cost in both money and resources to support, maintain, and update the software on that new platform. That is a cost that will continue forever. Updating testing and implementing new versions of that software on Linux will likely take the same amount of effort as it would for the Windows and Mac versions but with a lot smaller return.
To be successful on Linux they would need to be confident that there would be thousands of very likely customers, with the existing customer base willing to pay for annual upgrades for each new release, and new customers buying into their software every year. Frankly I don’t see it happening on Linux.
This is a business after all. There is significantly more potential for additional business growth if they stick to Windows and Macs. Porting to a little used platform to satisfy the needs of a small audience generally averse to paying for their software would be a poor business decision. Supporting Linux would be nothing more then a distraction from their core business model.
I know that’s a disappointment to people that chose, and love, to use Linux. But that is a conscious decision that people made knowing full well the dearth of commercial applications available on that platform. The situation is not likely to change much in the foreseeable future. if using some of the top commercial software currently available is important to you, then you may have to rethink your options.
Fascinating discussion.
The first RAW editing software I purchased was Bibble (4, I think), under Linux. I loved it, but I loved a lot less screen calibration and printing pictures. In fact, I just could not get it done. I am not sure whether I would do better now. Anyway, Bibble eventually did not work so well anymore, and I also figured I would do better processing my pictures under another OS.
As mentioned above, the diversity of Linux distributions, versions and environments is a major problem too (I had issues with Bibble, due to some distros not setting up certain libraries like the developers expected).
Maintaining good support across many distros would be a lot of work, so I am not sure it would be worth it in the case of DxO. I say this as someone who loves Linux and use it everyday for many things, including development. That being said, of course, if I see it happen, I will probably want to try it quickly.
Since it seems important, a few words of disclosure first: I’m a developer of commercial software based on .NET, using an older version of DXO and would love to see it on Linux.
As for the discussion;
The argument “which distro would you support” doesn’t really apply anymore than “which windows version would you support?” for the .net reasoning. There are at least 6 versions of just windows 10 and then myriads of different patch levels…what is done is that you support a specific .NET -version, and then possibly test on the windows versions you want to “support” (and find workarounds when MS has messed something up until they fix it or forever). That’s one of the main points of .net, you run the same clr on different architectures or OSes…the difference now is that they bring it to linux as well, but the idea stays the same, you don’t support a linux version or distro, you support a .net version, and possibly test a few to have a response when MS messes up.
