13 ms·
It's a bunch of stuff. - Standard libraries for audio software are less tested on Linux and tend to be buggier there, and we have to chase down those bugs. -
by wheels 3y ago
It's a bunch of stuff.
- Standard libraries for audio software are less tested on Linux and tend to be buggier there, and we have to chase down those bugs.
- There are a bajillion versions of Linux, and even if you say you only support some subset of those, Linux users will ignore that and still write to support when it doesn't work on their distro / version. Sometimes that even extends to writing bad reviews online.
- Packaging on Linux is annoying since there are a bunch of different formats, and nothing will make everyone happy.
Because of those issues, we probably get almost as much support mail for Linux as for macOS or Windows, despite Linux being 1% of our user base. From a cost perspective, it just doesn't make sense to keep supporting it. Ironically, we'll keep developing it, because we do some of the development on Linux, and often test on Linux before Windows. From what I know of other companies: we're not unique there, and I thought some of this might be interesting info for pro-audio Linux users.
- hammyhavoc 3y agoIf it were me, I'd just focus on making sure it works well on Fedora. PipeWire is in the OOBE, and is where a good chunk of pro audio folks have settled due to top-notch PipeWire support.
- wheels 3y agoThis is exactly the sort of email we get. Except that this would literally be the first request for Fedora, exactly exemplifying our problem. (Most of our users request things with Ubuntu / Debian bases.)
- hammyhavoc 3y agoI'm not requesting you do anything, I'm not even a customer (as far as I know), I'm saying what I would do in your position if I was aiming to support an audio product on Linux. I explained why I would choose Fedora. It's your business, you run it as you see fit. I can't help but feel that maybe you read too much into what you get sent. I would pick one "known good config" (your distro of choice), present it as what it gets tested on prior to release, and for anything else, you're on your own, that's part-and-parcel to being a Linux user. It assumes some kind of ability to troubleshoot and solve your own problems. It isn't for everybody, just like a Mac or Windows isn't for everybody. Officially, DaVinci Resolve has CentOS as the only supported config, but it runs on other Linux distros just fine. Doesn't seem to hurt its popularity on Linux. If you go back a couple decades, there was even developer-recommended hardware for specific DAWs. People will use whatever gets the job done, your product or not, Linux or not.
- gcr 3y agoI'm glad you have good ideas about how to run a business like this. But I think your post illustrates the idea that, unlike windows or mac users, us Linux users have opinions about how the business should be run, and we tend to voice them through support channels and emails and social media etc. Every time we voice our ideas, it generates a ticket or a discussion or a decision that they have to deal with. The act of contemplating these possible ideas, whether that means discussing them with the team or deleting the email or reading the slashdot thread, costs the business time and attention. Each time a company makes a concrete decision -- "let's only support Fedora," to use your example -- 80% of the potential linux user base you'd even care about (ubuntu users) is out of luck. This is true for "let's care only about pulseaudio," "let's only care about ALSA," "let's only care about this specific glib version," etc. The intersection of these decisions is very small. You're saying they should pick one supported configuration and stick with it. That's what they're already doing though -- it's just not an operating system that runs Linux. We can keep around a hardware-specific Windows box just the same.
- hammyhavoc 3y ago> us Linux users have opinions about how the business should be run I have plenty of public and private opinions about software that aren't Linux-centric too, and you're assuming that I only use Linux, when no, a good chunk of my workflow is on Windows because that's where the software and driver support is for cutting-edge DSP and video game middleware. Some companies even pay for consulting. Businesses that rely on selling licenses to customers for new features heavily rely on feedback. You only have to look at the likes of the Native Instruments forum or their Centercode to see that providing feedback isn't some kind of Linux-exclusive phenomena, and most reasonable businesses encourage that feedback. However, triaging and actioning that feedback is an entirely separate matter. > it generates a ticket or a discussion or a decision that they have to deal with They don't need to do anything. That's entirely a business decision whether or not a company acts on user feedback, or feedback from potential customers. A lot of biz pay handsomely to run focus groups and and actually get feedback whatsoever, but there's a big difference between a social media discussion, a feature request, a bug report, or a support request. Who is seeking support for an audio product via Slashdot? Nobody. It's an audio product on Linux. You go where the devs are, where you'll get help, if they're indeed offering it. If somebody buys a license for x on Linux, you ask them what OS they use on the contact form in a dropdown or text field on a webpage, you use conditional logic where if it isn't the supported OS, you tell them "sorry, this isn't a supported configuration, if you would like support, please use the recommended/supported distro, here's a link for more info". This eliminates almost all cases of this. Prior to selling them a license, you have them tick a box that acknowledges that Linux support is limited and that unless they use x, they're SOL on getting support unless they go to the forum or elsewhere and figure it out either on their own, or with others. > The intersection of these decisions is very small. Then those customers or potential customers make a decision as to whether it is appropriate for them, if they are willing to compromise by changing x, or if Linux as a whole is appropriate for them in the same way that people who want to use Logic Pro will need to buy a Mac, and that might mean that some plugins aren't available on macOS, or aren't yet natively supporting Apple Silicon, and if you want to exclusively use Windows then you aren't going to be using MetaSynth. Renoise users on Linux seem to be able to navigate this without problem in terms of understanding `glibc >= 2.27 (e.g. Ubuntu 18) on x86_64 based systems`. Same goes for Bitwig Studio users understanding Flatpak and PipeWire support. If people want to have their cake and eat it, they use Windows or macOS. Linux users are accustomed to compromise, it goes with the territory. > That's what they're already doing though And he dismissed my response as being an example of the problem when that wasn't what I was saying whatsoever, thus was not an example of his problem, and now you're latching onto that like I didn't already understand this. Furthermore, this entire discussion ignores the potential for a return by supporting an influential power user on a niche OS. It may actually bring significant sales of something in. That's strategy. E.g., DaVinci Resolve may not have the majority of its userbase on Linux, but because x major director or studio uses DaVinci Resolve on Linux on y film, they will sell a lot of software and hardware to z users on other operating systems. Audio has some of the most influenced people niche-wise because they want to use what their hero uses, hence signature versions of guitars, hence signature guitar pedals, hence VST collabs, and preset and sample packs from x producer. tl;dr: it's a misconstrued and misrepresented throwaway comment that of course turned into an irrelevant tangent because it's HN. On an interesting thread for an interesting product launch, and this is what people focus on? Exhausting.