Why Linux is Becoming Impossible for Audio Developers to Ignore

[ghettosynth]

Also don't forget that there are other negligible niche OS down there contributing to that bucket as well, plus (as you note) bots.

Nothing wrong with niche OS - I am actually a huge fan of NetBSD and I have used linux as my sole desktop a few times in the past, including a two year stint on gentoo.

But arguing that you are an unignorable market force is pretty funny and not the way you will gain developers.

To be clear, I'm not arguing that. I don't care what devs do or don't do. I have never supported casual users on Linux or promoted the idea that people should switch. If you have to ask, then no, you shouldn't switch, it's still not for you. It's for us, and we don't want you peeing in the pool.

What I did say is that there are aspects that are driving greater acceptance. That doesn't necessarily convert to audio users. However, gaming on Linux is become a thing, in part, because people want set-top box style control and Linux+Steam provides that. Valve's devices are going that direction and have been for a while. Microsoft's AI and crash nonsense this year is fostering a lot of conversation that is angry. Whether that is just threats, I don't know. However, it was my impetus to switch my main desktop back this year. Eurocentric digital sovereignty is also having some impact. None of these are smoking guns, in fact, the second point is as likely to drive Mac sales.

But, you can't say "other negligible niche OS" without a reasonable discussion of what they are. I have a FreeBSD system across the room. It's not a dekstop OS though. NetBSD/FreeBSD, Amiga whatever, are tiny tiny niches as compared to even weird Linux OS systems. Which niche OS's exactly do you think are driving those numbers? All of them taken together that aren't Linux won't move a needle.

The probe that I gave you the link for is a data point. You can use it, or not use it, I don't care. Of the probed 15M desktops about 6% were Linux. Not NetBSD, not AmigaOS, but Linux. Of course there's an error rate associated with that, as there is with any number.

[stoopicus]

Probably yep. But it's still not a significant desktop market, which is my whole point in this thread. There's lots of ways for linux to attract developers, but claiming unignorable market share is not one of them.

Sell it instead on the merits of the development environment, or even just that it sucks less than Windows for some things. As a software developer I love linux for day to day development - but have also been doing development on unix based operating systems for decades. You need to sell the merits because to many developers they are not obvious at all.

I'll go with that. The title of the thread was my doing, so I'll take full responsibility. However, did you notice that MacOS adoption has been dropping over the last several years? It's currently less than 9% total. Windows has been dropping adoption like crazy too. So, for the last several years, while Windows and MacOS, has been dropping in adoption, Linux has been growing in adoption. You don't find that an interesting trend?

They are doing a weird thing where they split MacOS and OSX.

[stoopicus]

The probe that I gave you the link for is a data point. You can use it, or not use it, I don't care. Of the probed 15M desktops about 6% were Linux. Not NetBSD, not AmigaOS, but Linux. Of course there's an error rate associated with that, as there is with any number.

Can't look right now but one thing I am curious about is differentiation of desktop vs non-desktop usage there. I can look later but any quick breakdown on how that worked? Like my lan at home has multiple non-desktop linux devices, etc

[Tiles]

Well if we are going to lean in to the "Unknown" then maybe linux users can use The Force to mind trick some developer adoption. I'm not the one that brought up StatCounter but if it is going to be used as the data source (like those quoted articles above did) then it is fair game to go take a look at.

Mh maybe a little thought here. Statcounter uses the same method since years. Be it right or wrong. And if the values remains nearly unchanged, then that's a clear sign.

Either way, percentages can be misleading. A rise from 2% to 4% can be presented as "Linux gained 100% market share" and is growing like crazy. But going from 2% to 4% does not necessarily matter if a software vendor needs 10% reachable market share just to break even.

A 100% increase sounds impressive until you look at the actual numbers.

[ghettosynth]

The probe that I gave you the link for is a data point. You can use it, or not use it, I don't care. Of the probed 15M desktops about 6% were Linux. Not NetBSD, not AmigaOS, but Linux. Of course there's an error rate associated with that, as there is with any number.

Can't look right now but one thing I am curious about is differentiation of desktop vs non-desktop usage there. I can look later but any quick breakdown on how that worked? Like my lan at home has multiple non-desktop linux devices, etc

From this article.

https://www.zdnet.com/article/think-lin ... otherwise/

Lansweeper's data, according to CTO Guido Patanella, comes from its agents and scanning of 15 million consumer systems and 3.5 million corporate systems. Much of the consumer data comes from Fing, a home and small office network monitoring and device inventory company that Lansweeper acquired in 2020. The company defines consumer PCs as computers that are workgroup-based, standalone, or bring-your-own-device.

The following is almost certainly true. It's been true for me for a number of years. I should say that even when Windows was on my desktop, there was always WSL instances installed. Windows without WSL is an intolerable OS situation for me.

There are several reasons why Linux is finally gaining popularity. One I haven't looked at recently but deserves mention is the rise of AI programming. Linux has become the default operating system for AI and machine learning workloads due to the prevalence of leading AI open-source frameworks, such as TensorFlow, PyTorch, and Hugging Face Transformers, and OpenAI's new open-weight models.

Now, from the same article, this article which references a study is more interesting.

https://www.zdnet.com/article/linux-has ... heres-how/

This is access logs to government websites. This is better data than stats counter as it's more likely to be representative. Even here though, it likely undercounts. Many people who use Linux have a Mac or a Windows machine around that is only used for this kind of access. Edge has two good uses, an extension free browser when nothing else works, and to read and mark pdf files.

At any rate, you can play with the views on the data in real time, have fun.

https://analytics.usa.gov/

Now the AI desktops could be overcounting Linux "users", but that doesn't change the percentage of desktops, so that may not mean that the number of customers follows directly.

[camsr]

There is probably more statistics from the user-agent string the browser broadcasts. I woe the day that user-agent string becomes a "second-class citizen" for retaining it's freedoms.