I tend to disagree.

You may see GNU folks as crackpots, but in many aspects they are right, for business reasons.

I have an increasingly hard time to admit in my core stack anything that is not open-source (even if not free-as-in-freedom). Commercial offerings come and go, change their terms, etc. They can do something that goes contrary to my interests, and I have no recourse, and nobody has.

This is on the top of problems while troubleshooting. Being able to look at the source, understand what's happening, and maybe even quickly patch it is indispensable when running a production system, and suddenly encountering a problem.

That is, I don't use open-source programs because they don't cost money. E.g. I've donated more to Mozilla than many commercial products cost. I happily install paid versions of open-source apps on Android, as a way to pay the developers. I won't mind to buy paid support for open-source products.

But the openness is very important for something that is not a quick fix for a temporary problem.

Being able to look at the source and troubleshoot it is not mutually exclusive to a product being proprietary (ie, source available would give you this).

There is also plenty of business I know that do everything in their power to ensure that if they go away the customer isn't left helpless (this is very popular in some parts of industrial engineering where it's common sense to hand over the blue print along with a service contract).

It's just very popular for software vendors to behave without care for the user but it's not the only business model and I think that's a common mistake of some people (ie, believing that FLOSS and proprietary is a binary property that will automatically mean certain things like you mentioned).

Sure, if you're a big enough customer, you can have access to (relevant parts of) source code, and various forms of preferential treatment. I've seen his firsthand. This is the norm with big vendors that are on the market for many decades, and not going anywhere (e.g. Microsoft).

If you are a smaller fish, you have fewer of these options, because they just cost money.

Of course it's often cost-effective to quickly start using a piece of commercial software or SaaS, e.g. Slack or Gmail, as opposed to hosting your own chat or email with opensource solutions. As with any software where you put your important data, you need to know your way out, and a way to transfer these data to different software / SaaS. With email, it's reasonably easy; don't know about Slack. With open-source solutions it's usually reasonably easy, too, and it's easier to find, or at least to produce, a migration tool.

In some cases, there's just no free alternative; e.g. if you are writing a PC game, you can't realistically stay away from Windows.

As I mentioned, this isn't about existing vendors but potential business models (or rather, rare business models since I do know vendors that do this even with very small hobbyist users).

As one of the "Linux for FLOSS" users--and, yes, quite loud about it--I am inclined to agree.

I would pay good money for a high quality application on Linux. But in large part, one of the less-philosophical, more-practical things that I value about FLOSS is the opportunity to fix things if something doesn't work right. While I would gladly purchase an incredible, closed source application, I would much more value the distribution of the source code (with a workable build chain) along with that purchase.

In that manner, I KNOW I'm not just a buck in their wallet. It protects my purchase, to some degree, and demonstrates a good faith relationship with the developer. Even if I'm not allowed to distribute the code, and much prefer FLOSS, the right to maintain my purchase is incredibly important to me.

Of course, that's almost never an option, but to my mind, that would be a fair compromise...

~A Loud FLOSS Supporter

I've always thought that this is the key to bridging the gap between proprietary and OSS, and it's what we do with our software. In our case, the source code does cost extra, but that's because we're selling developer libraries and we typically need to provide extra support to customers that purchase the source code because they are also (typically) customizing the source code in some way.

I really like how Epic is handling Unreal Engine 4. The source is available, and the standard license is revenue share. It's not necessarily open source, but it reaped so many benefits from public contributions. You can modify it and extend it for your own use. The revenue share means they are actively interested in your success.

Of course video games are highly public consumer products. So you can't actually abuse their model because if your game is successful it will be well known.

I think you may be on to something here. For smaller companies like us that are in niche markets, the source code being available isn't a big deal because the level of piracy is fairly low (and you don't get support).

For larger companies in larger markets, it's much, much easier to see when someone is abusing a license or pirating the software without paying.

> I would much more value the distribution of the source code along with that purchase.

Everyone would until they realize how exceptionally hard and draining it is for a vendor to deal with blatant rip-offs and trivial repackaging of said source code, sold under a different name for a fraction of the price by some random outlet in China.

That would be another options, yes, though I don't think I've seen many employ source-available while also giving access to a working build chain/script.

> I would probably not be wrong when saying that most Linux users will happily pay a developer for the binary of an application if the dev is "doing it right".

Hm, this reminds me of the Caddy, the HTTP server with Let's Encrypt integration. It's still open source, but a while ago they made the binaries paid (if you want it for free you need to compile it yourself). They got a lot hate for that; people were even accusing them of having become closed source.

I stopped used Caddy on that event. It wasn't purely for the binary, the server also has a number of issues when mainly used as a proxy and a number of other problems for more advanced use cases.

I think the problem really is, atleast what I thought, was that the binary wasn't something I'd pay for. I could trivially compile it myself. Additionally the price was IIRC quite high for some hobbyists like myself. Plus putting advertisements in the header (or was that a separate event?)

If it had been a 10$/a license for the binary I would likely have done that but at the time it seemed cheaper to just switch to nginx.

Also keep in mind the different audience. There is server admins and developers. When I wear my server admin hat and license software, that is usually because there is no or not comparable free alternatives that I couldn't setup myself in a few hours of work. When I wear my developer hat I'm lazy and I will pay for software to save a few minutes of work if it's good enough.

Of course, I don't agree with the people who said it had become closed source, that is silly.

And then when Let's Encrypt's certificate renewal server was down for a few hours, Caddy wasn't able to start because it required to establish a connection with LE to check the certificate and it wasn't able to determine that there was an old, still valid cert available on disk. The bug tracker on Github went wild because people were unable to bring their websites back up but the author was unwilling to acknowledge it as a bug and kept closing the issues. [1]

That was when Caddy lost all its reputation for me.

1: https://github.com/mholt/caddy/issues/1680

I am not on any side in this battle, but the issue was closed only once, and after users disagreed with the author's reasoning for why it behaved the way it did, the author fixed it the same day it was closed...

Doesn't seem as bad as you are portraying it... certainly not "lost all its reputation for [you]" bad?

I think the Caddy mess had a bunch of factors complicating it, and isn't a good comparison. E.g. they promoted themselves as an easy and powerful solution for people that aren't really sysadmin/server types, and then hurt that exact demographic by from one day to the other requiring them to either pay a not-insignificant amount of money or learn how to compile a product they choose because it's really easy themselves. People really don't like bait-and-switch.

I cannot see someone trying to get paid for their work as a "bait and switch".

Then see it only as "surprising and major change, negating the previous unique selling point, makes people unhappy".

I can't really see it as that, either. The source is still open, it's free to check out and compile.

> free to check out and compile

Which isn't a problem for you and me, but is for the kind of person that was attracted to it because it's easier than nginx+certbot, as I said in my initial comment. I saw 2 major users of it: this kind of user, and people wanting a small simple thing to stick in their container-based setup for HTTPS/reverse proxy duties. The latter more or less automatically updated to container images that do the build instead of loading a binary and didn't have a problem with even having to consider paying. For the first it was a problem, especially since the cost wasn't really in the ballpark of "I'd happily pay that to host a small blog".

I started using Linux for FLOSS reasons about 15 years ago, but now that I'm older and mostly just care about my computer working the way I want it, without bugs and distractions, I have really shifted away from my old free software opinions.

I still use Debian Linux for my home computers. My wife even uses a XPS Developer edition. I would like to switch to Windows 10 for the battery life upgrade, but I find that Windows is slow and has a lot of distractions (eye candy and advertisements) compared to my Linux computers. The virtual desktop management isn't as convenient as it is on Linux (Mate desktop). Emacs, Git, Elfeed, and a lot of the other Emacs-related features/plug-ins I like to use everyday are also dog-slow on Windows, which is the real deal-breaker for me.

I use it partly for FLOSS reasons. But I'm not an absolutist. I still pay for JetBrains IDEs, and I pay for Mailspring (https://getmailspring.com/), among others.

FLOSS is also arguably more important in some software. OS is most important. Technically IDEs are very important too, but there just isn't an open source alternative for JetBrains IDEs. And they definitely deserve the money.

In principle I'd happily pay for a good application. But not being able to use my normal package manager to install it and keep it up to date would already be imposing a significant cost.

That's not totally unsolvable. Flatpak leaves this kind of stuff open so you could implement a storefront in flatpak.

PPA's on Ubuntu would do to but other distros are less equipped to deal with it, in that case an AppImage might be better.

Every year end when I do my charitable donations I go through the applications I’ve been using and donate 10-20$.

It’s not a ton but it’s something. Libre office/Mozilla/ eclipse foundation/ sequel pro/ aqua emacs...

I think they do even if they don't know it. After all, GNU/Linux is the way it is because it is FLOSS. You might just as well say people don't use cars because of the wheels.