5 ms·

Because OSS doesn't work at scale, the economic incentives are not aligned. Sure, it may work sometimes with the goodwill of someone who cares, but 90% of OSS

Because OSS doesn't work at scale, the economic incentives are not aligned. Sure, it may work sometimes with the goodwill of someone who cares, but 90% of OSS code is a dead portfolio the authors built with the hope of landing a tech job or skipping some algorithm questions. Sure, OSS allow people to experiment with crap for free (even though it's mostly big corps benefiting from OSS) but what about the negative effects OSS produce on small businesses? How many developers could spend their life maintaining small parts of software instead of working in soulless corporations, if giving away your code for free (and without maintenance) wasn't so common? How much better would this code be compared to the wasteland of OSS projects? How much more secure could the entire ecosystem be? How many poor developers are working for free just in the hope of getting a job someday? We need to stop dreaming the OSS dream and start making things fairer for the developers involved.

jchw 3y ago
In the long term, software is not very novel. Needs evolve quickly in the genesis of a new category of software, but it doesn't take very long for it to stabilize for many categories. That's why, in my estimation anyways, open source is actually more sustainable in some cases: because when software becomes a commodity, it makes more sense for stakeholders to collaborate and collectively benefit from it. There is no "OSS dream" anymore—today, there is an OSS reality. We have some open source stuff that objectively works: there are business models that are more or less proven, at least as proven as any internet or software-oriented business model, and plenty of highly successful projects that deliver immense value to the world. Then again, some of it doesn't seem to work, and there are a lot of unknowns about how it works, what the dynamics will be, etc. But, if we're to call open source into question, we shouldn't forget to call proprietary software into question, too. Proprietary software has many seemingly-endemic issues that are hard to mitigate, and the business model has been shifting as of late. Software is now sold more as a subscription and a service than it is a product. The old business model of boxed software, it seems, has proven unsustainable for many participants. The main issue open source seems to have is really funding. It works well when open source software acts as a compliment to some other commercial business, but it works poorly when the software itself is where all the value is. After all, if you, for example, just host an open source piece of software in exchange for money, you're effectively competing in the highly competitive web hosting business. It can work since you can obviously provide some value, but it's a precarious position. Thus very few companies really have a lot of money to put into the Linux desktop, or at least if you wanted to compare it to Windows or macOS. It's complimentary to some of them who use it as a developer workstation or something, but there are only a couple companies who I think genuinely have a good model. System76 is definitely one of them, to be explicit about it. But rather than give up entirely, I propose something else: we should literally fund open source collectively. Obviously a lot of us already do: you can donate to projects like Blender or Krita monetarily, and you can donate your time and code as I'm sure many of us also do for random projects. But also, I think that open source should get (more) public funding. These projects arguably end up adding more value to the world than you put in them, and I think governments should take notice. Of course in some cases this has already taken shape for one reason or another. Consider Ghidra. Clearly released for the benefit of NSA's PR, but wait, why not release Ghidra? It's immensely useful, especially given that even at a fraction of the functionality of Hex Rays products, it's still extremely useful for many parties who could simply never afford the annual cost of maintaining IDA Pro licenses, especially now that it is only available as a subscription. The way I see it, software and computers in general are still moving quite fast even though it's clearly stagnating compared to where it once was. But, as things stabilize, software will simply need to be updated less, because we simply aren't going to have reasons to. As it is, old versions of Photoshop are already perfectly serviceable for many jobs. And at that point, we're only going to need one decent open source release for some given category of work. Things will need occasional improvements and updates, but c'mon, there's not an unlimited amount of potential to squeeze out of e.g. an image editor, any more than you can a table saw or a hammer. At some point you hit the point of diminishing returns, and I think we're nearing it in some places, hence why the switch to subscription models is necessary to sustain software businesses. It's a myth that open source is held up by poor developers starving for their ideology. I'm sure a lot of that exists, but a lot of open source is also side projects, work subsidized by companies for one reason or another, projects with healthy revenue streams or donations, etc.

jchw 3y ago

In the long term, software is not very novel. Needs evolve quickly in the genesis of a new category of software, but it doesn't take very long for it to stabilize for many categories. That's why, in my estimation anyways, open source is actually more sustainable in some cases: because when software becomes a commodity, it makes more sense for stakeholders to collaborate and collectively benefit from it. There is no "OSS dream" anymore—today, there is an OSS reality. We have some open source stuff that objectively works: there are business models that are more or less proven, at least as proven as any internet or software-oriented business model, and plenty of highly successful projects that deliver immense value to the world. Then again, some of it doesn't seem to work, and there are a lot of unknowns about how it works, what the dynamics will be, etc. But, if we're to call open source into question, we shouldn't forget to call proprietary software into question, too. Proprietary software has many seemingly-endemic issues that are hard to mitigate, and the business model has been shifting as of late. Software is now sold more as a subscription and a service than it is a product. The old business model of boxed software, it seems, has proven unsustainable for many participants. The main issue open source seems to have is really funding. It works well when open source software acts as a compliment to some other commercial business, but it works poorly when the software itself is where all the value is. After all, if you, for example, just host an open source piece of software in exchange for money, you're effectively competing in the highly competitive web hosting business. It can work since you can obviously provide some value, but it's a precarious position. Thus very few companies really have a lot of money to put into the Linux desktop, or at least if you wanted to compare it to Windows or macOS. It's complimentary to some of them who use it as a developer workstation or something, but there are only a couple companies who I think genuinely have a good model. System76 is definitely one of them, to be explicit about it. But rather than give up entirely, I propose something else: we should literally fund open source collectively. Obviously a lot of us already do: you can donate to projects like Blender or Krita monetarily, and you can donate your time and code as I'm sure many of us also do for random projects. But also, I think that open source should get (more) public funding. These projects arguably end up adding more value to the world than you put in them, and I think governments should take notice. Of course in some cases this has already taken shape for one reason or another. Consider Ghidra. Clearly released for the benefit of NSA's PR, but wait, why not release Ghidra? It's immensely useful, especially given that even at a fraction of the functionality of Hex Rays products, it's still extremely useful for many parties who could simply never afford the annual cost of maintaining IDA Pro licenses, especially now that it is only available as a subscription. The way I see it, software and computers in general are still moving quite fast even though it's clearly stagnating compared to where it once was. But, as things stabilize, software will simply need to be updated less, because we simply aren't going to have reasons to. As it is, old versions of Photoshop are already perfectly serviceable for many jobs. And at that point, we're only going to need one decent open source release for some given category of work. Things will need occasional improvements and updates, but c'mon, there's not an unlimited amount of potential to squeeze out of e.g. an image editor, any more than you can a table saw or a hammer. At some point you hit the point of diminishing returns, and I think we're nearing it in some places, hence why the switch to subscription models is necessary to sustain software businesses. It's a myth that open source is held up by poor developers starving for their ideology. I'm sure a lot of that exists, but a lot of open source is also side projects, work subsidized by companies for one reason or another, projects with healthy revenue streams or donations, etc.