3 ms·

Time for the real hard truths. Almost all major open source success are the product of subsidized development, either by a commercial arm (Linux via Red Hat et

Time for the real hard truths. Almost all major open source success are the product of subsidized development, either by a commercial arm (Linux via Red Hat et al., gcc by Google et al., Java by Sun et al., etc.) or a research arm (OCaml by Inria, Haskell by MSR, etc.). Open source by hobbyists doesn't work (at scale).

ScottBurson 14y ago
Could crowdfunding change that? https://bountyoss.com/ https://bountyoss.com/
rjzzleep 14y ago
wait a sec, a bunch of researchers doing what they want is subsidized development? maybe in the states. in europe a bunch of researchers doing what they want is just that. a bunch of researchers doing what they want. and then comes a us company making money off of it. "sorry dude, you're not a hobbyist, you don't count as free, because you have a day job, and get paid to do what you want." /facepalm not to mention that most got developed FIRST, and then got funded after showing off their value proposition
- Locke1689 14y ago
  I have no idea what you're trying to say. I'm saying that major open source success only happen when a funded entity, e.g. a corporation or a research institution, pays for someone to develop that project full-time. In other words, successful projects have someone paying for full-time salaried programmers.
berlinbrown 14y ago
Corporations should do a better job of adopting open projects (see Google) and Open source projects should do a better job of recruiting corporations to get involved.
- bunderbunder 14y ago
  Corporations should do whatever is in their best financial interests. That's not a philosophical assertion, it's a legal mandate. Consequently, open-source projects that are critical to companies' success are already well-supported by those same companies. And most the major open-source projects have plenty of commercial support - in a nutshell, what organizations like the Apache Foundation and the Eclipse Foundation are. The only spot where open source really frays at the edges is consumer software such as Linux desktop environments or office software. That's sort of inevitable - programmers are expensive, and from a company's perspective hiring developers to keep GIMP up to speed is orders of magnitude more expensive than a stack of Photoshop licenses, and won't necessarily deliver the goods until months or years down the line (rather than minutes). So projects like that don't end up being supported as well as we'd like because it's not really in many companies' financial interest to do so.
jiggy2011 14y ago
From my understanding nobody seriously involved in FOSS expects hobbyists to scale apart from on small or niche projects. The term "open source" rather than "free software" was coined at least partly in order to make it easier for business to understand and adopt the software and incentivise them to throw paid manpower at it. It seems to have broadly worked too, "we should adopt this solution because it is open source" is something that MBA types actually say now.
- Locke1689 14y ago
  I don't know if that's really the case. The existence of the AGPL implies that many people see open source as a development model quite different from the one used by open source inside corporations. In other words, how open source is something which is used and modified internally by company X, but rarely gets changes released because company X is exclusively SaaS? There are people who have very different opinions on what open source means. Some of them think that a lot of open source business models are bad for open source. It is important to point out that a lot of these licenses are the reason why certain open source projects are widely successful.
  - jiggy2011 14y ago
    The AGPL makes sense for either dual licensed things (like MongoDB) or for projects that simply don't have the need for any direct revenue generation such as people's side projects. AGPL seems to be to the cloud what GPL is to desktop. There are certainly those such as Stallman who are basically opposed to SaaS, though I'm not entirely sure where he draws the line between "in house software" , websites and SaaS. It does seem to me that asymptotically software with freer licenses "win". For example the Linux kernel. I would not be surprised if facebook was eventually replaced by something that existed under an AGPL license. However such a system would probably be powerful enough to be considered a platform in itself with it's own DSLs and therefor this would become where proprietary software lived. From a pragmatic standpoint though the increase in SaaS seems to have vastly increased the amount of open source stuff available (rails etc) and perhaps this trend will continue as higher abstraction levels of the stack become "freer".

ScottBurson 14y ago

Could crowdfunding change that? https://bountyoss.com/ https://bountyoss.com/

rjzzleep 14y ago

wait a sec, a bunch of researchers doing what they want is subsidized development? maybe in the states. in europe a bunch of researchers doing what they want is just that. a bunch of researchers doing what they want. and then comes a us company making money off of it. "sorry dude, you're not a hobbyist, you don't count as free, because you have a day job, and get paid to do what you want." /facepalm not to mention that most got developed FIRST, and then got funded after showing off their value proposition

Locke1689 14y ago
I have no idea what you're trying to say. I'm saying that major open source success only happen when a funded entity, e.g. a corporation or a research institution, pays for someone to develop that project full-time. In other words, successful projects have someone paying for full-time salaried programmers.

berlinbrown 14y ago

Corporations should do a better job of adopting open projects (see Google) and Open source projects should do a better job of recruiting corporations to get involved.

bunderbunder 14y ago
Corporations should do whatever is in their best financial interests. That's not a philosophical assertion, it's a legal mandate. Consequently, open-source projects that are critical to companies' success are already well-supported by those same companies. And most the major open-source projects have plenty of commercial support - in a nutshell, what organizations like the Apache Foundation and the Eclipse Foundation are. The only spot where open source really frays at the edges is consumer software such as Linux desktop environments or office software. That's sort of inevitable - programmers are expensive, and from a company's perspective hiring developers to keep GIMP up to speed is orders of magnitude more expensive than a stack of Photoshop licenses, and won't necessarily deliver the goods until months or years down the line (rather than minutes). So projects like that don't end up being supported as well as we'd like because it's not really in many companies' financial interest to do so.

jiggy2011 14y ago

From my understanding nobody seriously involved in FOSS expects hobbyists to scale apart from on small or niche projects. The term "open source" rather than "free software" was coined at least partly in order to make it easier for business to understand and adopt the software and incentivise them to throw paid manpower at it. It seems to have broadly worked too, "we should adopt this solution because it is open source" is something that MBA types actually say now.

Locke1689 14y ago
I don't know if that's really the case. The existence of the AGPL implies that many people see open source as a development model quite different from the one used by open source inside corporations. In other words, how open source is something which is used and modified internally by company X, but rarely gets changes released because company X is exclusively SaaS? There are people who have very different opinions on what open source means. Some of them think that a lot of open source business models are bad for open source. It is important to point out that a lot of these licenses are the reason why certain open source projects are widely successful.
- jiggy2011 14y ago
  The AGPL makes sense for either dual licensed things (like MongoDB) or for projects that simply don't have the need for any direct revenue generation such as people's side projects. AGPL seems to be to the cloud what GPL is to desktop. There are certainly those such as Stallman who are basically opposed to SaaS, though I'm not entirely sure where he draws the line between "in house software" , websites and SaaS. It does seem to me that asymptotically software with freer licenses "win". For example the Linux kernel. I would not be surprised if facebook was eventually replaced by something that existed under an AGPL license. However such a system would probably be powerful enough to be considered a platform in itself with it's own DSLs and therefor this would become where proprietary software lived. From a pragmatic standpoint though the increase in SaaS seems to have vastly increased the amount of open source stuff available (rails etc) and perhaps this trend will continue as higher abstraction levels of the stack become "freer".