Ask HN: Taxpayer funded open source?
5 comments
This is already happening here and there. The DOE made a policy decision to use open source on their supercomputers so that they would be vendor-indpendent; this has resulted in open source software like Lustre and SLURM. The E community was funded by DARPA. I'm sure there are others.
I don't know of any umbrella initiatives to fund open source in general. Given the ridiculous amounts of "stimulus" money being given out, go for it.
I don't know of any umbrella initiatives to fund open source in general. Given the ridiculous amounts of "stimulus" money being given out, go for it.
Let's say we pay 1000 programmers $50,000 a year to work on open source projects.
What makes you think that $50k/year is enough to attract good developers? We're talking about people who could easily earn double that -- and many of the top open source developers spend most of their paid time writing open source code anyway.
What makes you think that $50k/year is enough to attract good developers? We're talking about people who could easily earn double that -- and many of the top open source developers spend most of their paid time writing open source code anyway.
Then double it. I'm brainstorming here.
The chance to benefit society could be a compelling factor. Why do people become school teachers?
Also there's the fact that hackers like to share code and ideas. Actually, this is probably the biggest reason I would choose to participate in something like this. Hell, I'd do it for 10K.
Yes there are currently people doing open source work, but there could be far more. Also the areas in which they work are limited to those that will provide direct benefit to the funding companies (Mozilla, the Eclipse Foundation).
The chance to benefit society could be a compelling factor. Why do people become school teachers?
Also there's the fact that hackers like to share code and ideas. Actually, this is probably the biggest reason I would choose to participate in something like this. Hell, I'd do it for 10K.
Yes there are currently people doing open source work, but there could be far more. Also the areas in which they work are limited to those that will provide direct benefit to the funding companies (Mozilla, the Eclipse Foundation).
So that's it? I propose a straight-forward plan to revolutionize software development and all anybody has for me are a few lame objections?
How disappointing.
How disappointing.
Whenever the idea of government getting involved with startups or silicon valley comes up it makes me sick to my stomach. Picking the right startups, open source or otherwise is a hard thing to do and especially for the government. They have done that in the past, the government trying to fund the semiconductor industry in the 90s, and failed remarkably. I wish they learnt from their mistakes but alas seems like they don't.
As far as your idea of funding open source only projects, that already happens in universities. Most of the good open source project you probably use comes from these universities that get federal funding.
As far as your idea of funding open source only projects, that already happens in universities. Most of the good open source project you probably use comes from these universities that get federal funding.
Ok, so government should fund code development, not startups. Companies can then take the resulting code and figure out how to turn it into consumer value. I'll change the title to "Taxpayer funded open source".
Yes universities are a good source for OSS, but the obvious problem is that they're restricted to university students and faculty. Plus academic bureaucracy is probably not conducive to effective software development.
Yes universities are a good source for OSS, but the obvious problem is that they're restricted to university students and faculty. Plus academic bureaucracy is probably not conducive to effective software development.
The government should fund code development? How would individual programmers pitch their projects to the government? Not all code that's created, even if it's free, is ever useful to anybody. Figuring out how to "turn code into consumer value" is a very hard problem, as I'm sure any entrepreneur would tell you. And it's sort of a backwards way of looking at how software, and the people that make it, can benefit themselves and society best. First, there needs to a problem. Then, you solve it.
Anyways, I think the kind of thing you're talking about already exists in a more general way - government contracts. The government decides they have a problem, comes up with a list of specifications, and bidding on the proposed solution is open. The most qualified company that solves the problem for the least amount of money or in the shortest amount of time wins. (At least, that's how it's supposed to work.)
Many contractors that develop for the government, and bid on these contracts, already use FOSS because it gives them a competitive edge when bidding. For instance, some versions of the FBCB2, a communications system for humvees and other military vehicles, run on Red Hat: http://en.wikipedia.org/wiki/FBCB2 and http://74.125.95.132/search?q=cache:LLwHBOS8JmYJ:www.cjmtk.c...
I think that academia and savvy businesses are really where it's at with open source. As bad as academic bureaucracy can be, government bureaucracy is in my experience much, much worse.
So, all things considered, I'd think that even if the government had a direct interest in improving open source software, the best way to do it is through more funding for education - get more people to school, and get more people programming.
Anyways, I think the kind of thing you're talking about already exists in a more general way - government contracts. The government decides they have a problem, comes up with a list of specifications, and bidding on the proposed solution is open. The most qualified company that solves the problem for the least amount of money or in the shortest amount of time wins. (At least, that's how it's supposed to work.)
Many contractors that develop for the government, and bid on these contracts, already use FOSS because it gives them a competitive edge when bidding. For instance, some versions of the FBCB2, a communications system for humvees and other military vehicles, run on Red Hat: http://en.wikipedia.org/wiki/FBCB2 and http://74.125.95.132/search?q=cache:LLwHBOS8JmYJ:www.cjmtk.c...
I think that academia and savvy businesses are really where it's at with open source. As bad as academic bureaucracy can be, government bureaucracy is in my experience much, much worse.
So, all things considered, I'd think that even if the government had a direct interest in improving open source software, the best way to do it is through more funding for education - get more people to school, and get more people programming.
The most obvious way for a programmer to pitch to their projects is to use a committee of experts, like YC does. I'm sure we could come up with other methods too; maybe some sort of web-based voting scheme, for instance.
Government contracts are nice for solving certain problems, where you have a narrow goal to achieve, but don't address other problems, such as those that call for a more general, trial-and-error, research type approach. For example, say you wanted to work on Emacs, or the Leo editor. I doubt you could get a government contract to do that under the current system, despite the fact that your work has the potential to benefit all the programmers who use those programs.
You may be right about academia being the best way to go. I myself dropped out of college after two years. It felt like a big waste of time to me. You have to do a lot of irrelevant work and it's expensive as hell. The prospect of enduring another two years of that, followed by a Masters program and then a PhD just makes me cringe.
I don't see any reason why we can't create an organization with minimal bureaucracy that allows hackers to do their thing.
Government contracts are nice for solving certain problems, where you have a narrow goal to achieve, but don't address other problems, such as those that call for a more general, trial-and-error, research type approach. For example, say you wanted to work on Emacs, or the Leo editor. I doubt you could get a government contract to do that under the current system, despite the fact that your work has the potential to benefit all the programmers who use those programs.
You may be right about academia being the best way to go. I myself dropped out of college after two years. It felt like a big waste of time to me. You have to do a lot of irrelevant work and it's expensive as hell. The prospect of enduring another two years of that, followed by a Masters program and then a PhD just makes me cringe.
I don't see any reason why we can't create an organization with minimal bureaucracy that allows hackers to do their thing.
Guess Arpanet was a huge mistake because it was government funded.
Arpanet wasn't a startup, it came out of a defense program. What I am saying it trying to pick winners in silicon valley is a bad idea.
The government picks winners all the time when awarding contracts and has a lot of experience doing so (when the process is done right). For Arpanet they picked a winner with BBN.
If governments funded various software projects you would remove the need for those projects to turn a profit, thus removing a big obstacle to the sharing of code. I believe this would benefit society greatly.
If the code is released under a business friendly license, corporations could effectively have their R&D costs outsourced for free. This would allow them to focus on their customers more. This would be good for the economy.
I could see excessive regulation being a problem. I'm not sure how much oversight would be appropriate. I wonder what would happen if you just gave 100 hackers $20K over the course of a year and let them choose what to work on, just so long as it was open source.
A more realistic pilot program would not take much funding at all. Let's say we pay 1000 programmers $50,000 a year to work on open source projects. Toss in another $10 million for regulators and whatnot and that's $60 million total. Peanuts, really -- less then .01% of the U.S. federal budget.
What do you think? Is this a good idea?