Should we do more to incentivize open source?
jackealtman.com
jackealtman.com
This is why I'm really skeptical of things like gittip or other "we'll pay you pennies to spend hours of your life building open source projects" approaches.
I much more excited by the cultural shift that seems to be happening around "my github account is my resume". It's not just a "write open source as resume padding to get a better job" thing, it's an increase in the amount of social capital resulting from contribution to open source. Basically if people find out I contribute to a "good" open source project more people care and the people who care, care more than they used to.
I think this is an excellent incentive scheme. A popular open source project cannot generate enough revenue to pay me what my contribution is worth so it's impossible for them to incentivize me with money. But if my reputation is increased because my contribution is recognized and my peers think more of me, my ideas are more easily heard and my financial career path opportunities are expanded because of that reputation then I am highly motivated to contribute.
I think there's a big place for the patronage/sponsorship model as well (and the incentives to be a patron are quite similar), especially for getting the necessary but boring/less sexy things done. It can also push back, in one way, against the problem of reputation based incentives undervaluing boring but high-value work.
That said it's a great way to let people actually get what they want most from a project by putting something on the line to make it happen. For single person projects or small teams who are all getting funded I think it's nearly perfect. Certainly very exciting.
The only thing worrying is that it's much more likely for those funded projects to stop development when the money stops. Who knows if that will be an actual problem or not.
For larger projects, when everyone can't get paid from the pot, it's a bit tricky. I'm pretty sure that we'll see some successful projects completely implode from the internal arguments and demotivation resulting from allocating crowdfunded money.
Anyway, I could dream up fun scenarios all day but it's just predicting the future so it's probably all wrong. I think kickstarter type models are really promising for all sorts of things so people should go for it.
One thing to always keep in mind though: A human will happily do something for free for all sorts of reasons, if you start paying them they will be even happier for a while, then will normalize to just as happy as before. If you stop paying them they will never go back to being happy doing that thing for free.
The UK government, for example, has recently mandated a preference for open source: (http://www.computerweekly.com/news/2240179643/Government-man...)
Unfortunately, governments are still getting caught in proprietary solutions. Here in New-Brunswick, Canada for example, the government has just chosen an officially sponsored electronic medical record system for the province.
Unfortunately, they picked a proprietary SAAS system which means that if doctors signup, everyone's medical records will be uploaded to some private company's servers.
The company charges $24 000 per doctor upfront plus $400 a month subscription for continued access with a price guarantee for only three years. After all our medical records have been converted to their proprietary format at a huge cost (not included in the $24000), this company will be in a position to charge whatever they want.
The government gives incentives to doctors to signup to the proprietary system ($16000) and have given monopolistic exclusivity for at least three years to this system when it comes to connecting to hospital lab results.
There is actually a good open source alternative developed in other Canadian provinces. The OSCAR open source EMR is quickly becoming the most popular system in those provinces. Support plans often costs less than $100/month (it would be free if you were a geeky doctor willing to install and support it yourself). Somehow, the vast marketing and sales budgets of proprietary companies compared to the more minimal OSS sales push was able to sway our government into giving a quasi monopoly to a costly proprietary system.
No, this is false. Free as in speech vs. Free as in beer. Make all the money you can, just keep the software free (as in speech).
As far as I can tell, very, very few businesses actually make any money off writing and selling free software. They either open-source some tertiary part of a larger non-free piece of software, or they just give the software away for free and try to make money on support and consulting contracts.
I've not seen much evidence of this so far, but StormCloud, a weather applet available on Ubuntu and Chrome, is one example (more or less - I'm not sure that the "Don't be a dick public license" meets strict definitions of freedom).
But, if the choice had instead been between a $1 closed source version on the store or an open source manual compile, I think I would have taken the build/install option.
As far as what should be done, that's a lot trickier. My inclination is that it should mostly be limited to universities and government having something of a bias for open source with very liberal licenses. That's a big vague, but I'm a bit tired, so it'll have to do. I do not think the government should get involved in funding software it doesn't actually need in a big way.
One mode which does seem to work very well is Google's Summer of Code. For pretty modest investments, it seems to yield a lot of useful functionality, and because it piggybacks on existing projects & communities, the improvements seem to generally stick. I've watched the Haskell-related Summer of Codes for years now (http://www.gwern.net/Haskell%20Summer%20of%20Code) and there's been some great stuff accomplished via them.
What kind of open source should we incentivize?
For example, do we really need to incentivize ten thousands of ruby on rails testing frameworks?
I would argue that grassroot open source needs help in discovery and reaching their audience more than they need "incentivizing", and discovery means filtering.
People work on open source projects because they are driven and motivated by the challenge of creating great software that lasts, the mastery they develop while doing it, and the fulfillment and satisfaction of making a contribution to the community of fellow hackers.
But money isn't why people start open source projects, for the most part. What's their motivation? Pride? Political freedom? Helping the world? Depends on the person. Incentives need to be aligned with the individual being incentivized. But honestly, money isn't a good open source motivator. If someone wants to get rich, there are usually better ways via closed source.
Years ago, I read a biography of a prostitute and political activist. She wanted prostitution decriminalized, not legalized. She felt legalizing it would create problems. She cited Nevada as an example, where there is a fair amount of regulatory burden on the women but the pimps and johns still essentially do as they please. She just wanted the right to charge for services without risk of arrest.
I will suggest that since there is already open source, incentives are probably not a good idea. But removing barriers of some sort might do a lot.
I have thought a lot about such things over the years, not open source per se but, for example, how to make money from making information freely available. It is a challenge but can be done and has been for years (broadcast radio, broadcast tv, etc). So I would suggest you work on removing barriers rather than providing "incentives." Or at least balance the incentives piece with also removing barriers.
My 2¢.
Edit: Someone want to kindly explain the downvote? Is there some problem with suggesting a removal of barriers instead of providing incentives? Or this just a helping of lurv from one of my many anti-fans? Thx.
Also, broadcast radio, broadcast TV are not freely available, so they're poor examples for you to use. (They have licensing terms, prohibit rebroadcast, embed commercials that are not desirable with no legitimate way to remove them, you have to pay to get a DVD or watch on iTunes or Amazon, etc.)
Note to self: Folks on HN routinely think sex and prostitution references are irrelevant, not on point, not a meaningful example, ad nauseum. Maybe someday I will remember it gets perceived that way, even though I think we are all adults and it should be nbd.
But thanks for the feedback. Have an upvote.
My main interest is in the general model of making info/entertainment available to an audience for "free" while still finding a way to monetize it. Broadcast TV and radio have done that for a long time and there are some websites that exist which do much the same. There are no membership fees. You can read/participate for free. Yet the site owner somehow makes money. Some take donations. Some use advertising to pay the bills. Some sell merchandise. (And one invests in startups, which is way more complicated than my modest goals, so not a model I am looking to emulate but it does exist.)
Perhaps I am merely a fool, tilting at windmills. But that is the general model I am shooting for. The details on how that works varies, but that model has been around for decades, both before the Internet and, in recent years, on the Internet. It can be done. It is being done. I have yet to make it work, but it is a viable model.
A company needs an OSS feature so they hire one of the devs to work on the software, full-time.