Dear readers, please help me brainstorm here. What instances can you think of where a business has released semi-mature/mature software under an open source license, where previously it was proprietary? AND the business managed to foster a community of developers around it, while their developers continued to work on it?
There are a few examples where companies have “dumped” software under an open source license where they no longer cared to look after it themselves (acquired software, or just discontinued products). And there would be some examples where a company has began development of a package as open source from the very start (Laconica, Drizzle?). But what about the middle case?
The only examples I can think of are
- OpenOffice.org by Sun
- the programming language Java also by Sun
- Launchpad by Canonical – not yet complete
So, regarding the problem of how you get a developer community interested, Java and Launchpad seem to have it lucky, as their users are typically programmers. OO.org has a more general user base, but OTOH they also had a well-defined and much-loathed proprietary enemy, to galvanise developers into contributing. What if your userbase is not technical, but you don’t have a clearly defined enemy in this way?
And moreover, what if you are not a gigantic company that already has some cred in the open source world? What if you are a small business? How do you demonstrate your good will and get peopled interested?
And is that case any different to the case of individuals wanting to increase awareness and development of their pet open source project?
AOL/Netscape released Mozilla (11 years and one day ago now — http://weblogs.mozillazine.org/gerv/archives/2009/03/happy_birthday_mozilla.html )
They certainly intended to continue development within the next 3-4 years before they spun it off into a separate non-profit organisation…
— Chris Neugebauer · Apr 1, 06:44 · #
Yeah I was about to say Mozilla has got to be the prime example.
Zope (well Zope 2) was originally the closed source Principia application server.
HP’s Tesseract is sort of an example: they don’t seem to be supporting a community around it, but Google has taken over development as open source. However, see below about not actually having a community!
Wikipedia’s “Commercial open source applications” probably has more examples for you but I’m not sure.
You’re created something of a false dichotomy between “successful continuing project with community contributions” and “abandoned project available under Open Source licence”. The licence may be open source and the project ongoing but the contributions (and sometimes the process) may remain very much in-house. Sun products have been accused of this, I don’t know with what level of accuracy. Launchpad will be undecidable on this criteria for at least a year after release and possibly longer.
There’s also the common “there’s a more featureful proprietary version, which is by definition developed in-house”, there’s examples at Wikipedia’s “Business models for open source software”. Ghostscript used to be under this model too.
There’s also Open Source projects that were adopted by a company (sometimes by hiring the sole developer). IronPython is an example of this, its very early development was community and then it became a Microsoft project.
The *BSDs are another weird example: a community couldn’t form around them until various commercial interests stopped legal action against each other.
— Mary · Apr 1, 08:17 · #
Blender kind of counts. The company built to market it went out of business, and the lead developer then started a non-profit group specifically to release and foster it as an open source product.
— Peter · Apr 1, 08:31 · #
A few more examples: many of iD software’s/John Carmack’s commercial game engines have been released as Open Source after some time, although I don’t know how much time the company spends in the community. I don’t know quite where that fits in your hierachy.
And Nokia is intending an Open Source release of Symbian, I believe.
— Mary · Apr 1, 11:22 · #
Regarding how to get people interested if you’re a small company, well, I guess:
1. developing a user community and paying proper attention to it: one example would be hiring a community manager and letting her/him actually do the advertised job rather than being sucked into yet another dev role.
2. consider carefully some kind of nicely documented and featureful API allowing the development of extra features and their distribution as plugins. This seems to be something of a gateway drug, people will end up poking around in the main codebase to see why their plugin won’t work, and in the meantime they get to develop features and own them rather than having to fight with other developers about core features right up front. See Slim Devices here, who seem to have had perhaps 5–10 devs prior to their acquisition by Logitech.
Also, it’s probably just not going to work out hugely well if you can’t build the bottom layer of your community pyramid: passive users. If your product has very few users I wouldn’t think building an open source community would be easy.
— Mary · Apr 1, 11:46 · #
@Peter: As the business went bust, Blender isn’t a great example for my purposes, but thanks. :) I wasn’t aware the lead dev stuck around, I thought it was a case of end of life “dumping” actually.
@Mary: thanks for all your examples. I will look into them some more. Re: the possiblity of open source lip service (development continues mainly in-house), that occurred to me but it also seems something to avoid, not really a genuine commitment to open source.
Having an open API as a “gateway” is something that did occur to me too. But I guess this is also slow going as API users may often not congregate around the product (eg join the mailing list) – after all the whole point of the API is to “get it out there” rather than draw people in. Kind of distributed community, if you know what I mean.
If your product has very few users I wouldn’t think building an open source community would be easy. Yes, well, this may be the cold harsh truth.
— pfctdayelise · Apr 1, 18:58 · #
Given how tightly Sun holds the reigns on OOo, and the way that has caused the creation of Go-OOo.org to pop up to accept the patches Sun won’t…I wouldn’t really put them on the list of companies that did well at that.
— Mackenzie · Apr 2, 05:58 · #
One project I regularly use as an example, which has not been mentioned here already is Eclipse.
In many ways, despite being somewhat overburdened with bureaucracy (everyone who borrows too much from Apache suffers from this), IBM chose a brilliant strategy for Eclipse: taking the bits they couldn’t monetise, spinning it out into a separate organisation, and allowing it to create a new ecosystem (which, meanwhile, happens to support and promote mostly things that IBM likes, and competes in markets that IBM would like to mix up a bit).
— Jeff Waugh · Apr 6, 04:05 · #