1. dump the code into a zip and release it under a free licence
2. step 1 + answer to emails
3a. step 2 + provide a hangout for interested developers
3b. step 2 + actively work on the code base
4. steps 3a and 3b
5. step 4 + actively sell the project to devs
For example, id Software is on level 1 (or 2?) by releasing old engine sources. That is already much better than throwing the code away. Most open source projects themself are only at 3b, because they are hobby projects and don't care about users.
iOS currently makes for a good example, because all it takes for some old apps to use the Retina display is to recompile the very same source with newer dev tools. Sometimes it just takes that, or a small fixes to make software run on a new OS.
I'd pay more for software if the authors promised to open-source it the moment they drop support.
( Of course, the App Store has a problem similar to Ruby Gems - it is hard to transfer control of the main repository.)
When the announcement that Copycopter was to be shut down both Crowdtap and Iora Health reached out to Thoughtbot about keeping the service alive. Dan Croak facilitated communication between us and began the thankless job of putting the application in a state that we could deploy and maintain the codebase.
I believe this transition was a ton more than just opening a repository on Github. Dan engaged interested members of the community and has helped us get to the point in which we can carry the service forward. It was handled very well from my point of view.
If there was interest by some other developers, there was always some new open source project/community building itself up around the released code. And if not immediately, maybe a few years later.
E.g., some examples (where I was partly also involved in):
* OpenLieroX http://openlierox.net
* ioquake http://ioquake3.org/
* many projects here: http://icculus.org/projects/
* iodoom3: http://www.iodoom.org/ http://git.iodoom.org/ (still not much content there)
* https://github.com/id-Software
* Commander Genius: http://clonekeenplus.sourceforge.net/
* https://github.com/LaPingvino/The-Puzzle-Pits
...
oh wow, I worked on puzzle pits for a while trying to get it to compile.
I ended up figuring out that the original binary ran perfectly fine in dosbox (which I played for a bit) and that the original code was pretty much useless and it would be easier to do a complete rewrite of the game to get it to work with a modern system. Not to mention this doesn't even use a recent standard version of C. So I ended up losing all motivation because of this and because of being a 1 man team.
It was a fun project though, it was this project that introduced me to the hacks that made dos games possible :)
Being open-sourced doesn't prevent any codebase turning into some kind of abandonware. Certainly for me I have used a lot of software that turned out to be abandoned no matter they are open-sourced or not. But here is the problem, if it's a service, it's worse than abandoned. You simply can't use it. If it is just some desktop software, I can stick with an ancient version until I fill sick maintaining the running environment. For a service, that is it.
So to my own end, open-sourcing it, is at least, better than simply vaporize the service (which many of Google's services have been like that for the past months.)
For other ideas, maybe a new team will actually pick up the idea and continue to work on it as a company. I am currently working a startup project and we found a company, that was working on the same idea for 3 years and ultimately gave up. We emailed them and are hoping that they can share some lessons, but imagine we they open sourced their code. That might have really given us a jump start. But then again, as a developer I prefer writing by own code from scratch, to working on someone else's (by definition) crappy (most likely) PHP code base.
This is a worrying trend. Most of the projects on github have become 'orphans'. I see no docs and worse - no activity for years even on some pretty interesting ones. Contrast that to Apache projects and their communities.
@patricksroberts I think @nirvana meant this in a more general sense.
It really depends on what you're making. If the software is popular, there's little need to help form the community, facilitate the process, and so on. It might help, but just dumping the sources online is way better than nothing.
Sure, the developer can decide to release the code as open source. That's fine. But to make an open source project requires a lot more commitment than that[1].
An open source project is kinda like a startup. You need your product (the code) you need your marketing (letting people know the code exists) you need your customer development (is the code useful for a lot of developers?) you need sales (dude, you really want to help with this code!) you need passionate early adopters (dude! you actually wrote unit tests for us? Tubular!) etc.
Compared to all that, putting the code up on github is the easiest part.
Really, starting an open source project is about building a community of people. Not people who use the code, but people who care enough about the code to maintain it and contribute to it.
If you don't do that, you've still got abandonware, its just that the code was open sourced before disappearing, never to be seen again.
Spending this effort to build a community may not be viable for projects being shut down- especially if they are failures. The people behind them may be burned out or soon to be unemployed or soon to be handcuffed by NDAs.
Maybe the open source community could use its own equivalent of the ASPCA-- the American Society for the Prevention of Cruelty to Abandoned Projects.
They could take them in, like pets, give them a home for a few months or maybe a year. IF the original developers are willing they can provide as much or as little support as they are emotionally or physically capable of. After a year, if the project has attracted enough interest it is spun out as its own open source project, and if not, it is "euthanized" by being marked as shut down and put in cold storage.
This, at the very least, would serve to bring light to these projects that could use some love. And since the organization would only handle a limited number of projects at a time, they could provide life support during that transitionary year, without burning themselves out trying to save every abandoned project, lost and starving on the side of the road.
[1] I speak theoretically, I'm in the process of making my first open source project and have been thinking about this a lot lately. My previous open source experience has been as a contributor.
Edit to add: this isn't meant as a commentary on Copycopter, just general thoughts on the topic of "don't shut it down, open source it."