So you’ve heard about this great piece of open source software and you are considering either contributing to it or using it a very serious way, maybe even a business critical way. What are some of the questions you should be asking about the project?
- Is it good code and is it well architected?
- Who are the founders, contributors, and users?
- What are the motivations and behavior of each?
- What is the form and governance of the community?
- Is there a single dominant player that is controlling the direction or is it a more democratic community?
- Are there intellectual property issues involving copyrights or code provenance?
- What about that license?
You need the confidence that the code and the community that supports it is fit for your purpose, are sufficiently stable, and have no “gotchas.” That is, there should be no surprises around the intellectual property involved and no unexpected strangeness around the community and its leadership.
What other things do you look for and what other advice you would give along these lines?
The suggestions here are good, but *before* I do any of that I look to see if there’s *any* documentation, and if it looks like I can get things done with it, because most projects have no docs, or the docs they have are outdated or missing coverage of key features. No amount of community support makes up for having no documentation. You don’t have to be on par with Django’s all-singing-all-dancing docs, but it needs to be useful and insure that I’m not having to dig through code to get basic things working.
I guess I find a quick documentation review to be less work than figuring out if the code is well-architected. It can also be just as telling.
i tend to skim the bug reports and the forums to see what type of activity is happening – who’s responding and timeframes. I just like to get the pulse on the community.
Anymore, I feel like if the Apache license is involved, it might protect my interests a bit more…I tend to favor that licensing- especially if I think I might make my own modifications down the road and plan on distributing it or using it in a project.
As a sysadmin, my concern with some random piece of open source is how good the software is.
* Cross-platform = robust. An open source project with ports to every Unix and Unix-like there is and possibly a Windows port too, is likely to be of robust internal construction. One that works on Red Hat *and* Fedora … not so much.
That’s really the biggest consistent indicator of quality I’ve seen. There are exceptions at either end, but for a random small library or whatever that you’ve never heard of that may be just the thing … this is a good one.
Oh, and:
* If the last update was in 2006, that’s fine if the web page is straight un-CSSed HTML.