Open source is like this…
I was working on a plug-in for Eclipse (open source), which was using Maven (open source) under the hood. And I thought, well, there’s already a plug-in that embeds maven into eclipse, so I could use that.
That’s open source. A free plug-in that allows redistribution, less work for me, better experience for everyone.
So I install the plug-in, state it as a dependency and went on happily writing code. Until I notice that my plug-in can’t use the other plugin’s classes. Weird. Really.
But open source is also like that (with exceptions). Made by volunteers who don’t have much more time than me, and that just keep going with limited testing and a reduced scope (there’s no time for more).
Probably no one else had tried to use the plug-in in the same was as me… chances are that the author forgot to export some packages. But it is open source, right? I could easily fix that.
Now I get the code… and everything works. No fixing needed, the latest version already has been modified. But not released.
That also is open source. Developers need some time to package a release, and time is a finite resource.
So I’ll just go along with the bleeding edge version from the repository. After all, I guess that I won’t be releasing anything before they do. If just time was open source, I could hack something…
It’s sometimes perplexing how many tools we’ve grown to depend on which are free and open source. Just for a little task I’ve used a handful of them. Not counting engine powering the browser I’m using, or a great deal of the Mac OS X it is running on.