Project Forking..?

 Recently I've been using a small tool which puts transparent icons upon my desktop.

 This compliments the simple window manager I use IceWm and gives me a nice working environment.

 However the program has some buggy behaviour, (like crashing with >10 icons - or when dragging icons off the screen). I've fixed these bugs, tidied up the code a lot, and added an Autoconf/Automake build system.

 Like a good user I mailled the author about each of the problems and described my fixes. No response. No updates to the code - which has only had one public release.

 So, what do I do? I'd like to package up his code + my changes/enhancements and make it publically available. Is this the right thing to do, or not? Intuitively I feel that it is, as it lets other users take advantage of the bugfixes, but it feels almost like I'm hijacking an existing project. So I'm torn.

My Project

 My streaming MP3 server now has support for plugins which I've used to reorganise a lot of the existing code. I'm hoping that people will contribute plugins - but I expect this won't happen...

 I made a new upload for the Debian project and happily there have been no new bug reports. I'm strangely disappointed. I'm not daft enough to believe it's bug free and I rely upon other people pushing the code to make them visible; either through misconfiguration or running the code in interesting ways.