Recently we've had a fair number of "big" projects proposed, both on the -core list and on the -dev list, including herds, udder, refinement of -stable and -stats (now subsumed by herds?), uid/gid handling, new types of init scripts, hardened gentoo, using "delta"s when upgrading gentoo packages, and probably a number of other things that have dropped off the scope. Unfortunately, we have no single repository of these proposals, which means that it's very easy for both good and bad ideas to be proposed and then quickly forgotten. My proposal is that we adopt Python's concept of Python Enhancement Proposals (http://www.python.org/peps/), which serves as Python's system for handling large-scale proposals. (Incidentally, Gnome has also adopted this idea, calling them Gnome Enhancement Proposals: http://developer.gnome.org/gep/list.html.) I've made a start. Take a look at http://www.gentoo.org/proj/en/glep, and you'll see two informational GLEPs already present. (If you want to make your own GLEP proposal, see the README.txt file in http://cvs.gentoo.org/cgi-bin/viewcvs.cgi/gentoo/xml/htdocs/proj/en/glep/ as well as GLEP 2. Incidentally, the reason for choosing ReST, barely-marked-up text, as the base format is to make proposals simple to write without too much worry about the formatting. It should be relatively simple to convert ReST to our guide-xml, if desired.) Here's the comments that have been made so far about this GLEP proposal from other developers: * Numerous votes in favor. * It's been suggested out that GLEPs will solve the problem of a moderator on the mailing lists to summarize where current proposals are, as that's now the job of the GLEP authors, help increase signal-to-noise on -dev, and help consolidate pros and cons. * Existing Gentoo devs are willing to help with infrastructure and creating any needed new tools. * One comment that perhaps it would make more sense for people to develop GLEPs (until ratified) on the new gentoo development wiki. I'm biased, but I prefer a more formal proposal submission method because I think people are more likely to spend time polishing it before submitting it. Now it's time to open up this idea to the general Gentoo Linux community. Thoughts? Comments? Suggestions? -- Grant Goodyear