On 2/9/07, Mike Frysinger <[EMAIL PROTECTED]> wrote:
forking the package is retarded. maintain backward compability and there's no reason to fork it. baselayout isnt Roy's package, it isnt my package, it isnt anyone's. it belongs to Gentoo as a whole which means changes to it affect everyone in the distribution. -mike
I totally agree with Mike here. Roy, here is what I suggest you do: 1) maintain the existing baselayout and don't change things at all. 2) start a new package called fastlayout and do whatever you wanna do with it. Be as innovative as you want to be with it. Change all the stuff in it that you want to change. Get people to test it and work on it with you. 3) once fastlayout is done and mature and is obviously better/faster/more wonderful than the existing baselayout, *then* let's talk about it becoming baselayout-ng. I have learned from experience that you never want to start a project and call it "something"-ng. Doing so takes for granted that it will be the official replacement for the existing baselayout. Thus, every proposed change is now a potentially disruptive change, and people will want to review every design decision you make - and with good reason. People don't like change unless they can see the obvious benefits of such change. And it is hard for people to see these benefits when you are just in the planning stages. That's what we call a catch-22. Ever wonder why every official "portage-ng" project would die within weeks of being launched? This is why. The pkgcore/paludis model, where neither is blessed as being the eventual successor of portage, is a model that works. "something-ng" is not a model that works for any meaningful change for Gentoo. You have every right to scratch an itch, go scratch it. But it will be slow going if you refer to this effort as the eventual successor to baselayout. Making incremental changes to baselayout with no clear roadmap is a bad idea and people will be critical of anything you propose. It's *far* better for you to work independently and create something that, when it's ready, we can all switch over to because it is clearly better, well-integrated - and done. Structured this way, "fastlayout" is certainly a project that sounds like a great idea, and would I enjoy working on in some capacity - I have some ideas about this. I also think it would be a good idea to check out what other distributions are doing in this area. -Daniel -- [email protected] mailing list
