Bruce Schultz <brulzki <at> gmail.com> writes:
> (This has ended up hard to read; I hope it's not my tablet that's messed > up the message threading, but apologies in advance if it is) Nope. I use gmane as a front end and it is aggressive in trying to keep posts small. It also rambles a bit so here is my condensed sentiments:: The install process can be broken down to (2) main parts (imho):: (1) Low level hardware, file systems and the mimimum of configs to get a working reboot-able (withoout the installation medium) installation. My "base-install" is a btrfs, dual disk raid-1 base system, with fstab, efi gpt and grup-2 issues solved, configured and able to be "studied" as to how configuration was accomplished. If it can be done without mdadm and lvm the that's even better. For this I am willing to *pay* the right person to develop this open source software. Note a base system is probably going to be smaller than a 'default profile system', but it's the same idea, except from it you can build up a myriad of targets. This would essentially consist of the hard drive setup including grub2, gpt, btrfs, raid1, fstab/mtab, mbr-or-efi type of issuse. Menu drive would be keen or a gui installer (YaST). (2) Decide if the device is embedded, tablet, laptop, workstation, cluster, server, firewall or whatever target. Set up the configs and install the appropriate packages that complement the chosen system profile. This part could easily remain manual. It would be an exercise for the noob/user/expert to develop a set of ansible (or whatever) rules to control the build out of the target system. It would continue to be the 'noob filter' that so many experienced gentoo folks seem to like. The self development of automation for multiple system, beyond the functional core (baseline install) is the responsibility of the gentoo user to develop. So it would filter out those 'undesirables' both devs and long term gentoo members seem to abhor. (Not my issue either way). Other key points:: Read up on stage 4 installs as that work is very doable and is being pursued and has the blessing of the dev community, so it is most likely to happen. YaST, ported to GENTOO would be a warmly received project, imho. Better still we have part of yast in gentoo already, just 'eix libyui'. [1,2.3] Besides opensuse others use Calculate Linux or Sabayon Linux then just convert to gentoo:: hth, James [1] http://libyui.sourceforge.net/ [2] https://github.com/libyui/libyui [3] http://michal.hrusecky.net/2011/08/libyui-in-gentoo/
