On Wed, Nov 21, 2007 at 04:04:50PM -0900, Jeff Spaleta wrote: > Because for other, established, usage cases that have seen significant > discussion, having a lot of stuff you don't need installed has costs > in terms of security and bandwidth consumption when doing things like > updates. Hmmm. Not much of what is in there is servers or reuires any kind of priviledges though. As for bandwidth, that's rather cheap too as long as it's not upload. I understand it can be an issue for mirrors though. > What you are doing is firmly outside of the usage cases for > which any coherent plan has been even described let along attempted to > be implemented as part of this project. A subproject to make > everything or nearly everything installs work from a kickstart file or > from a specialally designed spin doesn't sound like something that > could not be done as part of F9 release prep. But there has to be > people willing to drive this forward. Are you one of these people? > You clearly care about the issue, but are you prepared to step up and > chart a course to see improvement in a family of usage cases that > matter to you? Well, I need Core back. Not Core as in "my package is holier than yours", that we can do without, but core as in "all these packages work well together, can all be installed together, and you will be able to do mostly anything with them". Jesse Keating told last june on the list, after yet another of my bitchings, that the "Fedora" spin, aka the packages on the DVD, were supposed to replace that. The F8 release shows it's not the case anymore. I'm not sure how technically and socially a core-like package selection could be organized. Any experience in that area? > > Disk is cheap, time isn't. > This statement goes against the fundamental axioms by which graduate > students come into being. > You just need more graduate students. Actually, graduates students are getting expensive in this part of the planet. Science is not that interesting anymore as a studies, it seems. > -jef"is anyone working on a Fedora spin meant for cluster usage already?"spaleta Oscar does. It kinda sucks though, because they forgot that disk is cheap, and you can't predict what is going to be used on the cluster. Sometimes the job includes command-line php scripts. Sometimes the end stages include some scripted rendering though TeX or gnuplot. Things like that. So instead of using their spin I just slightly patch my standard "everything goes" one. OG. -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
