Date: Sat, 12 Apr 2003 15:47:06 +0930 From: Malcolm Kay <Malcolm.Kay@internode.on.net> To: Brooks Davis <brooks@one-eyed-alien.net>, ian j hart <ianjhart@ntlworld.com> Cc: stable@freebsd.org Subject: Re: 4.8-Release disk3 and disk4 Message-ID: <200304121547.06679.Malcolm.Kay@internode.on.net> In-Reply-To: <20030411215257.GA23072@Odin.AC.HMC.Edu> References: <20030407194038.GA18372@qcislands.net> <200304112003.04157.ianjhart@ntlworld.com> <20030411215257.GA23072@Odin.AC.HMC.Edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 12 Apr 2003 07:22, Brooks Davis wrote: > On Fri, Apr 11, 2003 at 08:03:04PM +0100, ian j hart wrote: > > What exactly is "the effort involved". The packages are already built= , > > right? > > Solving a more complicated[0] form of the nice little NP-complete > problem known as bin packing. > > -- Brooks > > [0] This variant adds a dependency graph problem so how large a package > is, depends on what other packages are already there. It also requires > that you assign values to each package to determine which ones have the > highest priority since you can't fit them all on anything short of a > dual layer DVD (and I don't expect that to hold much longer). I'd be > fairly suprised if you could find two people who gave the ranked > ordering of the importance of the seven thousand plus ports. With the increasing prevalence of broadband internet connections the loss of packages from the distribution is of reduced importance. On the other hand the normal hard disk capacity has increased enormously so it is usually feasible to transfer the all packages from a distributio= n to=20 harddisk, where upon the dependency tree problem, I believe, largely=20 disappears. Perhaps unsorted distributions would be a better compromise. Malcolm Kay
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200304121547.06679.Malcolm.Kay>