IPv6 is site specific, but it has been important since April 24th of this year to support IPv6, since that was the date that Cisco released code to support it in all their supported routers. But right now, we all know that widescale acceptance of IPv6 is going to have to come in at the client level, with Microsoft driving the deployment process.

It wouldn't hurt if someone were to build a highly efficient NAT box for IPv6<->IPv4, so that once Microsoft CDROMs could be pressed by AT&T @HOME or some other severable network provider, that IPv6 deployment could go forward a large chunk of the net at a time, instead of being an all-or-nothing crash-fest. Take this as a hint to the IPv6 advocates in the audience that they need to do something.

NFS is also important. NFS is hard to load as a driver, and keep the LEASE code working. NFSv4 is looming on the horizon, and it appears to finally fix locking, for once and for all, for non-coelescing clients, as well as for stacking VFS layers including an NFS VFS somewhere in the stack. There is some possibility that it will actually be useful to Windows systems.

I think it's time to look at supporting "drivers" floppies, and the pain in having no floppies is certainly incentive for someone to do the work, should that become stated policy of the project to support most things through driver floppies that are loaded post-boot.

To Unsubscribe: send mail to majo...@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message