Date: Thu, 25 Mar 1999 12:18:42 -0000 From: paul@originative.co.uk To: chuckr@mat.net, dillon@apollo.backplane.com Cc: jkh@zippy.cdrom.com, toasty@home.dragondata.com, hasty@rah.star-gate.com, dm@globalserve.net, hackers@FreeBSD.ORG Subject: RE: NFS - Will it ever be fixed? Message-ID: <A6D02246E1ABD2119F5200C0F0303D10FE7A@octopus>
next in thread | raw e-mail | index | archive | help
> -----Original Message----- > From: Chuck Robey [mailto:chuckr@mat.net] > Sent: 23 March 1999 03:25 > To: Matthew Dillon > Cc: Jordan K. Hubbard; Kevin Day; hasty@rah.star-gate.com; > dm@globalserve.net; hackers@FreeBSD.ORG > Subject: Re: NFS - Will it ever be fixed? > > > On Mon, 22 Mar 1999, Matthew Dillon wrote: > > > :3) I personally think that $25,000 would very easily be raised, > > :especially when companies using FreeBSD can make targeted donations > > :towards a goal that they would very clearly benefit from > (nfs). I think > > :the amount that could be gathered in, say, one month will probably > > :surprise everyone (I'm not trying to hold you to $25,000 > here, because I > > :think that might well be too small a guess). > > : > > :4) This wouldn't even have to be completely restricted to > nfs, but it > > :would have to be restricted to a very small, well defined > problem set, > > :of general interest. The funds collected would have to be > funnelled > > :into a fund whose goals are not general, but specific to > this effort. > > > > It isn't quite so easy. It's one thing to raise money to hire > > a programmer familiar enough with the code to be able to fix the > > problems. *Finding* that programmer is a whole > different deal, and > > if you make a mistake that's $25K down the drain. > > I'm not sure he'd be available, but are you saying that if he > had time, > and we were able to offer enough, McKusick couldn't do it? > He's not the > only one I'd have sufficient faith in (read that, if they failed, I'd > feel it was well worth the risk). I'd kinda lke to pay Heidemann to > implement stackable filesystems here. You see what I mean. > This isn't > something we could do a lot (there is a limit to how much > money could be > collected this way) but I think once or twice could be done, and with > potential terrific effect. I think both John Heidemann and Kirk > McKusick would like to do it (although I couldn't guess at > their current > calendar commitments). I don't know how much it'd cost us ... > I think one of the flaws in the thinking here is that this is somehow a "one-off" task. While NFS may be really crufty at the moment and cleaning it up may be a one-off task keeping it working is a long term commitment. As examples, if David and John had been hired to write the unified VM system how long do you think it would have stayed working after their contracts had finished? There are only two ways, in my mind, that FreeBSD can solve these problems. You have people on board who take responsibility for these tasks over the medium to long term just for kicks, which is how we currently work in the main, or you have some commercial company run with the ball and provide commercial support for a particular subsystem. This is largely starting to happen on the user side (sendmail, Sleepy Cat, etc) but as yet hasn't happened much on the kernel side. Paul. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A6D02246E1ABD2119F5200C0F0303D10FE7A>