Date: Thu, 19 Mar 1998 18:48:25 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: "Lee Crites (AEI)" <leec@adam.adonai.net> Cc: hackers@FreeBSD.ORG Subject: Re: SCO (was Re: hi terry) Message-ID: <20031.890362105@time.cdrom.com> In-Reply-To: Your message of "Thu, 19 Mar 1998 20:08:49 CST." <Pine.BSF.3.95.980319193539.19449D-100000@adam.adonai.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> First off, I don't know where the "wishlist" is, if, indeed, it > exists as an entity which can be "seen." I really haven't done a Not really, it's more of a collective conciousness thing - the sum total of all the various "gee, wouldn't it be neat if..." postings that go through -hackers and -current. > Second, I don't think anyone will just present something as > complicated as clustering to the group and say: "clustering is > cool, please check out my FreeBSD implementation!" There has to > be a lot of groundwork laid and a fair amount of emphasis from > the top down before any person (or group of people) would take on > something of that size. Now I think I understand what Jordan was Fair enough, but I'm also not sure that there's anyone "at the top" who's even done this kind of work so you're not likely to get any genuine leadership there. We're a pretty sharp bunch of guys as they go, but our talents don't extend into every domain of computer science. :) If someone wanted to take on a leadership role with something as substantive as clustering, I'd say they'd simply be given the reins and told to go for it, much as John Fieber was with the Web pages or John Birrell with the Alpha port. You don't need to be a member of core to lead a substantive effort in FreeBSD - far from it. > Third, the potential list of "somebodies" who could really "do > the job" is fairly limited. While I do have a copy of the entire > tree available to me, it is really those of the core team which > are doing the bulk of the development. Somewhat over a year ago The bulk of the development in certain areas, yes, but see above for at least one of those areas for which core really can't be relied upon for leadership. > I sent out my second message offering to help on a project. It > was literally the first time I had offered free help and had the > offer turned down. Jordan responded to my message of dismay I'm not sure it was quite that cut-and-dried, though my memory on this one fades. I think it was more a case of you wishing to help and us having nobody lined up *for* you to help, so we had to simply say "uh, there's no existing team for you to play on, sorry!" > What someone like me would have to do is develop something, lock, > stock, and barrel, then present it -- as an outsider -- to > someone on the inside, which I will have to convince to take it > on, in order to get it approved and integrated. I'll do that for > something rather small. Perhaps a driver or utility. But, quite > frankly, I won't put in the time to do anything close to clusters > without some significant support from the "inside." I think all of us would like to see clustering, just as we'd all like to see FreeBSD running on the ALPHA or the SPARC. However, just as is the case in those two examples, we lack the manpower physically within core to do anything substantive about it and have to rely on outside leadership to take things forward. If you came up with a really decent clustering implementation which didn't bodge the heck out of FreeBSD then I think there'd be no question about it being brought in. If you made a real hash out of it then, on the other hand, I daresay you'd be first among those asking that it be left out of the tree until a superior solution could be implemented. It really comes down to quality of work, and (for the most part) this is no different than things are done under Linux. Somebody writes something that the users really like and sticks it on an FTP site someplace and, given a little time, it becomes canon. If the users turn their noses up at it, on the other hand, then the choice is also pretty clear. > So we're back to my original comment: those people who control > the direction of FreeBSD need to take a step back and see where > the industry is moving and see where they want it to be. We'd like FreeBSD to be many places and many things. The issue still remains, however, a far more pragmatic assessment of where we CAN be. It does no good to wish for pie-in-the-sky features that have little chance of happening when we could be spending our time in pursuit of far more tangible goals. Jordan 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?20031.890362105>