On Wed, 2003-04-02 at 23:14, T. Ribbrock wrote: > On Wed, Apr 02, 2003 at 04:13:35PM -0800, Cliff Wells wrote: > > Sure. But have you ever downloaded a vanilla Linux kernel and run RH on > > it? Works just fine, sometimes better. If you are concerned about RH's > > QA, then just do that. > > Kind of defeats the purpose of buying a distro, doesn't it? :-}
Not at all. That's like saying that installing a newer version of apache or samba from a tarball defeats the purpose. The purpose of a distro is so that you can get a bootable machine with most everything working without a big hassle. It doesn't mean leaving it in 'pristine' condition. The difference between starting with a distro and rolling a few of your own packages (including the kernel) and starting from scratch is remarkable. Take a look at the directions for one of the BYO linux "distros". While it would certainly be.. er, a learning experience, I think I'd rather hit myself in the head with a 40+ floppy install of SCO Openswerver 3.0. Okay, maybe not, but you see my point ;) > Over the past two years, RH lost quite a lot of that appeal. In my > eyes, they cared less and less for the "ordinary" user - the folks > they started off with. It's little things, like bad support for the > mailing lists, changes in release policies, etc. > I myself am in two minds about this: On the one hand, I understand > that they're a commercial entitiy, which needs to survive. On the > other hand, I'm missing the "idealism" (for lack of better word) from > the early times. Granted. But this is probably a stage nearly every company goes through as they go from small to large, enter new markets, get a broader base of customers, etc. It isn't uncommon for early adopters to feel alienated as the tides change. I just don't think you'll find it much different elsewhere. I myself have cursed RH at various times (and will again, no doubt, probably before lunch), but for me the their idealism hasn't changed in one important aspect: they give back to the community. rpm can screw me, they can f* up my python install, and all that, but I can't deny their commitment to the community (the Linux development community, that is). To me that far outweighs any occasional packaging snafu on their part. > What I'm driving at: Making such statements publicly is one way of > getting a discussion going - and maybe even a way to influence things. > Yes, I know about the "vote with your Euros" thing, but that's not the > same. I'll grant this as well. I was just being snippy. > [...] > > Debian (which is what I'd recommend, if you want stability at the > > expense of new features, ...) > > Well, RH's balance in that regard was perfect for me in the past, > that's what I liked about them. And I'm not quite ready for the Debian > bigots... <donning asbestos underware> :-) Ah, some of them are nice people, if a bit misguided <puts one leg in> > (it's ironic anyway, as I'm running OpenBSD as well...) Speaking of bigots... <puts other leg in and hopes a ;) stops the flames> > I'm not going to argue there - you do have a point. I myself was > waiting for 8.2... ;-) So now you'll wait for 10? <wink> > There's nothing "magic" about it. x.0 *was* a beta release for me (and > for others). Those were the releases were the new stuff hit public > scrutiny and the biggest bugs and problems were shaken out. It's been > that way for at least the time I'm using RH. Yes, there is a beta > program. Yes, those bugs *should* be found during that time. In > reality, that's not the way it works - the x.0 will always have more > people trying it out (and hence more eyes watching it) than any "real" > beta release before it and I honestly cannot see a way to improve > this - nor is RH the only company suffering from this problem. RH just > was honest enough to admit the faults in the first releases and follow > it up with improved, compatible releases. I'd chalk it up to protecting their reputation rather than honesty. I don't think they'll change much in that regard. They've always made it clear that they consider their name to be their most valuable asset (which is why they've never tried the Caldera tactics of not GPL'ing portions of their code). I seriously doubt they'll let their name get drug down with half-ass releases. I admit that there might be some hitches to their new scheme that either 1) they haven't worked out or 2) they haven't told us of their solution to. Either way, I'm willing to give them a shot at it. > > Off-the-shelf software is rapidly becoming a > > thing of the past anyway. Why kill a tree (and a couple of plastic > > things) to put *software* in a pretty box in the store when it will need > > to be upgraded before the shrink-wrap has hardened? We are moving into > > the age of broadband (or are already there, if you ask me) and RH > > certainly must recognize this. > > "the age of broadband" is still a some way off. It's still far too > expensive and in fact not reachable for many (think outside USA, or > even outside the "Western World" (for lack of a better word). Even in > Europe, there are still many, many people dialling in and many are > still paying by the minute for it. Never mind lesser developped > countries. Actually, from reports I've seen, broadband in the US lags behind some European countries (although I agree that it certainly isn't ubiquitous by any stretch). > > Perhaps a middle-of-the-road solution would be for RH to make occasional > > updated releases available via their website (mail-you-a-CD type) for > > people who don't have the bandwith for a major download. Proposing this > > (or some other solution) to RH is far more constructive than stamping > > your feet ;) > > Point taken. Your suggestion above is actually an interesting one! Maybe we can lobby them. Even if they don't like that idea, perhaps they'll address it some other way. > [...] > > So you wget the rpms and rpm -hUv them? How is that different than what > > the automated updates do? I know a lot of people do just this (in fact, > > I used to be one of them), but it doesn't really make a lot of sense, > > IMHO. If you are rebuilding src rpms or from a tarball, that's a bit > > different, but when you get the binary rpm (and all of its dependencies) > > you are, in fact, using an automated update system, only you're doing it > > the wrong way <wink> > > Ok, admittedly, part of it is lazyness. wget to the fileserver, then > updating the machines I want to update (four at the moment) works so > well, that I had no reason to use any other system. New system, new > bugs, simple as that. ;-) This is a common problem with the introduction of new tools (and I'm certainly not immune). I *know* I could be a lot more productive if I'd commit myself to learning certain key tools, but often as not, I don't just because I don't want to (or can't) invest the time to learn them. I know the issue with automated updates isn't really a matter of learning, it's rather a matter of trust (i.e. "wtf is this program doing to my computer?"), but it's the same sort of problem with the same end results. However, if you're interested in anecdotal evidence, I've been using apt on all my RH machines for several months (7.2, 7.3 and 8.0) now and have only hit a couple of snags (which I attribute to problems with rpm itself rather than apt). > Nonetheless, I'll wrap the discussion up at this point. I've enjoyed > the exchange (sometimes I think people who are able to have a good > discussion are getting rare these days, but maybe I'm just getting > old...), and you've certainly convinced me that not everything is as > black as I've made it out. One some other points we'll probably have > to agree to disagree - but there's nothing wrong with that, either, > IMO. Likewise. Regards, Cliff -- Cliff Wells, Software Engineer Logiplex Corporation (www.logiplex.net) (503) 978-6726 x308 (800) 735-0555 x308 -- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-list
