On Fri, 11 Apr 2003, Rodolfo J. Paiz wrote: >> > Also, developers paid by Red Hat to work on >> > GTK, GNOME are doing important new development, not fixing bugs after >> > the final release have been shipped. >> >>So the bugs never get fixed while we get new important features. Am I the >>only >>one who doesn't see this as a good thing? > >I'll weigh in here. I see Red Hat's RHEL strategy as a wonderful thing, >primarily since making a buck or two in profit will allow them to thrive >and succeed and grow which is the only way they'll be around for a long >time. The one thing I don't like right now is that there is currently no >middle ground between RHL and RHEL. I would like to use RHEL for my office >and RHL for my home; however, there is no way I can afford RHEL for a >15-computer office. The home and enterprise markets are being served, but >the SOHO market is out on a limb. > >As far as bugs getting fixed or not, I note that there has been no >statement from RH on this, so really all we have is the theories of a few >users who may be right or wrong but are, at best, guessing. In my opinion >(FWIW), important new features are nice, but stability, robustness, and >proper functioning are CRITICAL. If there's a bug that affects me (or lots >of users) in a package that RH includes in RHL, I expect it to be fixed! >"We'll fix it for the next release in six months" is really not an >acceptable solution, so I hope that RH keeps up with past tradition and >fixes things that were broken when a distribution was released. In some ways I agree with you. Fixing a certain type of bugs is something very important. Critical security bugs are one category, however there are others also. In reality, it is not possible for Red Hat to fix every single bug in every single piece of software that ships in the distribution. Red Hat ships a _lot_ of software, and we do fix a lot of bugs in that software and contribute those fixes to the given upstream projects. We also do a lot of development in many upstream projects such as GNOME, GTK, and tonnes of other stuff. Ultimately we must strike a balance between fixing some bugs, developing software such as our config tools, and spending time doing development of software in upstream projects such as GNOME, GTK, XFree86, the kernel, nautilus, and tonnes of other projects. This means that some bugs just will not be fixed until a future release, and some other bugs may be defered by us for someone else to fix such as the upstream developers that maintain the official source code for a given project. There is not a single Linux distribution out there in existance nor any Linux company that has the manpower resources to completely and totally maintain every single piece of software in their entire distribution including debugging, troubleshooting and fixing 100% of every single bug that any user of that distribution encounters and reports. It just is not possible, because users will _always_ find more bugs and problems that there are engineers and engineer hours to invest in fixing those bugs. There will also always be a high priority bug queue or queues which will trump working on low priority things. That is just the reality of things. Users can wish/hope/pray all they want that every single bug they find will be fixed in 10 minutes, and they can even demand it and jump up and down and shout from mountaintops about it in a temper tantrum. Nonetheless the bottom line is what reality dictates: 1) No distribution out there, has enough engineers to spend countless hours dedicated to fixing every single bug reported in every single software package in the entire distribution, and that includes Red Hat. Upstream developers who maintain their software are expected to maintain their software, and while we do contribute, we dont and wont do it all. 2) If users do demand that a given distribution such as Red Hat absolutely must fix every single bug, and we were to give that guarantee you would quickly watch the distribution shrink from the current package count of approximately 1500 or so packages down to a very small package count of things we can spend countless hours on, and run and jump on every problem reported and fix it with time to spare for a game of tuxracer. The problem with #2 however, is that most of the software in the distribution that many users use every day and rely on would disappear. The distribution would be nowhere near as useful as it is bugs or not. Users would leave in flocks to other distributions that do still have the software pileup that they want, and those same users while they might complain about bugs here and there, would use the distro anyway because it has the stuff they want. We try to balance fixing as many problems as possible with new development and features. But if any user thinks we can or will fix every single bug reported and do so immediately and release erratum right away, they are greatly fooling themselves and I challenge them to find _any_ other Linux distribution or operating system that is either bug free, or fixes bugs on a dime the second they become known and instantly releases bug fixes. That distribution exists only in a fantasy world I'm afraid. -- Mike A. Harris ftp://people.redhat.com/mharris OS Systems Engineer - XFree86 maintainer - Red Hat
