Hi! El jue, 29-04-2004 a las 14:48, Carlos Garnacho escribió: > Hi Tiago, > > As I told you in IRC, while it might be a sensible solution and that all > distros should unify in a single configuration schema, I don't think at > all that we should give the first step here, my main reasons are: > > - distro specific tools & apps will rely in the distros' config files, > causing data duplication, misinformation, errors, user headaches, etc... > > - we still should parse config files (at least for creating the initial > home-made config file), and at some point, we still should rewrite the > config files, for data consistency > > - this is only valid in the network config issue, IMHO the GST look for > a broader point of view for config stuff > > - This would be more a maintenance pain than a bless for the gst > developer(s?) > > Of course, these problems should vanish if the distros agree on a common > config file, but I expect more to see cows in the moon than distros > agreeing on something like this :) > I think it is very very difficult for distributions to change its own way to do things. Maybe from some group like Linux Standar Base we can push distros to use similar config files, but I find it at the present very very improbable and something like GNOME System Tools can't depend onto. Cheers > Regards > > El mié, 28-04-2004 a las 01:55, +0000, Tiago Cogumbreiro escribió: > > I would like to propose to the list an effort to try to create a XML > > specification file that would provide the network configuration details. > > > > Why? > > Unix has a know problem, a miss of standard. Because of the openness of > > Open Source each of us tries to create the best solution for a given > > problem, while this is great for some things it is not that good for > > other, because it may lead to fragmentation and disorder, therefore > > losing effectiveness. > > > > GST, and you should know this more than i do, suffers alot from the > > "each distribution has it's own spec" effect, in what concerns to > > network configuration (from now on netconf) each distribution has it's > > location and syntax on debian it's a file in /etc/network/interfaces in > > redhat it's located under /etc/sysconfig/networkm under slackware in > > rc.network, it is an extraneous operation to add and keep up to date > > support on each of this platforms. > > > > The solution to this problem, IMHO, is not that difficult. Why not > > provide a new spec for this type of file, and not a spec based in bash > > or any other syntax, we need something technically advanced such as XML, > > it gives us all the benefits that all of you must already be aware of. > > Some might say that it won't solve all problems, and that's true, one > > thing that has to be created is a service that should be called uppon > > system boot that would load the configuration file, but i don't think > > that this is harder then supporting all configuration files available on > > every distribution (and operating system) that GNOME might used on. > > > > We want an universal configuration tool, why not depend on a spec that > > is more accessible and open? > > > > With one configuration file it is up to the back-end to take care of how > > it is implemented, whether it's on Linux or on FreeBSD. This way the > > work of back-ends is focused on configuration and not parsing data. > > > > This case study was made on netconf, i leave it to you to think about > > the following: Is it possible to have other tasks GST currently does > > with configuration files? > > My guesses are that in some cases, as services, it isn't possible, the > > only viable solution is the LSB kind of approach. > > > > Tiago Cogumbreiro > > > > _______________________________________________ > > system-tools-list mailing list > > system-tools-list gnome org > > http://lists.gnome.org/mailman/listinfo/system-tools-list > > > _______________________________________________ > system-tools-list mailing list > system-tools-list gnome org > http://lists.gnome.org/mailman/listinfo/system-tools-list >
Attachment:
signature.asc
Description: Esta parte del mensaje =?ISO-8859-1?Q?est=E1?= firmada digitalmente