Mike Frysinger wrote: > people need to bring up their outstanding issues now and get them > addressed Hello all, I'm currently a Gentoo "power user" and I'm being mentored by kloeri to become a developer. During this time, I've had the chance to study accurately the Gentoo developer's handbook and from what I can understand, respecting the Etiquette policy and being polite to other developers and users is part of the Quality Assurance of the Gentoo project. I understand that some matters are particularly "hot" to discuss, but please help this project be different from other OS projects that have a high quality code but miss to treat their users and developers adequately and don't make it confortable to interact with: it's really part of the user's experience and if we can achieve to keep this debate from degenerating into a flame, maybe more people will try to step in and help find a solution. Furthermore, if suspending the project with a definite timeline to reconsider the next steps may help to discuss it with more serenity, I encourage the Council to do so. That said, I'd like to be proactive and share with you my experience about OpenBSD, another project that has a "tree" of unsupported packages compiled from source through a series of scripts (ports). In OpenBSD, ports tree is clearly claimed as unsupported. In my opinion, the percentage of users that miss to understand that ports do not receive the same QA as the core system ranges between 5 and 15%. You may also want to consider that OpenBSD core system is quite limited in terms of quantity and youth of softwares included, but it's not simply a collection of packages: instead their target is to assure the quality of all the code, not just the packaging system. Furthermore, ports are maintained partly by the OpenBSD team and partly by the developers of the ported software. This aspect is interesting: if the developer of a tool is given the possibility to maintain their port or packaging scripts for various Operating Systems, there's a chance that they will implement them with a good quality, because they know the packaged software certainly better than an external developer from the OS team: most people like keeping their car polished, but not in working for a car wash. After all, most ebuilds under Gentoo are not going much further the statement that "it works fine". Obviously, this a two-edge blade: unlimited free commit to any quantity and quality of ebuilds is given, when the unofficial overlay grows larger, its quality will evenually decrease to an unacceptable level. So maybe a good compromise could be to limit access of users to one specific software or series of softwares, giving priority to those who actually develop the software they want to package under Gentoo. This way, we would be able to improve sinergy between herds and software developers and maybe lift some work to external sources while avoiding the risk of malicious code injected into widely-used packages. In OpenBSD, external developers must show their diligence and knowledge of ports system, before they're given CVS access to their port.
While I'm not upholding the idea that we should conform to OpenBSD in its kind of management (Actually, I'm not expert enough to have a solid opinion on wether Sunrise project should be aborted or continued), the above hints may be useful to this discussion, in order to understand what could happen or not happen in the future if we go through a certain way and they may be useful to formulate new ideas or proposals. Last, IMHO we should avoid the word "support" regarding Sunrise, because this word is ambiguous, since it has two different meanings: a) Yes, we support it, meaning that we endorse it and spend some of our resources to make it work. b) No, we don't support it, since we don't give any warranty regarding its quality. Sincerely, Giacomo 'jwk' Cariello -- gentoo-dev@gentoo.org mailing list
