On 2007-11-21, Jeremy Maitin-Shepard <[EMAIL PROTECTED]> wrote: > 1. You do not want the reputation of Ion (and your work) to be hurt due > to distributions associating the name Ion with old or patched (and > therefore possibly worse) versions of Ion. > > 2. You do not want to have to deal with support requests for old or > modified versions of Ion.
Mostly right. > In addition, you might also want to ensure that everyone using Ion is > using the best possible version (which you assume to be the current, > unmodified version), No, just that people wouldn't install old versions when a newer one is already available (on a particular branch, i.e. Ion1, Ion2, Ion3, ...), unless they really know what they're doing. If they're already using an old version that works for them, fine, just don't come asking for help before upgrading. > I'll address the first two objectives below, but I'll explain first why > this last objective does not need to be considered: I think that it > really is reasonable for distributions to decide what version of Ion to > provide users; it doesn't make sense to try to force users of some > "stable" distribution, like Debian stable, to use the very latest > version of Ion. Ion3 is not "stable"; it should not be in "stable" (read: "static") distributions. So you forbid distribution of development snapshots? The herd would still cry that it's "not free", and these same distros even refuse to fix bugs in "stable" software. They don't trust authors, and authors can't trust distros. > One significant reason users pick a distribution is > because they _want_ to follow the judgment of the distribution packagers > in selecting particular versions of packages and particular patches to > apply. If they use their own judgement, they should take full responsibility of supporting the package. That means making sure their users are aware that the package no longer has anything to do with me. > I don't think you should be concerned about users being > given convenient access to "sub-standard" versions of Ion. I should, when The Party, i.e. the big distros, have de facto central control on convenient access. Closed source operating systems are more decentralised in this area. > As far as the first objective, it is common knowledge that distributions > often make significant patches to the packages they distribute, and > sometimes those patches can change the behavior of the package in > certain ways, or in the worst case introduce bugs. It is also common > knowledge that distributions typically don't always distribute the > latest version of a package. It is not common enough knowledge. Maybe among old-time users, but not the flood of newbies. And certainly it is not common enough knowledge that upstream will have nothing to do with the distros' corrupted and ancient versions. The distros don't even bother to properly indicate the development snapshot status of the software they distribute. > More importantly, though, > you can ensure that this fact is very clear by requesting that > distributions include a very prominent message in any packages of Ion > that are modified or not current. They're not going for that. The "new version in 28-days" clause in the present license has that as an alternative, but the FOSS/distro herd considers such simple reasonable requests "unreasonable" and the software "non-free". So why have such a clause when nobody acts by it? > As far as the second objective, the problem seems to come down to > eliminating what you would consider to be effectively junk e-mails > (support requests for old/modified Ion versions) from the Ion mailing > list. The mailing list is not the only contact address you know. > Ion mailing list that you are willing to provide support for the current > official Ion version, but you don't have time to provide support for old > or modified versions of Ion, The current "stable" branch is also somewhat supported despite newer developments. > I think distributions should certainly try to respect your wishes > regarding the packaging of Ion They don't. They've become a powerful central Party that just uses authors as workhorses for their own ends. > I'd like to think that > there can be a friendly relationship between distribution package > maintainers, users, and free software developers. Dream on. That is unlikely to be possible until the middle-man has been removed and authors have regained control of the primary distribution chain; until distributions have been reduced into providing base systems and customised versions of select software for their audience, from their present position of central control over conveniently installable software. Distros are the record/film companies and their lobby groups of FOSS! > someone merely looking to "check out" Ion is less likely to > do so if it requires manually obtaining the source code, building it, > and installing it) And with distros doing whatever they will, someone just checking out Ion will get crap not representative of the project's present state, or the project at all. -- Tuomo
