[maemo-developers] Package promoting
From: Polyvertex polyvertex at gmail.comDate: Thu Sep 23 11:10:08 EEST 2010
- Previous message: Package promoting
- Next message: Package promoting
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Wed, Sep 22, 2010 at 22:51, Attila Csipa <maemo at csipa.in.rs> wrote: > On Wed, Sep 22, 2010 at 11:28 PM, Polyvertex <polyvertex at gmail.com> wrote: >> >> If I am a lambda user, I just want to get applications on my phone and >> don't want to spent time for downloading an another application B to >> note or vote for application A. > > I personally have no problem with that. I actually even support that - you > don't want to test, you don't want to get karma, just plain Apps - cool ! > That's still better than mobbing people into voting for apps. And technically speaking, how do you 'support' casualness exactly ? Please don't turn this into a flamebait, this is not my intention... > >> >> If I feel unhappy with the latest application I've downloaded, I just >> want a simple/unique place to say it to the maintainers and eventually >> have a voting system to vote for/against a (group of) developer(s) >> according to the quality of its/their releases, that would make sense >> here. > > You have that, but given the way HAM works, lambda users will have > difficulties finding it and what you get is a talk thread with a 'Maemo > sucks' title. It seems everyone here is happy with the idea of creating an extra-application to vote for/against an another application. Even if, of course, I do appreciate the effort (your effort in this case) to improve the community living, I still consider to be a non-sense the idea of needing an application to do voting only because the server-side partially failed to do the job. I mean, I can understand it is maybe better for the user to be able to immediately indicate a mood about a given application but don't you think the initial problem is that users still don't know where to report issues ? Package page ? Garage page ? New talk thread ? Find developer's e-mail address ? Also, don't you think this KISStester thing should be completely integrated into the application manager so users can see feedbacks from others *before* installing ? In this case, this would totally legitimate this new feature. > >> If you want such a perfect QA process, why not building a big >> Ovi/AppStore, hire some testers and application approval team, and >> stop saying Maemo.org is an open community ? > > Hey. That *is* a bit harsh. Ideas are of course always welcome. But just as > I feel we don't have the right to do the above, I also feel we don't have > the right to shove down updates/apps down the throats of unsuspecting users > who are not interested in half-baked software. I mean, Extras-devel IS > public and advertised - anybody interested in bleeding edge software can get > it there, and that's OK. OTOH There ARE N900 owners who really don't want to > reflash every other week and think hunting down processes and > broken/oversized packages is not fun - and that's who Extras is for. So > there, to each to his own. Yes I totally agree with you but you missed my point. My initial point is about stuck releases into extras-testing, not about pushing an alpha version from extras-devel directly into extras. I consider this extras-testing step to be too constraining for the developers and the users : if a user wants bleeding edge versions, he will use this advertised 'extras-devel' repository, otherwise, the 'extras' one should be good enough for him. Regarding the 'reflash' problem due to a very bad application, this is where a good enough feedback system should be so the number of impacted users should be quickly limited. Again, users don't know where to report. > >> I would prefer the Maemo.org community to act like a regular open >> community and let people vote for/against an application *after* it >> has been released. Thus, only 2 repositories should be enough : >> 'extras' and 'extras-beta'. > > I must be missing something - that has nothing to do with community > openness, and in fact I'm inclined to say there isn't a single distro that > works that way. Imagine Debian said 'ok, everything the devs want to promote > gets promoted to stable, and if there are complaints, we'll remove it from > the repo'. We kind of had that with Chinook/Diablo. It didn't work, really. > We had a lot (LOT) more complaints about borked installs. I cannot discuss about Chinook and Diablo distros since I never used them but that is my point, maybe it didn't work because there was no integrated feedback system into the application manager (or its equivalent) ? You are comparing Debian releasing *cycle* with the Maemo.org's package promoting process. You should not for obvious reasons and also, I am happy I don't have to wait 3/4 years to get the next stable version of gPodder. I don't criticize the releasing cycle of the Debian stable, in fact, I heavily rely on it everyday but I am happy that applications on my N900 are released a bit faster (not quite enough actually). Generally speaking, I can see goodwill everywhere on Maemo.org. This community is very strong and this thread tends to demonstrate it, but I feel like everyone want to "Do The Right Thing" without any pragmatism. I mean, as a community oriented project, you cannot have such a perfect QA process, and the way the Debian project is organized seems indeed to be good for such a thing but Maemo.org does not work this way at all. Also, this does not mean it has to become anarchy and application/virus/bug bombing just because there are 2 repositories instead of 3. Well, I'll stop here, my English is not good enough to clearly elaborate my point of view and I can see that continuing again and again against the same arguments is a non-sense (i.e.: "we've already done that, it doesn't work"). Though I'm a bit disappointed, at least I understood how things work in here and I am even happy to have a view of things to come. Cheers, Jean-Charles
- Previous message: Package promoting
- Next message: Package promoting
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]