On Mon, Jul 15, 2002 at 05:57:15PM -0700, Mark Vojkovich wrote: > On Tue, 16 Jul 2002, Jos� Fonseca wrote: > [...] > > > > The XFree86 developer page (http://www.xfree86.org/developer) says and I > > quote: > > > > "When requesting to join the XFree86, the most important qualification is > > not your experience level but your keeness on contributing to the project > > and climbing the uphill road to learning and mastering XFree86." > > > > And now a core member completely states otherwise!? > > I am particularly addressing architectural issues. There is a > difference between saying that everyone can contribute something > and saying that all problems are easy enough for anyone to solve. > Moving the project forward in terms of innovation and features is > rarely done by newbies. I don't think it's reasonable to expect > someone with little experience to jump in and write an extension > to provide window transparency, for example. People ask why we > don't have stuff like that yet and it's precisely for that reason. >
I agree that it isn't reasonable to expect someone with little experience to do big things right away, but if they have the _long term_ ambition to do such a big thing, there is no reason to expect that it will fail on that aim. But resuming the main issue here, it seems to me that one reason for the difficulty in adding new features - which require a profound knowledge of the architecture - is that those who have these knowledge are too busy with tasks that could be delegated to developers less experience but yet competent. I'm for instance referring to testing, debugging, and patches submition, i.e., Quality Assurance. For instance, the XFree86 should have a scheme with several CVS branches, each one with a small set of responsables and specific to a card or family. These developers would receive patches related with these cards, analyze them, test them, and then report to the core team for the merge with the trunk. This the core team would had their job facilitated and have more time to dedicate themselves to architectural changes. Note that in these cases one could give CVS access to more people, with the condition that they would'nt submit to the trunk directly. (This is how is done implicitly with DRI, i.e., developers ask for feedback before submiting somthing directly to the trunk). Something like this would function better with a proper bug database, but I still believe that it would be a better model without it than the current development model. (I would like to step in and setup a Bug database that could be used both for the XFree86 and DRI projects, but no matter how much I would like that my time stretched, it doesn't, so I can't take responsabilities which I wouldn't be able to keep.) [...] > > > > I'm sorry to say that is the kind of attitude that you (and others > > like you) have towards potential new developers that is holding the > > XFree86 development down. You fail to realize that there is a thin line > > between the experienced and not experienced, and that those who do have > > the experience also have the power to quickly transform an unexperience > > yet motivated soul into an experienced one. > > > I don't have an "attitude" towards potential developers. I'm > just being realistic and I'm speaking from experience. Everyone > can contribute, but good intentions aren't enough to get all the > features that some people are asking for - we need people who know > what they are doing. There are two scenarios: > > 1) Newbie joins with plans to do big things. Finds out that this > stuff is hard and quits. > > 2) Newbie joins and does small things for a year or two and builds > up to doing big things. > > Number one seems to be much more common than number two. I don't > know what we can do to get more people into the number two category. One thing that could help get more people into the number two category, is facilitate and incentivate people to start doing small things. AFAIK, patches go to a big black hole which is [EMAIL PROTECTED] and very few come out or are acknowledged... this goes back to what I said before. [...} > > > > > <dream> > > Give CVS access for more people, open up the development, close > > the closed development mailing lists, substitute the central development > > model for effective QA, incentivate people to help, and make sure their > > involvement is appreciate... > > </dream> > > > > We need more people worthy of CVS access. You might think this is > an elitist attitude, but you haven't seen some of the patches we've > been getting. CVS access is granted only after a strong history of > good patches has been established. We actually do have a large number > of people with CVS access. I don't know how it compares to something > like Linux kernel, but I wouldn't be surprised if it was similar. > > ps. There aren't really any closed development lists anymore. Or rather, > there are some but nobody uses them. I really don't know then. For several months I have been monitoring the xpert mailing list and I still have little clues to how the development is headed besides some keywords one hears once in a while like RandR, RENDER, ... but I can't tell how far these things are done, who's working on them, how can some one test or try to help... For example on DRI just by monitoring dri-devel I (or any interested person) can tell you who's working on what, what bugfixes have been commited lately, where's the development headed, and one just need to mention that experienced a bug to have users report whether they were able to reproduce it or not, and offering to make further testing. Unintentional or not, there seems to be a smoke curtain in the XFree86 development that surely isn't attracting new people. Jos� Fonseca _______________________________________________ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
