On Wed, Nov 6, 2013 at 10:54 PM, Alex Mandel <tech_...@wildintellect.com> wrote: > On 11/06/2013 12:41 PM, Ivan Mincik wrote: >> Dear Alan, at first thank You for Your work (and also for work of others in >> this project). +1
>> 1. This project definitely needs a strong leader with experience. >> > Or PSC, I don't think there needs to be one person in charge if there's > a group that meets regularly to set the items below. +1 > >> 2. UbuntuGIS lacks clear roadmap. Maybe it is loosing skilled hands just >> because they are not aware about fact they are needed. Preparing roadmap >> must by one of the most important tasks of project leader and PSC. >> > It doesn't seem like this would be hard to write up if a few of us met > online for a few minutes. +1 > >> 3. Other very important task is clearly stated work flow in terms how to >> cooperate with Debian on regular basis. >> > Agreed, seems recent meetings might have started down this path. Though I agree we should cooperate, I think that the actual bottleneck is having debian developers uploading packages to debian unstable. There is quite a long list of Geo-related packages on mentors.debian.net or experimental waiting for sponsors or transition. This is also important for ubuntugis: since the next release (trusty) will be a LTS release, so much more users will be using it for a long period, and also ubuntugis will have to work nicely with them. So I'd suggest that instead of only waiting for some transitions to happen in debian we should also consider sync requests [1] or feature for syncing some packages from experimental to universe. I'm especially thinking about gdal, because we now have a version of gdal in ubuntugis which is incompatible with the one in ubuntu universe. This means effectively means that ubuntugis is currently incompatible with all other gis packages around, including the ones included in ubuntu. This makes it eg impossible to install saga 2.1.0 from my ppa alongside qgis 2.0.1. >> 4. Still at least by my opinion, the PPA naming stable, testing, unstable >> is very confusing for all newcomers which automatically expect the same >> behavior as in Debian. If this schema remains the same, or it will change >> to something other, it needs clearly documented workflow how packages >> migrates from testing, staging to production and how often and under which >> circumstances production packages are upgraded. >> >I agree this is confusing and maybe the roadmap helps clear it up. Proposal: remove testing (which is mostly empty) and add experimental (which would be a more appropriate name for what we use it). I would not add a testing in debian sense as maintaining 2 repositories on the zillion ubuntu distributions is already enough work. > >> 5. As in other voluntary projects, there is a lack of manpower. But on the >> other hand, there are no rules which new contributors or uploaders must >> meet. Also I do not see any list of free tasks for new people. >> > I'm not aware of anyone being a designated maintainer, so its more a > free for all once someone is granted access. Maybe we need to keep a > list of who is responsible for what, of course when one project needs > half the packages rebuilt that gets tricky to coordinate/wait. The unclarity is bad. If we would send out a message on this list: let's start building for saucy, let's try rebuilding every package from raring on the testing (current naming) branch this week, so we can put them on -unstable next week at least I would be more stimulated and try to build some packages. Ubuntu releases are predictable, so we can make a schedule an plan ahead. Johan [1] https://wiki.ubuntu.com/SyncRequestProcess _______________________________________________ UbuntuGIS mailing list Ubuntu@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/ubuntu http://trac.osgeo.org/ubuntugis/wiki
