On 26-Jun-07, at 10:56 AM, Jason Birch wrote:
I personally really don't want to see this kind of complex permissions
structure in place. We need to start from a position of trust. If this
system is abused:
- We can re-evaluate our approach
- We always have SVN history to go back to
Two issues are being mixed up here...
1 - Administrative overhead - it seems that from a systems admin perspective it is prohibitive to implement or maintain. If this is the case, then SAC can say no, or we can find the methods or tools we need to get it done. What we have right now are some administrators of parts of the osgeo svn that want to partition access control. At present there are only 3 or 4 projects that really need this. I can't see it growing much more than that.
2 - Management need - This isn't about trust, it's about managing a logical organisation between several projects that are largely unrelated. It's more about empowering management groups of these projects to care for their areas and not be responsible for others.
In my mind, leaving it is as strange as having mapguide, fdo, geodata, openlayers, journal and gdal all in the same repository. Will future software projects be put into the osgeo repository? Who administers these permissions across all projects? I think you get the point. Having spanish book authors with full access to the web site theme subfolder (and vice-versa) doesn't make sense to me.
At this point there are very few people actually using this repository, but I believe it is better to think ahead than to deal with consequences later.
Gary - do you have any tips you can share? Can we look at how you have been doing this with QGIS - has it been a burden?
The group permissions tool we've had so far has been working great, but as far as I can tell we just need a couple more groups.
I won't belabour the point, but just wanted to lay out my thought process ![:slight_smile: :slight_smile:](/images/emoji/twitter/slight_smile.png?v=12)
Tyler