[geomesa-users] Branching Model + listserv

[Andrew Hulbert <ahulbert@xxxxxxxx>]

Since we're talking about releases...I was wondering what people thought 
about changing out branching model to create a maintenance branch for 
the 1.2.x series and then bump master to 1.3.0 for new features. For 
example, new datastores like dynamo or the native API would go onto the 
1.3 master and we'd fixate things at 1.2.x and actually follow the 
semantic versioning guidelines since technically 1.2.1 and 1.2.2, etc 
should only be bug fixes.

The ideal goal is to have all 1.2.x clients work with 1.2.x server 
iterators and plugins. Granted thats tough, but as we mature I think its 
a good goal. Accumulo guarantees that the 1.6.x series are all binary 
compatible with each other (and also don't introduce any breaking 
changes on the server side) so you're original client at 1.6.0 works 
with a 1.6.5 cluster.

Also, do we have a geomesa-dev listserv and archives? We should probably 
try to keep technical discussion on there as much as possible to show 
that we are active on the user lists and discussion forums as an open 
source project! Granted many devs are at CCRI, but the LT listserv 
should be more visible to people who are interested in the project. 
Also, since many clients are internal to CCRI it means that bugs and 
questions will be googleable(?) assuming we have a good crawlable(?) 
archive.

Any thoughts?

Andrew