Hi all,

I have opened a JIRA issue asking INFRA to configure the "used by" indicator
on our GitHub repos (native-git ones as well as svn-mirrors) such that they
reflect the "most commonly used" module and thus give a more reasonably
indication of the popularity of the different UIMA sub-projects.

  https://issues.apache.org/jira/browse/INFRA-19412

E.g. for apache-uimaj, the number of repos using it should not be 1 but
rather 914: 

  
https://github.com/apache/uima-uimaj/network/dependents?package_id=UGFja2FnZS0xNzk4MzkxNTI%3D

I wasn't really sure what to suggest for DUCC, so I picked the one which
I think has the most usages.

For other repos like addons and sandbox, there is really no single module
which makes sense to pick (I think). When/if these repos get updated at
some point, it would be worth considering either to split them up entirely
(i.e. one repo per component) or to set up one module which takes care
of version management and which people would then add to their POMs
into the dependency management section to import the versions of all
modules. Could also be some intermediate solution, e.g. splitting out
some modules (such as the SimpleServer) while keeping e.g. all 
analysis engines in one repo.



Cheers,

-- Richard

Reply via email to