Wait. How did this happen?
I subscribe to both math.CO and cs.DM on my feed reader. They are different.
Maybe from certain TCS perspectives, they could be seen as the same, but let's step back a little.
From my impression, articles with co tags often have more links to "pure" mathematics, like algebra, geometry, topology and include raw enumeration or extremal problems. In contrast, articles with dm tags are usually more related to "applied" areas, like optimisation, dynamical systems and statistics. For instance, Ramsey theory might fall under the former's, while many parts of graph theory would be under the latter's umbrella.
Surely, like any classification system, this is neither a covering nor a packing, but why the need for arbitrary tag minimisation? Are we overpopulated by arXiv tags?
And, if any merging was to be done, shouldn't it have been the tag that is more general (and, arguably, closer to TCS) to survive? (Actually, I'm not sure if I completely understand how this merging behaves now, but the question still stands...)