Don't know about CouchDB well, but personally recommend to go to more popular one (cassandra, mongo, hbase, riak etc) so you may find better documentations and more data points as far as nosql is concerned.
An interesting post to see how start could differentiate themselves in very hot but competitive space, especially where bigger players are trying to get in.
Interesting view. But I think this FB case, maybe it just happened that way..
Aside from the size of startup & investment communities, there are still lots of great early stage investors e.c. JBoss backer David Skok of Matrix. - one of my favorit VCs. At least these guys sometimes do very very early stages.
Looks they try competing against Cloudera and its HUE by having their own distro/management tool. Guess it's good news for hadoop eco-system.
1) "Red Hat vs SuSE" in linux analogy - such distro competition generated some confusions & lots of 'ahead power' to many sub-projects, e.c. kvm vs xen.
2) Fragmentation of piece of technology is dangerous? I don't think so. Hadoop already has a unified code base in Apache where both Hortonworks and Cloudera's contributions go back. I think "fragmentation" at the this layer where standardizations aren't always required is good thing.