I am using Couchdb in a "LYCE" application (you know, Linux Yaws Couch
Erlang), and I use Couchdb to store data.
I found two differents modules in the wiki able to communicate with
Couch, but both do json enconding then decoding, and uses http as transport.
I was wondering: could not be possibile to communicate with the couchdb
erlang process using Erlang mechanism for distribuited processing?
This way, you simply connect to the couchdb as a node, and talk straight
to couch, bypassing all the json encoding and decoding
(this is particularly important for me because I have to serve large files).
Any suggestion where to look to try to implement such a functionality?
Also, I think that exposing an erlang API could be the first step
implementing partitioning.
