While both are document stores, there are some major differences in their
data model, most notably that mongoDB uses an update-replaces mechanism,
while CouchDB allows you to access any version of a document, which brings
with it issues of transaction overlaps (who wins?) and having to
periodically compact your database.

CouchDB uses a REST interface for all interaction; mongo has programming
language-specific drivers (although there are also REST interfaces
available), which in many cases can increase performance.

Their querying approaches are differnet. Mongo is more akin to a "define an
index and use it when possible at query time." CouchDB is more of a "Define
a view beforehand and use that view".

Oops. I just found a better overview than I can provide, at

There are lots of other players in this space, too -- see


On Tue, Dec 14, 2010 at 9:12 AM, Thomas Dowling <[log in to unmask]>wrote:

> On 12/14/2010 07:58 AM, Luciano Ramalho wrote:
> >
> > I believe CouchDB will take the library world by storm, and the sooner
> > the better.
> >
> > A document database is what we need for many of our applications.
> > CouchDB, with its 100% RESTful API is a highly productive web-services
> > platform with a document oriented data model and built-in peer-to-peer
> > replication. In short, it does very well lots of things we need done.
> Amen.  Does anyone have helpful things to say about choosing between
> CouchDB and MongoDB?
> Thomas Dowling
> [log in to unmask]

Bill Dueber
Library Systems Programmer
University of Michigan Library