On 06/23/2009 10:17 PM, Dale Wiles wrote:
I've been following CouchDB for a while and I'm really impressed with the way
it's coming along.
However, like>90% of the users out there, I'm not a business and I don't really
care about replication and daemons and scary things that go bump in the kernel. I
just need somewhere to put my address lists and record collection.
Are there any plans, or is it even feasible, to make a serverless version of
CouchDB, in a manor similar to SqLite?
For those who don't know SqLite: CouchDB would be a file or set of files. It "starts up"
when the file is opened and "shuts down" when the file closes. If you can read the file
you can read the DB. If you can write the file, you an write the DB. Database locking is handled
by the DB.
I think a *lot* of potential casual database users would be interested in a no
hassle/no mystery version of CouchDB they could play with. It's something to
think about.
I totally agree with you that we should get CouchDB running in lots of
places, exactly the places where the delightful SQLite system runs now.
I don't quite agree that giving up listening on a network port is
necessary to do that though. One of the beautiful things about CouchDB
is the HTTP interface, and a network port is a big part of that. Even
programs running on my phone listen on network ports.
In Ubuntu, we're working to shrink down the Erlang packages to be small
enough to fit on the LiveCD, and in the next version of Ubuntu we plan
to have a CouchDB instance automatically running for every single user
on their desktop. You know, a nice place to put your address lists and
record collection :) If you'd like to follow along or help out:
https://wiki.ubuntu.com/DesktopTeam/Specs/Karmic/IntegratingWithUbuntuOne/#line-68
In fact, this effort to run a CouchDB for every single user account on a
computer running Ubuntu is why Stuart filed bug
https://issues.apache.org/jira/browse/COUCHDB-393 yesterday.
--
Elliot Murphy | https://launchpad.net/~statik/
