Jeff Hinrichs - DM&T
jeffh at dundeemt.com
Tue Jul 8 04:48:15 CEST 2008
I'll post more on it as I go. Right now the couchdb crew are making
some changes to attachments -- which is a good thing, since even
though attachments are supported, they are not available via Futon
(the web admin interface) nor couchdb-python module. All in all,
pretty good timing. I'm playing with a project to archive documents
(pdfs). While pdfs support meta data (to a degree) that support
varies wildly with what version the pdf is created. Other times that
I've done similar projects, I've stored the meta-data in a database
and hashed the storage location for the actual pdfs. It works, its
searchable, its usable, but its very fragile -- and then on top of it,
you need to write your access stuff for a browser based client. Not
to mention, replication. I've rolled my own before and lets just say
I see the value in having someone else do the heavy lifting on that
If couchdb works the way I think/hope/expect it will, this type of
project will become trivial. I am planning on using the stock couchdb
already Python and Ruby implementations to handle the map/reduce
functions for views. couchdb-python comes with the necessary script
to implement python based views.
Currently, things are on a slow burn, lots of admin stuff at work
right now and this project is definitely a skunk work style project.
I normally do the really cutting edge stuff after hours -- not all of
it pans out.
On a completely different tack, anyone out there use svk? We use
subversion internally but recently I've come to the conclusion that
some of our projects would be well served with some distributed vcs
capabilities. I've looked at hg and bzr but svk seems to get me what
I'm after and melds well with subversion. Anyone else using
subversion + dcvs?
On Thu, Jul 3, 2008 at 9:17 AM, Matthew Nuzum <newz at bearfruit.org> wrote:
> On Wed, Jul 2, 2008 at 12:43 PM, Jeff Hinrichs - DM&T
> <jeffh at dundeemt.com> wrote:
>> I have been playing around with CouchDB and Python. I can give a
>> lightening talk about what CouchDB is and what I've learned with the
>> Python interface.
> I'd be interested in hearing your experiences with this. I've played
> with it too, but just enough to excite the imagination, not really
> create anything useful. To me it feels like a very high quality ORM
> with out a need for the gritty SQL abstraction layer in the middle.
> Matthew Nuzum
> newz2000 on freenode
> Omaha Python Users Group mailing list
> Omaha at python.org
More information about the Omaha