News at Couch

Welcome to yet another installment of News at Couch, our review of what’s new with, on and around CouchDB.

CouchDB 0.8.0 was released! Biiig news. This is our first release with the ASF and we are all very pleased with this. Great work everybody! Changes include:

• Changed core licensing to the Apache Software License 2.0.
• Refactoring of the core view and storage engines.
• Added support for incremental map/reduce views.
• Changed database file format.
• Many improvements to Futon, the web administration interface.
• Miscellaneous improvements to system integration and portability.
• Swapped out Erlang’s inets HTTP server for the Mochiweb HTTP server.
• SpiderMonkey is no longer included with CouchDB, but rather treated as an external dependency.
• Added bits of awesome.

Christopher Lenz fixed HTTP 1.0 handling in our copy of MochiWeb and got his patches accepted at the official distribution. Congrats to Christopher and Benoit Chesneau who helped figuring out what went wrong. Why is that relevant? A lot of HTTP proxy servers talk HTTP 1.0 to their respective backends and the patch allows you to use CouchDB behind your favourite HTTP proxy.

This is not exactly news but I somehow missed to mention Paisley in the past. Paisley is a connector for the excellent Python framework Twisted. Twisted is king when you are developing applications that need asynchronous event handling (which is everything between network programming and GUIs). With Paisley, it is very easy to integrate CouchDB into your Twisted-based application. It might not work perfectly with the recent CouchDB 0.8.0 release, but with a few minor tweaks, it should be fine.

Speaking of language bindings, Kore Nordmann released PHPillow, the most comprehensive, best documented (it has docs :-) and unit-tested library for PHP. I highly recommend PHPillow over the (comparatively crappy) PHP library I released. And it is compatible with CouchDB 0.8.0.

Damien will be speaking at RubyFringe this month. Make sure the see the great man in action when you are going!

Oh conferences, yeah. I went to the Erlang eXchange in London last week and it was a blast! I will write a dedicated report but while you wait for that, check out the video of my presentation. I still do the ‘hands thing’ and I was mighty nervous with Joe sitting in the front row and Klacke way in the back. But it went well I guess, too bad the video doesn’t show the standing ovations I did not actually get. And I’ll be speaking at OSCON.

Chris Anderson of grabb.it fame posted a row of interesting things on his blog and even started working with CouchDB. Here’s a short rundown on what he’s been up to: Chris explains why CouchDB is relevant for him. Chris releases couchrest, a Ruby library for working with CouchDB (that has been gemified since). Chris explains how to do word counts with map/reduce. Chris creates Markov Chains with group reduce. Finally, Chris turns CouchDB into a self-contained application server. And finally finally, he co-founds the first CouchDB user group in Portland, OR. Has he been busy or what? Thanks Chris for all your contributions!

Todd Hoff posted a very informed article about why it is a good idea to trade disk space for scalability. This is not something CouchDB-specific, but a concept that CouchDB heartily embraces.

The aforementioned NodeDB project is looking for contributors.

Can CouchDB solve email client problems?

Chris Hartjes encourages you to never stop learning. The tools & techniques you use successfully today might be useless tomorrow. At least know what the competition is up to.

A very interesting research paper made the rounds suggesting that you can project JOIN operations on a map/reduce system. This might help to convince people that we are onto something.

Shawn connects CouchDB and ExtJS. His examples show how neatly designed ExtJS’s JSON store is and how perfectly it fits onto CouchDB’s RESTful JSON API.

And here is another one by Kris Zyp, he connects CouchDB to Dojo.

John Paulett released jsonpickle, a JSON to Python object converter that is a bit smarter than the regular simplejson library in that it can convert JSON objects into non-native Python types.

Over at Resourceful Idiot they have a nice list of RESTful databases.

Simon Willison quotes Steve Jegge. This is exactly why CouchDB switched from XML to JSON.

Ken Downs has a case against ORMs (Object Relational Mappers).

BradfordW has a smaller and a larger example on how to use CouchDB with DataMapper.

Bob Briski (of Yahoo!) explains how to count clicks with CouchDB:

I don’t lose any granularity, it takes almost no time to maintain and it’s incredibly fast.

Peter suggests that your middleware and architecture naturally follow your problem domain and cannot, or at least should not, be chosen independently.

Loren Segal released CouchIO:

[…] a simple VFS for Ruby, adding support to open, read and write CouchDB resource URI’s as if they were local files.

Yoan Blanc creates an Erlang based web chat system on top of CouchDB.

—

Thanks for joining me for this issue of News at Couch and until next time: Take couchcare!