Talk: Techniques for Building Third-party RESTful Web Services

I gave a talk today for Py Web SF called "Techniques for Building Third-party RESTful Web Services". Aside from covering a lot of material from the book "RESTful Web Services", I also covered a handful of tips and tricks that I discovered were necessary when creating a third-party RESTful Web Service meant for consumption by Web browsers. Here are the slides in format.


Heikki Toivonen said…
While I also use py-restclient, I have a few issues with it. 1) It is more work to override the transport to use something secure (like M2Crypto) than I'd like. 2) I was not expecting it to raise exceptions on non-2XX status codes. I understand the second point is probably a good point for some, but in my case I want to deal with all status codes in a uniform manner, and py-restclient's habit of raising exceptions forces me to translate them back to status codes. 3) It seems like it is also more work than I'd like to make py-restclient use disk cache.
Thanks. I haven't used py-restclient, so your comments are helpful. I only mentioned it because Drew Perttula said he really liked it.
If you're running OS X, I suggest NeoOffice to open the slides.
Brent said…
I believe that urllib2 has that same behavior of raising exceptions for non 2xx status codes, so I suspect that py-restclient is using urllib2 (or urllib?) under the hood. I had the same annoyance at urllib2's raising exceptions when using it for a little REST API testing framework I wrote.

The answer, I think, is to go down a level and use httplib directly instead of living with urllib2 and its assumption that all 403s and 404s (for example) should have exceptions raised for them. I've been too lazy to do it yet though. :)
Someone at the talk I gave last night also mentioned restkit.

Also, rather than have a script tag load content into a div above the script tag, it's possible to replace the script tag with an automatically generated div. said…
Cool talk JJ - the "third party" bit made it very practical.

Apparently the Firefox support for XHR across domains I mentioned is actually a standard. See and I haven't checked to see what other modern browsers do/will support it.
Thanks, Simeon.
aleax said…
Clicking on the "here are the slides" link I get a 404:

404 Not Found
The server can not find the requested page: (port 80)
Please forward this error screen to's WebMaster.
Sorry about that. My buddy must have stopped paying for hosting. Try this.