Comments on: RESTful Partial Updates

By: Subbu Allamaraju

Subbu Allamaraju — Sun, 17 Feb 2008 23:17:57 +0000

Nice attempt. You solved the code problem – not just the diff format problem. Wish there is a similar solution for XML.

By: rsayre

rsayre — Sat, 16 Feb 2008 03:53:57 +0000

We have hardware cache / SSL accelerators that don’t get along with Wordpress.

By: James Snell

James Snell — Sat, 16 Feb 2008 03:29:07 +0000

Thanks for fixing it. Unrelated: I’m not sure why but I’m seeing some strange caching issue with your blog. I left the comment, verified that it was there. Refreshed the page a few minutes later to see if you had responded, and it was gone. Refreshed a few more times, making sure that my local cache was cleared, and it was still gone. Then, it was back, with your response. I decided to come back to thank you for updating the post the comment (and your response) were gone again. Refresh the page again, and now they’re back.

By: rsayre

rsayre — Sat, 16 Feb 2008 02:36:16 +0000

Sure, PATCH is more explicit. In theory. In practice, you’d need a POST-accepting resource that didn’t want to apply it… hey, an Atom media collection is a good real world example of this. You’re right. I’ll update the blog post.

By: James Snell

James Snell — Sat, 16 Feb 2008 02:29:33 +0000

Definitely better than Joe’s suggestion but using POST still leaves the client tied to whatever meaning any particular server chooses to assign to that operation. That is, POSTing a x-application/json-sync to one URI (e.g an Atompub edit uri) can have an entirely different meaning than POSTing the same thing to a different uri (e.g. an atompub collection URI). With PATCH, on the other hand, the request means the same thing in both places.

By: rsayre

rsayre — Sat, 16 Feb 2008 01:48:08 +0000

Craig, I used POST.