Comments on: New Webcast: Understanding REST

By: Dan Rigsby » New Webcast: Syndication in .Net

Dan Rigsby » New Webcast: Syndication in .Net — Fri, 07 Nov 2008 12:54:05 +0000

[...] namespace to create and consume RSS or Atom feeds. It builds on top of the Understanding REST video, but understanding of REST is not necessary. This video also complements a recent presentation I [...]

By: Dan Rigsby » New Webcast: REST Support in WCF

Dan Rigsby » New Webcast: REST Support in WCF — Thu, 25 Sep 2008 13:58:02 +0000

[...] webcast up over adding REST support to a WCF application. This video builds on top of the Understanding REST video. If you have a decent understanding of REST, you don’t necessarily need to watch that video [...]

By: Dan Rigsby

Dan Rigsby — Fri, 12 Sep 2008 21:47:31 +0000

Good comments. With the XSD remark I was more refering the what operations were available, the structure of the urls, and automatic pull down of XSD. You could of course have another REST endpoint that returns back the XSD for the expected and returned types.

I did think about adding idempotent requests, but felt it was getting into the details too deeply (I tried to make this a 6 minute video but still failed). Thanks for adding this comment.

Thanks!

By: Jason Karns

Jason Karns — Fri, 12 Sep 2008 20:59:52 +0000

Overall I thought the Webcast provided a good summary and overview of REST. It's very hard to take REST, which is such a broad topic with very debatable/controversial aspects, and boil it down to a 101 level intro. I would like to counter two of the items you listed as disadvantages to REST, namely Metadata and the filesize drawbacks. You mentioned XSD for metadata in SOAP. There is absolutely nothing stopping one from creating and referencing a schema for the resulting XML by attaching the XMLSchema-instance namespace and adding the schemaLocation attribute to an XML document. In fact, linking to the schema in this manner follows RESTful principles by referencing other resources through links. With respect to the filesize issue, HTTP has a built-in mechanism for handling large files. This is accomplished by the server and client using the following headers: Accept-Ranges, Content-Length, Content-Range along with the 206 Partial Content Response code. One thing missing from the disadvantages, however, was the issue of authentication. I have yet to see a working example of high-REST that included authentication. Lastly, I think the issue of idempotent requests was missing when discussing the 4 main HTTP methods. To me, that is what distinguishes a POST vs PUT (PUT should be idempotent, while POST needn't be). Also, one should PUT an item directly to a resource URI to create it, while POSTing an item to its parent collection to create it. But this gets into implementation details that don't belong in an introductory webcast anyway! Great Webcast!