Comments on: RESTful Django practice

By: RaiulBaztepo

RaiulBaztepo — Sun, 29 Mar 2009 01:06:59 +0000

Hello!
Very Interesting post! Thank you for such interesting resource!
PS: Sorry for my bad english, I’v just started to learn this language ;)
See you!
Your, Raiul Baztepo

By: Martinez Marchand

Martinez Marchand — Sat, 28 Mar 2009 15:16:50 +0000

The issue with query strings is that the search may not be entirely accurate.

By: links for 2009-01-17 | zoia.org

links for 2009-01-17 | zoia.org — Sun, 18 Jan 2009 00:04:45 +0000

[...] RESTful Django practice | Refactor the Life (tags: django python rest webdev) [...]

By: Jon Moore

Jon Moore — Fri, 16 Jan 2009 21:37:27 +0000

(Whoops, did not escape my markup in the first post!)

In _RESTful Web Services_, one suggestion would be to have /bookshelf/books/{isbn} be metadata; this could be returned in a structured format (HTML, XML, JSON) and could contain elements pointing to the available representations, as in:

This makes the formats discoverable and “hypertext-driven”.

One downside to the use of the HTTP Accept header to specify format is that some clients cannot easily modify their headers (e.g. how do I make my web browser change that for a request)?

By: Jon Moore

Jon Moore — Fri, 16 Jan 2009 21:35:34 +0000

This makes the formats discoverable and “hypertext-driven”.

One downside to the use of the HTTP Accept header to specify format is that some clients cannot easily modify their headers (e.g. how do I make my web browser change that for a request)?

By: Adrian

Adrian — Fri, 16 Jan 2009 08:00:24 +0000

@Masklinn having the format in the querystring as an optional parameter makes it easier to “smoke test” services on the browser… Just my opinion, I think (and I’ve usually done that) you should have both, but enforce the header option.

By: bookstack

bookstack — Fri, 16 Jan 2009 04:54:08 +0000

@Masklinn:

If the user wants to check out the available eBook formats, but no intention to know the details of authors, publishers, … I am wondering whether HEAD could acts as a simple query? If not so, it can be dropped.

I also checked out roy.biv.com, he assumes that each URL should be representable, and guide the blind web crawler to find all the contents. In that sense, we need the following list:

/bookshelf/: list available object list: books, authors, publishers, etc
/bookshelf/books: list all available books with pagination. This is where query string shines: ?page=n