News for hackers. Stuff that matters.
Re:After reading this URL, I long for a Linus anti-REST email by callahad on Friday July 09, 2010 @03:28PM (#1501895)
I'm only just starting on reading O'Reilly's "RESTful Web Services" (Richardson & Ruby), and I've but skimmed the Felding dissertation, but the URI doesn't strike me as particularly RESTful.

In particular, custom verbs are embedded within the querystring (a=blob, a=blob_plain, a=history, a=tree, etc.) rather than relying solely on HTTP verbs and varying the base URI.

And that base URI never changes: it's always http://git.kernel.org/, regardless of what resource is being requested.

If I can't determine what resource I'm requesting from the base URI alone, how can it still be considered RESTful and not an RPC-style or hybrid architecture?

Re:After reading this URL, I long for a Linus anti-REST email by xenoterracide on Friday July 09, 2010 @04:05PM (#1502044)
I personally think you can learn everything you need to know by reading ch8 and on. the appendices are a great HTTP reference. seriously though the book seems to rehash the same stuff over and over, ch8 is more of a summary of that same stuff.
Re:After reading this URL, I long for a Linus anti-REST email by andrewbadera on Friday July 09, 2010 @03:57PM (#1502007)
"base" URI may consist of host+port+path not _just_ host or host+port. you don't have to switch subdomains or hosts in order to provide RESTful interfaces.
Re:After reading this URL, I long for a Linus anti-REST email by callahad on Friday July 09, 2010 @03:58PM (#1502016)
But the git web UI doesn't have a path component. As far as I can tell, it's always just '/' with a huge querystring on the end.
Re:After reading this URL, I long for a Linus anti-REST email by xenoterracide on Friday July 09, 2010 @04:06PM (#1502049)
it's a client to the restful interface... web browsers current suck at restful interfaces because forms don't support all the http methods, and you can use your own image for a button (or other form element).