- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Wed, 14 Apr 2004 15:54:22 -0700
- To: "Thompson, Bryan B." <BRYAN.B.THOMPSON@saic.com>
- Cc: www-tag@w3.org
On Wednesday, April 14, 2004, at 07:01 AM, Thompson, Bryan B. wrote: > Could you expand on your comment concerning creating an implementation > dependency between the client and the server? > > You wrote: > >> However, personally, I would avoid use of anything other than id's, >> since >> manipulation of the XML data tree will create an implementation >> dependency >> between client and server. Sure. Let's say that we come up with a new data format standard, TBTSSB, which is so incredibly better than XML that everyone wants to switch. If past references to those resources are via assignable names, rather than data structure, then nothing needs to break in the transition. We simply make sure that the new representation also defines those names. In contrast, if the clients are assuming that the server represents the resource as XML, then moving away from XML will cause them to break, and no amount of content negotiation can help that situation because the addressing mechanism itself has become format-specific. Note that, in spite of the fact that they are often called URLs and frequently considered "addresses", every bit of indentifying information in an http URI is an assignable name. That is, until someone decides to use a fragment identifier that is format-specific. ....Roy
Received on Wednesday, 14 April 2004 18:54:28 UTC