- From: Yaron Goland <yarong@Exchange.Microsoft.com>
- Date: Thu, 10 Feb 2000 22:33:45 -0800
- To: "'w3c-dist-auth@w3.org'" <w3c-dist-auth@w3.org>
- Message-ID: <7DE119D3D0E15543874F7561EECBDBED0261A15E@BEG.platinum.corp.microsoft.com>
The redirect spec repeatedly refers to servers, e.g. in the last sentence of the third paragraph and the last sentence of the second to last paragraph of the introduction. In general the term server is misleading because it focuses the user's attention on the box. For example, a reader of the redirect spec would most likely come away thinking that if two resources are on the same server then one doesn't need to link them but can use some other feature, maybe bind, since being on the same box somehow confers special relationships. As anyone who has used files across volumes knows, being on the same servers doesn't mean much of anything. In addition the reader is likely to believe that bind and similar features won't work at all across two different servers when advanced systems may be able to do this. The idea of a server has its place for certain HTTP transport issues but should never be mentioned in an application context. The actual physical box that a resource resides on is largely irrelevant in terms of what features are available between two resources. As such I move that the term "server" be universally replaced with phrase "unrelated systems".
Received on Friday, 11 February 2000 01:36:38 UTC