- From: Stefan Eissing <stefan.eissing@greenbytes.de>
- Date: Fri, 22 Mar 2002 13:22:16 +0100
- To: <w3c-dist-auth@w3.org>
Am Donnerstag den, 21. März 2002, um 20:16, schrieb Larry Masinter: > I'm reviewing the HTTP spec and the Cache-Control: no-transform > directive; RFC 2616 section 14.9.5. > > To prevent transformations of authored content, servers could > respond, when the client is "authoring", with a no-transform. That would set us back to sending Translate headers. > But I'm less sure of the utility of the Cache-Control: no-transform > request; does it mean "don't transform the request" or does > it mean "don't transform the response?". And would it apply > to an origin server as another way of saying a kind of "translate"? Hmm, 2616 is only speaking of content (and Content- headers). The examples are with server responses. I think modifying client requests is a no-no. Otherwise I hope Internet Explorer sets "no-transform" when I'm online shopping... > Note that authoring clients should be sensitive to > Warning 214 (Transformation applied). At least, then > there would be some belief that they wouldn't inadvertently > try to let the user edit something that wasn't the original > source. Everyday you learn some new HTTP Headers...;) > I'm not sure how a DAV server would know whether to supply > a cache-control: no-transform if clients distinguish between > the resource and its source by the "DAV:source" property. Can a "source" resources (listed in some other resource's DAV:source property) have own DAV:source properties? Must a resource designated as "source" be editable (e.g. support PUT) and will that PUT influence the content of its derived resource? When? I think we need clarification of the "source" concept. When do resources have a "source" relation? - ASP/JSP pages - shtml - executables, object files and source code - resources derived from a "master" with stylesheets - ... Which of these examples do we want to provide a solution for with DAV:source? > (I was also wondering about the interaction between > cache-control: no-transform and the vary header. A response > with no-transform and one without no-transform are different. > So if the decision to supply no-transform depends on a request > header, then should the result Vary by that request header? I'm not really a fan of new HTTP methods, but just for interest: Wouldn't a new method like "EDIT" which retrieves the editable content of a resource make our life easier? It could respond with a Content-Location where a PUT can be applied... //Stefan
Received on Friday, 22 March 2002 07:22:36 UTC