On 26/1/05 2:16 AM, "Anne van Kesteren" <[EMAIL PROTECTED]> wrote:
> I like the proposal in general though I wonder why an absolute URI is > required. One of the intended uses is for when a browser downloads a feed resource to disk, and then hands that file to an atom handler application. Once that happens, the original location context is lost. > Since aggregators already have to support xml:base it would be logical > to make that a relative URI. A relative URI would only be useful if there is in fact an xml:base in effect. With no xml:base then the relative reference would be useless. We could either write some spec text for that unusual requirement, or we could write "MUST be absolute". Simplicity. e.