James M Snell wrote: > Yeah, you're right, it's junk ;-). However, so is the current stuff in > the spec as all it does is echo info I can already find in RFC2616 > without telling me anything useful about what details I can expect the > APP server implementation to give me about an error situation (beyond > the status code).
The current spec text says what needs to be said re HTTP response codes, but maybe doesn't need to say anything more about response classes. Otherwise... if there should be APP errors or responses that are not HTTP responses (I doubt it) or if the APP needs human readable augments to responses (possibly), that needs to be specified separately. > How's this instead: > > HTTP defines different response status codes to indicate the > success or failure of an operation. > Consult the HTTP specification [RFC2616] for detailed definitions > of each status code. +1 > Entities contained within HTTP 4xx and 5xx > responses to APP operations SHOULD include a > human readable, natural language explanation of the error > situation. 0 - because I don't know; but why should they include more text? > A standardized format for > the reporting of APP specific error details is not defined." -1 - this last bit assumes APP specific error details exist, for which no format is defined. cheers Bill
