On Mon, Oct 12, 2009 at 4:43 PM, Simon Laws <[email protected]> wrote: > ...snip >> BTW, what's your goal here by adding these messages to the error >> messages ? Maybe one thing that could help us make the decision is to >> check what other frameworks usually do ? If you are using a Java EE or >> WS spec, do they show messages with spec related error codes ? >> > > Well I came across this while fixing the error messages for the otests > which I'm running through at the moment. I'm actually just asking that > we be consistent. > > As it happens I think there is some value in referencing back into the > specs as it gives the user a chance to see why the particular error > they are dealing with was defined. This will not be the case for all > errors of course. Just for those arriving out of compliance > statements. > > Of the top of my had I don't know what other frameworks do. A quick > glance at a couple didn't show them using spec references. However > that doesn't mean other frameworks don't. I think we should make our > own minds up. > > Regards > > Simon >
I think I agree. When I read the original email I took it to mean "if there's going to be a number in an error message then lets use the compliance numbers from the spec instead of the test number from the testing doc". That seems fine to me and it makes sense to try to be consistent so its obvious where the number is defined. There's still going to lots of Tuscany messages with no number, but if we do happen to know a compliance number associated with a message i don't see a problem with mentioning it in the message. ...ant
