Re: quota-03 spec review, was: I-D ACTION:draft-ietf-webdav-quota-03.txt

Lisa Dusseault wrote:

> I do object (but not to the point of delaying quota).  My reasoning is 
> that using the negative, to express the condition that must be met, is 
> hard to understand for the debugger/implementor/tester or support 
> person.  You see something like this:
> 
> <D:error>
>   <D:quota-not-exceeded/>
> </D:error>
> 
> and think "Huh?  My error is that quota is not exceeded? what's up with 
> that? "  I can even imagine bugs logged against implementors who 
> correctly follow the spec.

On the other hand, implementors may be accustomed with the terminology 
used by RFC3253, RFC3648 and RFC3744, and become confused by a new spec 
being inconsistent with it.

Also, people seeing DAV:error (and not already familiar with RFC3253) 
will hopefully find the specification, which will point them to RFC3253, 
section 1.6, which explains it all.

If server implementors want to make it super-readable, they can still 
add additional programmer-readable prose inside an XML comment in the 
response body.

> Furthermore, although this approach is consistent with the style in 
> DeltaV, it's not consistent with the overall HTTP error style, which is 
> to explain the error.  E.g. "404 Not Found" rather than "404 Document 
> Must Exist".

That's true, but I don't think it matters today. This is what RFC3253 
defines, and unless there's a *really* good reason to be inconsistent 
with that, we should stick with it.

Best regards, Julian

-- 
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760

Received on Thursday, 2 September 2004 18:57:50 UTC