Re: clarify some MUST requirements in HTTPbis part 1 section 3.3

On 07/12/2011, at 9:41 PM, Amos Jeffries wrote:

> Would it make sense to specify that proxies MUST (SHOULD? "ought to"?) fix headers which they pay attention to and identify as invalid?
> That leaves a huge loophole that they can ignore a header entirely for both usage and relay purposes without being in violation. But if they are smart enough to make use of it and validate they become part of the interoperability correction mechanism for HTTP.
> IMHO something stronger than MAY would be useful to encourage in-transit corrections toward validity. The process of erasing invalid headers pushes the valid formats more into view of those who dont read the specs.

I don't know if I like this. One implementation's "correction" is another's loss of information...

Mark Nottingham

Received on Thursday, 8 December 2011 01:03:38 UTC