Hi, now I like to try one of Mark's favourite arguments myself... :-) Christian Kratzer wrote: >I don't think we need a nother pleasenop extension or similar >though. Educating client vendors to NOOP often enough and >getting server vendors to sends expunge responses at the >earliest possible time is propably time much better spent.
In a perfect world, you are right: All client implementors read all relevant RFCs and dig into short paragraphs, fully understanding their implications. In reality though, some will only read (and implement) the core specs and never get to some sub-paragraph of "multi-accessed blablah" because time is money and time spent reading is nonsense :-) As an extension like "[NOOPFIRST]" or similar can only be mean something like "a client SHOULD noop and try again" (but cannot be a MUST so it does not break anything), it falls more into a clarification than a real extension. It does not bother any "perfect world" implementors, but may make "real world" implementors have a look at it. Christof
