[SIPForum-techwg] (Interop) SHOULD in RFCs...

[Peter Musgrave]

Hi, 

 

As part of trying to define specific interop requirements for use-case I
decided to start with looking at the SHOULD clauses in the RFCs from
Section (6). Past experience at SIPITs suggest to me that these clauses
cause a number of interop issues, in part because a number of parties
interpret a SHOULD as MAY....

 

To remind myself SHOULD means...

   may exist valid reasons in particular circumstances to ignore a
   particular item, but the full implications must be understood and
   carefully weighed before choosing a different course.

 

So how many SHOULDs are there?

RFC2246 - 0

RFC2283 - 6

RFC2782 - 8

RFC3261 - 276

RFC3262 - 9

RFC3263 - 16

RFC3264 - 27

RFC3311 - 8

RFC3323 - 23

RFC3324 - 0

RFC3325 - 2

RFC3489 - 26

RFC3581 - 5

RFC3725 - 19

RFC4028 - 18

 

So the winner by far is RFC3261 (which is also the longest spec). 

 

This represents a fairly large degree of freedom in establishing
interop. 

 

Should we be more specific in listing the requirements for interop? If
we *really* want guaranteed interop then every SHOULD which will impact
interop will need to be resolved. 

 

Do we have the collective appetite for this?

 

Peter

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://sipforum.org/pipermail/techwg/attachments/20070423/33e2eaf6/attachment-0001.html