Re: [v6ops] A common problem with SLAAC in "renumbering" scenarios

[Jan Zorz - Go6 <[e-mail]>]

On 23/02/2019 01:25, Brian E Carpenter wrote:
> On 2019-02-23 09:36, Jan Zorz - Go6 wrote:
>> On 21/02/2019 22:14, Brian E Carpenter wrote:
>>> Of course, SHOULD means "that there
>>> may exist valid reasons in particular circumstances to ignore a
>>> particular item, but the full implications must be understood and
>>> carefully weighed...". In other words, do RFC8028 and Rule 5.5
>>> unless there's a good reason not to.
>>
>> Unfortunately not all vendors understand SHOULD in this way.
>> Implementing more options is expensive and lowers the profits, so it's
>> quite easy to fall into "implement all the MUSTS, declare compliancy
>> with RFCs and ship the product" trap.
> 
> Sure. I dealt with a lot of development managers during my years
> at a 3-letter company, so this error in thinking is quite familiar
> to me. But there is nothing the IETF can really do about that.

Actually, there is.

If we feel that something must be implemented in end product, than we 
should not be polite about it and use a MUST :) That's the quickest way 
to get things back on track.

Cheers, Jan