John,
I'll remove the MUST.
thanks,
Peter
On 08/06/2023 15:05, John Scudder wrote:
Hi Peter and all,
On Jun 8, 2023, at 2:43 AM, Peter Psenak <[email protected]>
wrote:
A node MUST participate in a Flex-Algorithm to be:
- Able to compute path for such Flex-Algorithm
- Part of the topology for such Flex-Algorithm
This is an odd use of MUST.
what exactly is odd in it?
I don’t know what Paul found odd about it, but now that I’m looking at it
afresh, I also think it’s odd. My reason is that it’s not expressing a
requirement, it’s expressing a statement of fact, a natural consequence. An
analogy would be if we said “if you let go of your coffee cup as you’re lifting
it, it MUST fall down”. You don’t need a MUST there, gravity doesn’t care about
your rules. To continue the analogy, a more usual use of MUST would be to
express an actual requirement on the implementor — “you MUST NOT drink your
coffee through a straw”.
I haven’t gone and re-checked in the doc to be sure, but as I recall, it’s not
possible for a node to be part of the topology for a given FA unless it
participates in the FA, and this would be true whether the quoted MUST were
there or not.
$0.02,
—John
_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr
