Please find one follow-up inline, for completeness and correctness only. > On Jun 11, 2025, at 9:48 PM, Michael Richardson <[email protected]> wrote: > > > Carlos Pignataro <[email protected]> wrote: >> What is the incentive to request a Spec Required vs. an FCFS? Do you >> expect someone will ask for the more difficult choice, when they get >> exactly the same result? > > 1. It's not the same result. > > The entity that writes a specification winds up with a Specification. > That can be cited (like, in anotheer RFC), can be placed into RFPs (yes, > trade agreements call them "performance standards"), and can be discovered by > others. > > > 2. Aside from writing the specification, which is certainly more work, it's > essentially the same effort to get the number.
I am that entity. I write a specification. I complete the template: LinkType Name: LINKTYPE_CARLOS LinkType Value: TBA Description: Carlos protocol. Reference: URI to my specification of the Carlos protocol. I send a request asking for an FCFS value. Get an assignment without issues. Done. Had I requested a Spec Required value, the expert might have pushed back with potential interop and stability concerns for the spec and the pointer. Why would I do that? So, I am *not* writing a spec to get an assignment. I wrote a spec to have implementations. Since it’s the same work to ask for FCFS than Spec Required, would anyone not ask for FCFS? > >>>> 7. Shall there be a separate range for existing ones (<= 301) instead >>>> of lumping them in FCFS, since some have specification, etc? >>> >>> FCFS space will have uneven amount of specification, which historically we >>> have had. > >> Linking to the first comment, this would further encourage allocation >> requests with specifications to ask for FCFS… > > Worse case, in ten years, the FCFS is full. (Maybe full of junk) > And we decide to allocate another 10K numbers to it. Future me problem. I did not say it was a problem. If this is the perspective, and as you mentioned, are interested in not overthinking, the proposal does not seem to provide guidance on why or benefits of Spec Required, nor prevents from attacks to the registry. I was hesitant to continue this discussion — please treat these comments as thoughts for your consideration, intended to provide a perspective and not to micromanage, and do whatever you want with them. Thanks, and very best, Carlos. > > -- > Michael Richardson <[email protected]> . o O ( IPv6 IøT consulting ) > Sandelman Software Works Inc, Ottawa and Worldwide > > > >
signature.asc
Description: Message signed with OpenPGP
_______________________________________________ OPSAWG mailing list -- [email protected] To unsubscribe send an email to [email protected]
