Hi Job, I do understand your logic here, and I don't disagree with it taken by itself; however, I'm not sure YANG development in the IETF or at least the rtg-area can survive this requirement either -- we're barely relevant as it is b/c we take so long to publish these modules.
The main value I have seen in the field is to have well considered and constructed models developed by experts that vendors look to for guidance in creating their own vendor specific versions -- that's mostly how I've used them in FRR at least, although I did implement the keychain module verbatim :) To your point why not just publish experimental then; maybe we should. Will that dissuade the few people willing to work on YANG modules from doing so? I don't know, it may. In LSR we've made an effort and some good progress towards publishing YANG inline with the feature development, which I think has some value in itself. It would be a shame if we lost that. Perhaps this should really be a discussion that happens more generally in netmod/ops area, and guidance or requirements then provided based on the that. I do think that discussion should consider that certain requirements may also adversely affect what little YANG development and relevance we do have in the IETF. Thanks, Chris. > On Apr 1, 2026, at 07:03, Acee Lindem <[email protected]> wrote: > > > >> On Apr 1, 2026, at 5:13 AM, Job Snijders <[email protected]> wrote: >> >> Dear Acee, >> >> On Fri, Mar 27, 2026 at 07:08:34PM -0400, Acee Lindem wrote: >>>> The document shepherd write-up is deficient. In answer to question 4 (For >>>> protocol documents, are there existing implementations of the contents of >>>> the >>>> document? ) it says "N/A". But this is a Standards Track protocol document. >>>> YANG models are implementable and it would give significant credence to the >>>> completeness of this specification if this question had been answered with >>>> implementation details. (Of course, it would have been even better to see >>>> an >>>> Implementation Status section in the document.) >>>> >>>> It remains up to the WG and AD whether to pursue publication of an >>>> unimplemented YANG model. >>> >>> I'm not sure where you've been over the last ten years, but the IETF >>> YANG models have not been widely implemented. In LSR, we have chosen to >>> publish them without requiring implementation. >> >> It is concerning to read this practise has been going on for many years >> already, however it being a tradition does not equate it being a good >> practise. >> >> How does the WG know that the models are implementable, or even >> operationally relevant? >> >> The problem I see is that the LSR WG might take up review/editorial/IESG >> resources with their requests for RFC publication for documents that >> later on are discovered to be unimplementable works of fiction. What >> exactly is the purpose of rough consensus without running code? >> >> If the goal is to positively inspire vendors to copy parts of these >> unimplemented models, wouldn't "Experimental" fullfill the same purpose >> just as well? >> >> Why is Standards track considered appropriate for such documents? > > Independent of whether the IGP YANG models are implemented, they provide a > very useful > reference for routing protocols. Standardization of the configuration and > operational state > is essential to advance the IGPs and serves as a reference model for protocol > implementation. > > I've got much more important things to worry about than this. If you would > like to contribute > To the IETF YANG model implementation, Renato Westphal has implemented the > base IGP > models in his open-source implementation. > > https://github.com/holo-routing/holo/tree/master/holo-ospf/src/northbound > https://github.com/holo-routing/holo/tree/master/holo-isis/src/northbound > > Thanks, > Acee > > > > > > > >> >> Kind regards, >> >> Job _______________________________________________ Lsr mailing list -- [email protected] To unsubscribe send an email to [email protected]
