Andrew, On 30-Jul-19 15:04, Andrew Sullivan wrote: > Hi, > > As ever, I do not speak for the Internet Society though I am employed > by it. I'm posting this note especially because I have certain > experiences not widely shared: I was all of an IAB member, an IAB > chair, an IAOC member, and an IAOC chair during some portion of the > past where various bits of RFC 6635 applied. > > On Mon, Jul 29, 2019 at 07:15:10PM -0400, Donald Eastlake wrote: > >> The key word in RFC Series Oversight Committee is "Oversight". What do >> people think when they hear "oversight"? They think that a large part >> the job of whoever has "oversight" is to review and criticize. > > I don't think that. What I think is that it is the responsibility of > the oversight body to ensure that the thing to be overseen is > accomplished according to some conditions. I'd like to think that the > conditions are well-operationalized so that people are in a position > to know about this. > > The fundamental basis of oversight is the ability to decide whether a > given overseen thing is or is not adequately done. At a high level, > this generally works out to "hire and fire" capability. (Note you can > do it other ways. Gating-function, for instance, could do this, but > it seems unwise to implement.) > > That capability with respect to the RFC Series Editor goes back at > least to RFC 2850, which says, "The IAB must approve the appointment > of an organization to act as RFC Editor and the general policy > followed by the RFC Editor." This is of course far less concrete than > the full blown-out description found in RFC 6635. But the basic rule > is already there. No, I really don't agree. I was the editor of RFC 2850. It uses the term "oversight" explicitly in the phrases "Architectural Oversight" and "Standards Process Oversight and Appeal", but not in reference to the RFC Editor. It uses a quite different formulation: "The IAB must approve the appointment of an organization to act as RFC Editor and the general policy followed by the RFC Editor." This was not an accident of drafting. (At that time, the organization was ISI and the general policy was created by Bob Braden and Joyce Reynolds, but we didn't expect that to last for ever, and it didn't.) I now believe that the reformulation including "oversight" in RFC5620, updated in RFC6635, was a collective error, which I certainly missed at the time. Particularly, the phrase "The IAB retains its oversight role..." in RFC5620 was inappropriate IMHO, because there was previously no such formal oversight role to retain. Would you say that the IAB, whose membership is approved by the ISOC Board, is therefore under the oversight of ISOC? I think we got this wrong and now we are paying the price. IMHO the long term fix will be an "RFC Editor Model (Version 3)" that underlines and guarantees the independence of the RFC Series and its Editor. To be clear, of course the production and publication contracts need proper contract management, but that is more of an IETF LLC role. >> What if everything else we the same, but it had been called the RFC >> Series Support Committee? And everytime someone thought about or >> volunteer for or was appointed to the committee they were reminded >> that this is about supporting the RFC Series? > > I don't know what the world would be like in the case of > terminological change like you propose. But I would like to suppose > that everyone involved in the IAB, at least, and anyone I (at least) > ever asked to be on the RSOC regarded their role as making the RFC > series successful. I'm sure that's correct. > And making the Editor's general policy cohere with > the IAB's agreement is no innovation from 6635: it's right there in > 2850. It was also in RFC 1601 and RFC 1358. I am not too sure that > the responsibility of the IAB for this issue is grounded in anything > earlier than 1358, but I'm also not sure that a responsibility that > has been repeatedly affirmed in print since 1992 is one that we can > assume is inoperative. I still don't see how approving a policy morphed into oversight, which seems to have morphed further in recent times. By telling a group of people that they had oversight, we unintentionally set the scene for things to go wrong. > Elsewhere, Mike StJohns has claimed, "This is a senior person who > really should be co-equal with the IAB and IESG." I do not find the > documented tradition that suggests this is true. On the contrary, I > can find documents stretching back to at least 1992 (where I stopped > digging) suggesting that the RSE is in fact subordinate to the IAB. Well, that is obviously a matter of interepretation, but my preferred phrasing is that the relationship should be "collegial". (https://www.merriam-webster.com/dictionary/collegial) That matches what I observed most of the time until very recently. > That is not to suggest the relationship is some sort of > directive-management one. In my current job, I have plenty of > colleagues who know more about their area than I do (i.e. all of > them), yet I am responsible for their direction and in this formal > sense they are "subordinate" to me. If any of them messes up, they > are not responsible to my board: I am. Co-equal suggests that perhaps > the RSE ought to be picked by nomcom. I'm not too sure that is > desirable. Since the skill set is not one we commonly find in our own community, I agree with you there. A search committee model seems more appropriate. Regards Brian
