idnits 2.17.1 draft-resnick-on-consensus-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (September 29, 2013) is 3856 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'IETF24' is defined on line 600, but no explicit reference was found in the text Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force P. Resnick 3 Internet-Draft Qualcomm Technologies, Inc. 4 Intended status: Informational September 29, 2013 5 Expires: April 02, 2014 7 On Consensus and Humming in the IETF 8 draft-resnick-on-consensus-03 10 Abstract 12 The IETF has had a long tradition of doing its technical work through 13 a consensus process, taking into account the different views among 14 IETF participants and coming to (at least rough) consensus on 15 technical matters. In particular, the IETF is supposed not to be run 16 by a "majority rule" philosophy. This is why we engage in rituals 17 like "humming" instead of voting. However, more and more of our 18 actions are now indistinguishable from voting, and quite often we are 19 letting the majority win the day, without consideration of minority 20 concerns. This document is a collection of thoughts on what rough 21 consensus is, how we have gotten away from it, and the things we can 22 do in order to really achieve rough consensus. 24 Note: This document contains the musings of an individual. Right 25 now, it still likely has many holes that need to be filled in. 26 Even if those holes are filled, it's not clear that it should be 27 published as an RFC, and certainly is not intended in its current 28 form to be a BCP for a change of IETF policy. If it evolves into 29 such a thing, great. If it simply sparks discussion as an 30 Internet Draft, that's a perfectly fine outcome. 32 Status of This Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at http://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on April 02, 2014. 49 Copyright Notice 51 Copyright (c) 2013 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents 56 (http://trustee.ietf.org/license-info) in effect on the date of 57 publication of this document. Please review these documents 58 carefully, as they describe your rights and restrictions with respect 59 to this document. Code Components extracted from this document must 60 include Simplified BSD License text as described in Section 4.e of 61 the Trust Legal Provisions and are provided without warranty as 62 described in the Simplified BSD License. 64 Table of Contents 66 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 67 2. Lack of disagreement is more important than agreement . . . . 3 68 3. Rough consensus is achieved when all issues are addressed, 69 but not necessarily accommodated . . . . . . . . . . . . . . 5 70 4. Humming should be the start of a conversation, not the end . 7 71 5. Consensus is the path, not the destination . . . . . . . . . 9 72 6. One Hundred people for and five people against might not be 73 rough consensus . . . . . . . . . . . . . . . . . . . . . . . 10 74 7. Five people for and one hundred people against might still be 75 rough consensus . . . . . . . . . . . . . . . . . . . . . . . 11 76 8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 13 77 9. Security Considerations . . . . . . . . . . . . . . . . . . . 13 78 10. Informative References . . . . . . . . . . . . . . . . . . . 13 79 Appendix A. Acknowledgements . . . . . . . . . . . . . . . . . . 14 80 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 14 82 1. Introduction 84 Almost every IETF participant knows the aphorism from Dave Clark's 85 1992 plenary presentation [Clark] regarding how we make decisions in 86 the IETF: 88 We reject: kings, presidents and voting. 90 We believe in: rough consensus and running code. 92 That is, our credo is that we don't let a single individual make the 93 decisions, nor do we let the majority dictate decisions, nor do we 94 allow decisions to be made in a vacuum without practical experience. 95 Instead, decisions are made by (more or less) consent of all 96 participants, and the actual products of engineering trump 97 theoretical designs. Our ideal is full consensus, but we don't 98 require it: Full consensus would allow a single intransigent person 99 who simply keeps saying "No!" to stop the process cold. We only 100 require rough consensus: If the chair of a working group determines 101 that a technical issue brought forward by an objector has been truly 102 considered by the working group, and the working group has made an 103 informed decision that the objection has been answered or is not 104 enough of a technical problem to prevent moving forward, the chair 105 can declare that there is rough consensus to go forward, the 106 objection notwithstanding. To reinforce that we do not vote, we have 107 also adopted the tradition of "humming": When (for example) the chair 108 of the working group wants to get a "sense of the room", instead of a 109 show of hands, the chair asks for each side to hum for or against a 110 question. 112 However, in recent years we have seen participants (including folks 113 in IETF leadership) who do not understand some of the subtleties of 114 consensus-based decision making. Participants ask, "Why are we 115 bothering with this 'humming' thing? Wouldn't a show of hands be 116 easier? That way we could really see how many people want one thing 117 over another." Chairs are faced with factious working groups with 118 polarized viewpoints and long-running unresolved issues that return 119 again and again to the agenda. More and more frequently, people walk 120 away from working groups, thinking that "consensus" has created a 121 document with horrible compromises to satisfy everyone's pet peeve 122 instead of doing "the right thing". None of these things are 123 indicators of a rough consensus process being used, and are likely 124 due to some basic misperceptions. 126 This document attempts to explain some features of rough consensus, 127 explain what is not rough consensus, and suggest ways that we might 128 achieve rough consensus and judge it in the IETF. Though this 129 document describes some behaviors of working groups and chairs, it 130 does so in broad brushstrokes and it is not intended to dictate 131 specific procedures. It is intended to foster understanding of the 132 underlying principles of IETF consensus processes. 134 Note: This document contains the musings of an individual. Right 135 now, it still likely has many holes that need to be filled in. 136 Even if those holes are filled, it's not clear that it should be 137 published as an RFC, and certainly is not intended in its current 138 form to be a BCP for a change of IETF policy. If it evolves into 139 such a thing, great. If it simply sparks discussion as an 140 Internet Draft, that's a perfectly fine outcome. 142 2. Lack of disagreement is more important than agreement 143 A working group comes to a technical question of whether to use 144 format A or format B for a particular data structure. The chair 145 notices that a number of experienced people think format A is a good 146 choice. The chair asks on the mailing, "Is everyone OK with format 147 A?" Inevitably, a number of people object to format A for one or 148 another technical reason. The chair then says, "It sounds like we 149 don't have consensus to use format A. Is everyone OK with format B?" 150 This time even more people object to format B, on different technical 151 grounds. The chair, not having agreement on either format A or 152 format B, is left perplexed, thinking the working group has 153 deadlocked. 155 The problem that the chair got themselves into in the above case was 156 thinking that what they were searching for was agreement. "After 157 all", thinks the chair, "consensus is a matter of getting everyone to 158 agree, so asking whether everyone agrees is what the chair ought to 159 do. And if lots of people disagree, there's no consensus." But 160 _determining_ consensus and _coming to_ consensus are different 161 things than _having_ consensus. Consensus is not when everyone is 162 happy and agrees that the chosen solution is the best one. Consensus 163 is when everyone is sufficiently satisfied with the chosen solution, 164 such that they no longer have specific objections to it. The 165 distinction might be a bit subtle, but it's important. Engineering 166 always involves a set of tradeoffs. It is almost certain that any 167 time engineering choices need to be made, there will be options that 168 appeal to some people that are not appealing to some others. The key 169 is to separate those choices that are simply unappealing from those 170 that are truly problematic. So in the case of a working group 171 decision, it is most important to ask not just for objections to a 172 particular proposal, but for the nature of those objections. A chair 173 who asks, "Is everyone OK with choice A?" is going to get objections. 174 But a chair who asks, "Can anyone not live with choice A?" is more 175 likely to only hear from folks who think that choice A is impossible 176 to engineer given some constraints. Following up with, "What are the 177 reasons you object to choice A?" is also essential. Then the 178 purported failings of the choice can be examined by the working 179 group. The objector might convince the rest of the group that the 180 objections are valid, or the working group might convince the 181 objector that the concerns can be addressed, or that the choice is 182 simply unappealing (i.e., something the objector can "live with") and 183 not a show-stopper. In any event, closure is much more likely to be 184 achieved quickly by asking for and trying to accommodate the 185 objections rather than asking for agreement. 187 This also brings up an important point about reaching consensus and 188 "compromising": Unfortunately, the word "compromise" gets used in two 189 different ways, and though one sort of compromising to come to 190 consensus is good (and important), the other sort of compromising in 191 order to achieve consensus is actually harmful. As mentioned 192 earlier, engineering always involves balancing tradeoffs, and 193 figuring out whether one engineering decision makes more sense on 194 balance compared to another involves making engineering 195 "compromises": We might have to compromise processor speed for lower 196 power consumption, or compromise throughput for congestion 197 resistance. Those sorts of compromises are among engineering 198 choices, and they are expected and essential. We always want to be 199 weighing tradeoffs and collectively choosing the best set. However, 200 there is another sense of "compromise" that involves compromising 201 between people, not engineering principles. For example, a minority 202 of a group might object to a particular proposal, and even after 203 discussion still think the proposal is deeply problematic, but decide 204 that they don't have the energy to argue for it and say, "Forget it, 205 do what you want". That surely can be called a compromise, but a 206 chair might mistakenly take this to mean that they agree, and have 207 therefore come to consensus. But really all that they've done is 208 conceded; they've simply given up by trying to appease the others. 209 That's not coming to consensus; there still exists an outstanding 210 unaddressed objection. Even worse is the "horse-trading" sort of 211 compromise: "I object to your proposal for such-and-so reasons. You 212 object to my proposal for this-and-that reason. Neither of us agree. 213 If you stop objecting to my proposal, I'll stop objecting to your 214 proposal and we'll put them both in." That again results in an 215 "agreement" of sorts, but it also results in two outstanding 216 unaddressed issues, again ignoring them for the sake of expedience. 217 These sorts of "giving up" or "horse-trading" compromises have no 218 place in consensus decision making. In each case, a chair who looks 219 for "agreement" might find it in these examples because it appears 220 that people have "agreed". But again, answering technical 221 disagreements is what is needed to achieve consensus, sometimes even 222 when the people who stated the disagreements no longer wish to 223 discuss them. 225 Coming to consensus is when everyone comes to the conclusion that 226 either the objections are valid (and therefore making a change to 227 address the objection) or that the objection was not really a matter 228 of importance, but merely a matter of taste. Of course, coming to 229 full consensus like that does not always happen. That's why in the 230 IETF, we talk about "rough consensus". 232 3. Rough consensus is achieved when all issues are addressed, but not 233 necessarily accommodated 235 The preceding discussion gives an example where the working group 236 comes to consensus on a point: Either the objector is satisfied, or 237 the working group is satisfied. But that doesn't happen all of the 238 time, and it's certainly not the problematic case. Engineering is 239 always a set of tradeoffs. Often, a working group will encounter an 240 objection where everyone understands the issue and acknowledges that 241 it is a real shortcoming in the proposed solution, but the vast 242 majority of the working group believe that accommodating the 243 objection is not worth the tradeoff of fixing the problem. 245 So, an objector might say, "The proposal to go with protocol X is 246 much more complicated than going with protocol Y. Protocol Y is a 247 much more elegant and clean solution, which I can code much more 248 easily, and protocol X is a hack." The working group might consider 249 this input, and someone might respond, "But we have a great deal of 250 code already written that is similar to protocol X. While I agree 251 that protocol Y is more elegant, the risks to interoperability with 252 an untested solution is not worth it compared to the advantages of 253 going with the well-understood protocol X." If the chair finds, in 254 their technical judgement, that the issue has truly been considered, 255 and that the vast majority of the working group has come to the 256 conclusion that the tradeoff is worth making, even in the face of 257 continued objection from the person(s) who raised the issue, the 258 chair can declare that the group has come to rough consensus. 260 Note that this portrays rough consensus as an "exception". In one 261 sense, it is: As a working group does its work and makes its choices, 262 it behaves as if it is striving toward full consensus and tries to 263 get all issues addressed to the satisfaction of everyone in the 264 group, even those who originally held objections. It treats rough 265 consensus as a sort of "exception processing", to deal with cases 266 where the person objecting still feels strongly that their objection 267 is valid and must be accommodated. But it is certainly true that 268 more often than not in the IETF, at least someone in the group will 269 be unsatisfied with a particular decision. In that sense, rough 270 consensus might be closer to the norm than the "exception". However, 271 when a participant says, "That's not my favorite solution, but I can 272 live with it; I'm satisfied that we've made a reasonable choice", 273 that participant is not in the "rough" part of a rough consensus; the 274 group actually reached consensus if that person is satisfied with the 275 outcome. It's when the chair has to declare that an unsatisfied 276 person still has an open issue, but that the group has truly answered 277 the objection, that the consensus is only rough. 279 Now, a conclusion of having only rough consensus relies heavily on 280 the good judgement of the consensus caller. The group must truly 281 consider and weigh an issue before the objection can be dismissed as 282 being "in the rough". The chair of the working group in one of these 283 cases is going to have to decide that not only has the working group 284 taken the objection seriously, but that it has fully examined the 285 ramifications of not making a change to accommodate it, and that the 286 outcome does constitute a failure to meet the technical requirements 287 of the work. In order to do this, the chair will need to have a good 288 idea of the purpose and architecture of the work being done and use 289 their own technical judgement to make sure that the solution meets 290 those requirements. What can't happen is that the chair bases their 291 decision solely on hearing a large number of voices simply saying, 292 "The objection isn't valid." That would simply be to take a vote. A 293 valid justification needs to me made. 295 Any finding of rough consensus needs, at some level, to provide a 296 reasoned explanation to the person(s) raising the issue of why their 297 concern is not going to be accommodated. A good outcome is for the 298 objector to understand the decision taken and accept the outcome, 299 even though their particular issue is not being accommodated in the 300 final product. Remember, if the objector feels that the issue is so 301 essential that it must be attended to, they always have the option to 302 file an appeal. A technical error is always a valid basis for an 303 appeal, and a chair or AD has the freedom and the responsibility to 304 say, "The group did not take this technical issue into proper 305 account." Simply having a large majority of people agreeing to 306 dismiss an objection is not enough to claim there is rough consensus; 307 the group must have honestly considered the objection and evaluated 308 that other issues weighed sufficiently against it. Failure to do 309 that reasoning and evaluating means that there is no true consensus. 311 4. Humming should be the start of a conversation, not the end 313 We don't vote in the IETF. In some ways, we can't vote: Since the 314 IETF is not a membership organization, it's nearly impossible to 315 figure out who would get a vote for any given question. We don't 316 know who the members of any given working group are at any one time, 317 and we certainly don't know who all of the members of the IETF are. 318 So voting is not practical. We've also decided that coming to 319 consensus (albeit sometimes rough consensus) is an important thing to 320 do. So instead of voting, we "hum". 322 However, more and more we see people who think that a hum is a sort 323 of anonymous vote, with some chairs calling every question they have 324 for the working group by asking for a hum and judging the result by 325 the loudest hum, even saying things like, "There were lots of hums 326 for choice 1 and very few hums for choice 2, so it sounds like we 327 have rough consensus for choice 1." This really misses the point of 328 humming and is almost certainly mis-assessing the consensus. Hums 329 should not be used as votes. 331 So, why do we hum? The first reason is pragmatic. Quite often, a 332 chair is faced with a room full of people who seem to be 333 diametrically opposed on some choice facing the group. In order to 334 find a starting place for the conversation, it is often useful for 335 the chair to ask for a hum to see if one of the choices already has a 336 stronger base of support than the other. If choice "foo" has a 337 significant amount more support than choice "bar", it is likely 338 easier to start the discussion by saying, "OK, 'foo' seems to have 339 quite a bit of support. Let's have the people that think 'foo' is a 340 bad idea come up and tell us why it is problematic." At that point, 341 the group can start going through the issues and see if any of them 342 are showstoppers. It may turn out that one of the objections is 343 instantly recognized by the entire group as a fatal flaw in "foo" and 344 the group will then turn to a discussion of the merits (and demerits) 345 of "bar" instead. All that the hum does is give the chair a starting 346 point: There were likely to be less objections to "foo" than to "bar" 347 at the beginning of the discussion, so starting with whatever got the 348 most hums can shorten the discussion. 350 But couldn't the above could have been done with a show of hands 351 instead of a hum? Sure, but there are more formal reasons for using 352 a hum instead of a show of hands: Another reason we hum is because it 353 actually gives the chair the opportunity to take the temperature of 354 the room. A smaller bunch of loud hums for choice A and a larger 355 number of non-committal hums for choice B might indicate that some 356 people believe that there are serious problems with choice B, albeit 357 the more popular by sheer number of people. The chair might decide 358 that starting with choice A and getting objections to it is the 359 easier path forward and more likely to result in consensus in the 360 end. Remember, coming to consensus is a matter of eliminating 361 disagreements, so the chair wants to choose the path that gets to the 362 least objections fastest. A bunch of people who are not strongly 363 committed to B might have no real technical objection to A, even 364 though it is not their first preference. Taking the hum is to figure 365 out how to start the conversation. The chair could always be 366 surprised because the hum turns out to be unanimous. Otherwise, the 367 hum begins the discussion, it doesn't end it. Of course, a chair 368 could get the temperature of the room by doing a show of hands too, 369 and knowing who specifically feels one way or another can help a good 370 chair guide the subsequent conversation. However, a show of hands 371 will often leave the impression that the number of people matters in 372 some formal way. It takes a chair and a working group with a solid 373 understanding of how consensus works to do a show of hands and not 374 end up reinforcing the mistaken notion that a vote is taking place. 375 A chair can always take the hum and then later ask for specific folks 376 to identify themselves to elicit more discussion. The hum makes it 377 perfectly clear that the chair is simply figuring out the direction 378 of the conversation. 380 This also points to another misuse of the hum: If the chair is 381 already convinced that the group has come to consensus, there is no 382 reason to take a hum. In fact, taking the hum only serves to 383 discourage those who might be in the minority from voicing their 384 concerns to the group in the face of a large majority who want to 385 move forward. The right thing for the chair to do if they already 386 sense consensus is to say, "It sounds to me like we have consensus 387 for choice A. Does anybody have any concerns or objections to going 388 with A?" This allows folks to bring up issues to the group that the 389 chair might have mistakenly missed without having them feel that the 390 majority has "already spoken". 392 There are times where the result of a hum is a pretty even split. In 393 practical terms, that means it doesn't matter where the chair starts 394 the discussion. And in fact, we've had working groups where a coin- 395 flip decided which proposal to start with. That doesn't mean that 396 the coin-flip determined the outcome; if a fatal technical flaw was 397 found in the solution that won the coin flip, it is still incumbent 398 upon the group to address the issue raised, or abandon that solution 399 and find another. Rough consensus on the technical points, in the 400 end, is always required. Any way to find a place to start, be it the 401 hum or the coin-flip, is only getting to the beginning of the 402 discussion, not the end. 404 5. Consensus is the path, not the destination 406 We don't try to reach consensus in the IETF as an end in itself. We 407 use consensus-building as a tool to get to the best technical (and 408 sometimes procedural) outcome when we make decisions. Experience has 409 shown us has been that traditional voting leads to gaming of the 410 system, "compromises" of the wrong sort described earlier, important 411 minority views being ignored, and in the end worse technical 412 outcomes. Coming to consensus is what we do during the process to 413 arrive at the best solution. "Declaring" consensus is not the end 414 goal. Indeed, attempts to declare consensus at the end of a 415 discussion often get us back into the voting mentality that we're 416 trying to avoid. 418 We often hear chairs say that they are making a "consensus call". 419 Sometimes, they simply mean they are making a call _of_ the 420 consensus; that is, they are declaring the consensus that has, in 421 their view, been reached when the discussion has reached an end. 422 That's a fine thing and what chairs are supposed to do: They are 423 "calling" the consensus. Sometimes, when a chair says that they are 424 making a "consensus call", the chair is actually making a call _for 425 discussion_ of a particular point in order to reach consensus. 426 Although it's a bit odd to call that a "consensus call" (as opposed 427 to a "call for discussion" or the like), it is fine for a chair to 428 occasionally identify a particular point of contention and get the 429 group to focus discussion on it in order to reach consensus. But 430 more and more often, we hear chairs say that they are making a 431 "consensus call" at the end of a discussion, where the chair will 432 pose the classic "Who is in favor of choice A? Who is in favor of 433 choice B?" questions to the working group. That's not really a 434 "consensus call", and has the same potential problems as the "hum" at 435 the end of a discussion: It can be tantamount to asking for a vote. 436 Even talk of "confirming consensus" has this problem: It implies that 437 you can confirm that there is consensus by counting people, not 438 issues. The important thing for a chair to do is to "call consensus" 439 in the sense of declaring the consensus; others can always object and 440 say that the chair has gotten the consensus wrong and ask for 441 reconsideration. But the chair ought to be looking for consensus 442 throughout the discussion, not asking for it at the end. 444 There are some times where chairs will ask a question or take a hum 445 toward the end of a discussion in order to figure out the state of 446 consensus, but this must be done with extreme caution. This is 447 discussed in the next section. 449 6. One Hundred people for and five people against might not be rough 450 consensus 452 Remember that consensus is found when all of the objections have been 453 addressed. Because of this, using rough consensus avoids a major 454 pitfall of a straight vote: If there is a minority of folks who have 455 a valid technical objection, that objection must be dealt with before 456 consensus can be declared. This also reveals one of the great 457 strengths of using consensus over voting: It isn't possible to use 458 "vote stuffing" (simply recruiting a large number of people to 459 support a particular side, even people who have never participated in 460 a working group or the IETF at all) to change the outcome of a 461 consensus call. As long as the chair is looking for outstanding 462 technical objections and not counting heads, vote stuffing shouldn't 463 affect the outcome of the consensus call. So in a large working 464 group with over 100 active participants and broad agreement to go 465 forward with a particular protocol, if a few participants say, "This 466 protocol is going to cause congestion on the network, and it has no 467 mechanism to back off when congestion occurs; we object to going 468 forward without such a mechanism in place", and the objection is met 469 with silence on the mailing list, there is no consensus. Even if the 470 working group chair makes a working group last call on the document, 471 and 100 people actively reply and say, "This document is ready to go 472 forward", if the open issue hasn't been addressed, there's still no 473 consensus. It's the existence of the unaddressed open issue, not the 474 number of people, which is determinative in judging consensus. As 475 discussed earlier, you can have rough consensus with issues that have 476 been purposely dismissed, but not ones that have been ignored. 478 This brings us back to when a hum could be used (cautiously) at the 479 end of a discussion. A discussion may be ongoing for some time, and 480 a particular objection seems to be holding up the decision. A 481 diligent chair who's been carefully listening to the discussion might 482 think, "I have heard person X make this objection, and I've heard 483 responses from many other folks that really address the issue. I 484 think we have rough consensus. But the objection keeps coming up. 485 Maybe it's just the one person getting up again and again with the 486 same argument, but maybe we don't have rough consensus. I'm not 487 sure." At this point, the chair might ask for a hum. If only a 488 single hum objecting can be heard, even a loud one, in the face of 489 everyone else humming that the objection has been answered, the chair 490 has pretty good reason to believe that they heard the single 491 objection all along and it really has been addressed. However, to 492 say immediately after the hum, "It sounds like we have rough 493 consensus" and nothing else is at best being slipshod: What the chair 494 really needs to say at that point is, "I believe the only objection 495 we've heard is A (coming from person X), and I've heard answers from 496 the group that fully address that issue. So, unless I hear a 497 different objection than the one I've just described, I find that 498 there is rough consensus to move on." That leaves the door open for 499 someone to tell the chair that the objection was really on different 500 grounds and they misevaluated, but makes it clear that the chair has 501 found rough consensus due to the discussion, not due to the hum. 502 Again, it's not the hum that ends things, it's that the issues have 503 been addressed. If the small minority (even one person) still has an 504 issue that hasn't been addressed, rough consensus still hasn't been 505 achieved. 507 7. Five people for and one hundred people against might still be rough 508 consensus 510 This one is the real mind bender for most people, and certainly the 511 most controversial. Say there is a very small working group, one 512 with half a dozen truly active participants who are experts in the 513 field; everybody else is just following along, but not contributing 514 to the discussion. The active folks come up with a protocol document 515 that they all agree is the right way forward, and people inside and 516 outside the working group agree that the protocol is likely to get 517 widespread adoption; it is a good solution to a real problem, even if 518 the non-experts don't have the ability to fully judge the details. 520 However, one of the active members has an objection to a particular 521 section: The protocol currently uses a well-known algorithm to 522 address an issue, but the objector has a very elegant algorithm to 523 address the issue, one which works especially well on their 524 particular piece of hardware. There is some discussion, and all of 525 the other contributors say, "Yes, that is elegant, but what we're 526 using now is well-understood, widely-implemented, and it works 527 perfectly acceptably, even on the objector's hardware. There is 528 always some inherent risk to go with a new, albeit more elegant, 529 algorithm. We should stick to the one we've got." The chair follows 530 the conversation and says, "It sounds like the issue has been 531 addressed and there's consensus to stick with the current solution." 532 The objector is not satisfied, maybe even saying, "But this is silly. 533 You've seen that my algorithm works. We should go with that." The 534 chair makes the judgement that the consensus is rough, in that there 535 is still an objector, but the issue has been addressed and the risk 536 argument has won the day. The chair makes a working group last call. 538 Now the worst case scenario happens. The objector, still unhappy 539 that their preferred solution was not chosen, recruits one hundred 540 people, maybe a few who were silent participants in the working group 541 already, but mostly people who are at the same company as the 542 objector who never participated before. The objector gets them all 543 to post a message to the list saying, "I believe we should go with 544 the new elegant algorithm in section Z instead of the current one. 545 It is more elegant, and works better on our hardware." The chair 546 sees these dozens of messages coming in and posts a query to each of 547 them: "We discussed this on the list, and we seemed to have consensus 548 that, given the inherent risk of a new algorithm, and the widespread 549 deployment of this current one, it's better to stick with the current 550 one. Do you have further information that indicates something 551 different?" And in reply the chair gets utter silence. These 552 posters to the list (say some of whom were from the company sales and 553 marketing department) thought that they were simply voting and have 554 no answer to give. At that point, it is within bounds for the chair 555 to say, "We have objections, but the objections have been 556 sufficiently answered, and the objectors seem uninterested in 557 participating in the discussion. Albeit rough in the extreme, there 558 is rough consensus to go with the current solution." 559 There is no doubt that this is the degenerate case and a clear 560 indication of something pathological. But this is precisely what 561 rough consensus is designed to guard against: vote stuffing. In the 562 presence of an objection, the chair can use their technical judgement 563 to decide that the objection has been answered by the group and that 564 rough consensus overrides the objection. Now, the case described 565 here is probably the hardest call for the chair to make (how many of 566 us are willing to make the call that the vast majority of people in 567 the room are simply stonewalling, not trying to come to consensus?), 568 and if appealed it would be incredibly difficult for the appeals body 569 to sort out. Indeed, it is likely that if a working group got this 570 dysfunctional, it would put the whole concept of coming to rough 571 consensus at risk. But still, the correct outcome in this case is to 572 look at the very weak signal against the huge background noise in 573 order to find the rough consensus. 575 8. Conclusion 577 Although this document talks quite a bit about the things chairs and 578 working groups and other IETF participants might do to achieve rough 579 consensus, this document is not really about process and procedures. 580 It describes a way of thinking about how we make our decisions. 581 Sometimes, a show of hands can be useful; sometimes it can be quite 582 damaging and result in terrible decisions. Sometimes, using a device 583 like a "hum" can avoid those pitfalls; sometimes it is just a poorly 584 disguised vote. The point of this document is to get all of us to 585 think about how we are coming to decisions in the IETF, to make sure 586 we avoid the dangers of "majority rule" and actually get to rough 587 consensus decisions with the best technical outcomes. 589 9. Security Considerations 591 "He who defends with love will be secure." -- Lao Tzu 593 10. Informative References 595 [Clark] Clark, D., "A Cloudy Crystal Ball - Visions of the 596 Future", July 1992. 598 In 600 [IETF24] Davies, M., Ed., Clark, C., Ed., and D. Legare, Ed., 601 "Proceedings of the Twenty-Fourth Internet Engineering 602 Task Force", July 1992, 603 . 605 [Sheeran] Sheeran, M., "Beyond Majority Rule: Voteless Decisions in 606 the Religious Society of Friends", December 1983. 608 Appendix A. Acknowledgements 610 This document is the result of conversations with many IETF 611 participants, too many to name individually. I greatly appreciate 612 all of the discussions and guidance. I do want to extend special 613 thanks to Peter Saint-Andre, who sat me down and pushed me to start 614 writing, and to Melinda Shore for pointing me to Beyond Majority Rule 615 [Sheeran], which inspired some of the thinking in this document. 617 Author's Address 619 Pete Resnick 620 Qualcomm Technologies, Inc. 621 5775 Morehouse Drive 622 San Diego, CA 92121 623 US 625 Phone: +1 858 6511 4478 626 Email: presnick@qti.qualcomm.com