Minutes of the SIPB Meeting of 2007-08-13 The meeting was called to order at 19:30 by tabbott. In attendance were Voting members: andersk arolfe aseering ctl danjared jbarnold kenta kchen nelhage quentin tabbott tilia | geofft Associate members: asedeno awozniak belmonte | int18 Prospectives: mclamb Guests: Officer Reports: tilia: Spend money! tabbott: I’d like to defer discussion of the incident in the machine room until the end of the meeting. tabbott: There was an incident in the machine room during the Linerva outage Tuesday night. Ken refused to leave the machine room which eventually resulted in the CPs removing him. There was detailed email. kenta: You are heavily involved in this issue. At your option, you may wish to have someone else chair this part. tabbott: Are you happy with Nelson chairing? kenta: Nelson was also heavily involved. kenta: I would like to nominate jbarnold. jbarnold: I would rather not. tabbott: tilia, would you be willing to preside? tilia: Sure. tilia: [Threatens tabbott with the gavel.] kenta: So, I would like to speak presenting my version of the story. [reads prepared statement, quoted below] I am proud to have helped the Linerva installation in the following ways 1. There was a message sent to all logged-in users T−10 minutes before the outage using the “wall” command. This message was sent at my urging, after all the maintainers planned to, but ultimately forgot to send one earlier. 2. I helped remove its Ethernet cable before the machine was pulled out of its rack to install the new RAM. 3. I began to produce a series of notes about things that were done and perhaps should have been done during the install, notes that I hope would be useful in a future Linerva upgrade or provide details for the curious about exactly what makes up Linerva. The notes that I took before I left the machine room were sent that evening to linerva@mit.edu I will now present my story of the events that occurred in the machine room. Strangely enough, I would like to commend Mr. Timothy Abbott for his description of the incident. It is very objective and a fairly accurate description of the events that occurred. There are just a few omitted details, which I will use as the starting point. One thing omitted from Mr. Abbott’s description was just how high emotions seemed to be running during this incident. For example, after the demand was made to me that I leave, I said that I understood that I was perhaps asking too many questions and annoying them while they worked offered that I would like to stay and promised would stay out of the way as much as possible. This was rejected. I said that that in my experience, disk partitioning is one of the trickiest parts of an installation, and offered to leave after this step was completed. This was rejected. After these rejections, I began to believe that I was no longer dealing with reasonable people and that saying anything further would no longer be useful. I personally got the feeling this was no longer a struggle about how best to upgrade Linerva, or ultimately how best to Further The Goals Of SIPB, but purely a political power struggle of whether one man had the power to unilaterally eject another man from what he believed was his marked territory. I deliberately chose to remain silent in hopes that to the three of them that they would listen to themselves and would realize the silliness of the ultimata and threats they were uttering and it would shake them to their senses. At this point jbarnold came back into the machine room and attempted to moderate the conflict. I would particularly like to commend Mr. Arnold for this action, because in a situation where emotions were running high, the parties were failing to communicate usefully, moderation was exactly the right thing to do. We began to have what I believed to be the beginning of a fruitful discussion of what it would take for me to leave the machine room amicably. Mr. Arnold, once again, thank you. Sadly, it was only moments into the discussion that two officers from the MIT Police arrived. This highlights exactly the silliness of calling the police; once they arrive, it is no longer possible for us to work this out among ourselves and have a reasonable discussion. Certainly the police should not be forced to make a judgement of which members or not have the authority to eject which other members or not from the machine room. Although it was Mr. Geoffrey Thomas who took the initiative and made the call to the police, I do fault Mr. Abbott and Mr. Elhage as well as it was well withing their power to ask Mr. Thomas to call off the police, but they chose not to. As I’ve said, certainly the police ought not be forced to decide which SIPB member has the right to eject which other SIPB member, but unfortunately, they arrived and were forced to make such a decision. They arbitrarily made a decision that I should leave, at which point I peacefully exited. I chatted with the officers for a while afterwards. One of them remained utterly flabbergasted why they had been called. I was a member in good standing with the organization, I had been let in by some other member of the organization. To his eyes, nothing bad was happening. I asked whether I was to be charged with trespassing for being in a place I was not supposed to be. They responded that no, I was not to be charged with trespassing. I strongly suspect there exists personal animosity between Mr. Abbott and myself, and this contributed very strongly to the escalation of events that evening. Whatever feelings I have toward him, I was willing to put aside for the duration of the Linerva installation. I cite as evidence Mr. Abbott’s animosity toward me with his tendency to marginalize me and my work at SIPB. Today, in a e-mail, he wrote as that I am not an active SIPB member as evidence of why I should be regarded with suspicion. Nevertheless, I have attended two hack-a-thons this summer, of which I believe he was present at both. The one earlier this summer I succeeded in installing the Xmonad windowmanager into the ghc locker. However, the kicker is that Mr. Abbott is particularly privy to my work in SIPB, as he sought input in May for the SIPB Annual Report about what members have done for SIPB, and I sent the following message to him. Locker software installed or upgraded * Glasgow Haskell Compiler * haskell-mode Emacs mode * KDE * GNU Compiler Collection * XEmacs * six, a hex-playing game Two locker software I used to maintain have “graduated” to being maintained by ISNT: R and Pari/GP. The continued un-compatibility of linerva (lack of i386_rhel4 sysname and missing libraries) causes locker software maintenance headaches. Linerva maintainers are sometimes helpful, sometimes not (libtiff.so.3). Zephyr log search (glimpse locker) index regenerated at the beginning of term. New version of KNOPPIX image placed in AFS, new CDs burned (and CD-RWs rewritten) in the SIPB office. gcc and and kde are large particularly hairy pieces of software to make available in a locker, taking dozens or hours of work each to compile and install. However, my work was marginalized in the End-of-Term report to simply “various locker software packages were upgraded.” This, combined with with Mr. Abbott’s assertion today that I was not active in SIPB, is evidence that he has a personal agenda to marginalize me, which I believe played an important part in this incident. Since this report, I have completed one major project that I have worked on for several years now, which is the complete installation of Athena Linux into AFS along with all extra packages included with Red Hat Enterprise Linux 4. This brings us full circle to the Linerva installation: one of the reasons for this installation of all of RedHat into AFS was to provide a workaround for incompatibilities with the Debian libraries installed on Linerva. I would to thank you all for the opportunity to opportunity to address the Board and now yield to the chair. ctl: Assuming that nobody else has any questions, Ken, not speaking to other people’s behavior, what makes your behavior during this specific incident, and not before or after, acceptable? Or was it acceptable? kenta: Things escalated a lot. ctl: I don’t feel that answers my question. kenta: That makes it difficult to answer, but I will try. I feel that my actions were acceptable, because I feel that SIPB should be inclusive in projects. If it’s uncertain as to whether we should be inclusive or not inclusive, it should fail open as opposed to fail closed. By default, open. ctl: Are we scoping this discussion to Ken’s actions, or including the other participants’ actions as well? ctl: I’ll state for the record that I think Ken’s actions were unacceptable. I also have questions about the actions of the others involved. belmonte: SIPB is a collection of people who more than most can be quite rigid and rule-bound, which is great when everyone agrees what the rules should be, but not so hot when there are disagreements. Ken raises an interesting point that projects should be inclusive. There was a counterpoint raised in the email discussion that perhaps this wasn’t the time to make that decision in the heat of the outage. Leaving that aside, i think it may be worth noting that this doesn’t seem to be the first instance when a certain stubbornness caused things to, as you say, escalate. I refer in particular to the events of March 2005 with the Zephyr bot to -c geek. Again there was disagreement as to whether this was useful or annoying, and that disagreement was protracted and contentious because that zephyrbot was not taken down even temporarily. My question to Ken and to everyone is, (1) what should you do differently in instances like this, and (2) what should SIPB do differently? kenta: There are lessons to be learned. Certainly the fact that Mr. Arnold came in to moderate the argument was a very good thing. I believe that would have ended well had the police not arrived. The other thing is that MIT as a whole is a learning experience for all. What hurt this situation is that there’s this technical aspect of “we’re upgrading a machine” and there is this personal stuff. It is a skill that is useful for the rest of your life if you can learn to separate the two. I believe a failure of that happened. jhawk: I think a bunch of things went wrong, but those problems are self-correcting and not worth the board’s further time. quentin: I do agree that I don’t think there’s anything that really is wrong with our policy. Some people made some poor decisions that made this incident worse than it had to be. It would be sufficient to reiterate existing policies and say this was an isolated incident. tabbott: I don’t think you’ve answered this question of whether something should have been done differently. kenta: I feel like I answered it. The things that should have been done differently were that the police ought not to have been called, and people should learn to separate emotional things from technical things. arolfe: I would agree that the police should not have been called, however, I still think the right thing you should have done is not to keep arguing until you get your way. SIPB have different mechanisms for making decisions, and they would not have helped you right then and there to be part of the Linerva upgrade, but staying and arguing is not an okay way to do that. belmonte: I note the third-person phrasing of your response. My question had two parts, what should SIPB in general have done differently, and what would *you* have done differently? kenta: I’ve said before that I’ve reached the conclusion that I felt I was no longer dealing with reasonable people. The solution to that, to bring in a third-party moderator, didn’t occur to me. I wish I had thought of that a lot earlier. I wish I had thought of that before the police were even called. We might have been able to work this out. int18: Is there anyone here that would have done things differently in a similar situation? jhawk: I don’t think there’s much purpose to questioning in front of the board. Before any maintainers let kenta into the machine room again, they’re going to think carefully about that. I don’t think we’ve crossed the threshold that merits public questioning, and I think the problem is self-correcting. aseering: It would be nice if geofft were here. I would like to know if anyone present would have done anything differently. jhawk: I buried my head in the keyboard when CPs walked by. In retrospect, I could have gone and chatted with them and delayed for five minutes. ctl: Do you really think that the best thing you could have done in this situation was to work out your problem with Tim, Jeff, and Nelson right there? Can you not think of anything better that you could have done? kenta: I believe it’s within the realm of possibility that it could have been worked out in fairly short order. jbarnold: So I’m wondering what resolutions to the situation you might have found acceptable had they been proposed by a third-party moderator who you found to be reasonable, that would have been amenable to both sides. kenta: I was having difficulty understanding why I was being ejected. A third-party moderator would have helped work that out rationally as opposed to anger talking. That may have been enough to have me leave. jbarnold: One of the things I found a bit curious in the few minutes I talked with you, when I indicated that the police were on their way to the machine room, I suggested that you could leave the machine room and handle differences later, your response was to the effect that “I’m willing to leave the machine room, but what can you do for me if I leave?” I both didn’t know how to respond to that, how to “sweeten the pot” for you, and it was seconds after that that the CPs arrived. I’m wondering what you did mean by that and what were you looking for for leaving the machine room? kenta: In response to the second half of your question, I was looking for a reasonable discussion. Based on what you quoted to me, it seems that it came out wrong. I apologize that it came out wrong. I don’t remember what I said. jbarnold: I don’t remember exactly either, I don’t mean to quote you. kenta: It’s sort of a game of chicken, because the CPs can be called off at the last minute. Ultimately even after the CPs came, Tim and Nelson can say that we’ve decided to work this out. This is why I decided not to leave under the threat that the CPs were being called. I guess the point made is that the degree to which the non-inclusiveness was going on is best illustrated if the CPs actually came and escorted me out. This they ended up doing and thus the point was made. tabbott: Are you suggesting that your motivation for not leaving is to make a political point? That seems to be what you said and maybe I’m misinterpreting you. kenta: That the CPs were called is out of my power. tabbott: That is not entirely true. geofft told you he was going to call the CPs. It was somewhat within your power. kenta: Sure, but it’s questionable to claim that. That the CPs were called and they came is sort of that which happened and I am basically not going to do anything too special because of it. arolfe: Tim, do you think that campus police are an acceptable SIPB conflict resolution? tabbott: No, I don’t think they resolve a SIPB conflict at all. However, they are a mechanism for handling a situation that has gotten out of control in general. When people are refusing to leave an alarmed space and acting kind of strangely, hiding behind servers that we don’t own, having the CPs handle the situation is not a terrible way for things to end. ctl: How much time was there between the beginning of this dispute and the CPs arriving? tabbott: I think between 30 and 45 minutes. Do you agree? kenta: There was a point at which Tim asked everybody to leave, and a point at which I left. That time was about 30 or 45 minutes. The point that the CPs were called was maybe 5 minutes before I left. ctl: So you argued for half an hour. kenta: It didn’t seem that long. andersk: Do you feel that using your position in the machine room that you don’t have access to as a sort of bargaining chip to make your point was more effective than bringing it up at this meeting, or going to the Linerva maintainers later? kenta: I feel that SIPB projects should have low barriers to entry. It should not be necessary to bring it up at a SIPB meeting in order to be allowed to watch and take notes for the Linerva installation. If our positions had been reversed, I would have said, sure, watch along. If you haxxor the root password, I will try to get you kicked out of the university and pursue criminal charges against you. awozniak: Was the fact that you feel strongly about SIPB projects having a low barrier to entry, and that although you were not part of the Linerva team you wanted to be involved, was this part of the agurment? kenta: Yes. tabbott: That’s not what you said, you said that Anders and I were incompetent to do the upgrade and we “needed” your help. kenta: I don’t think that is an accurate description. I mostly tried to stay away as much as possible. It’s only when you started commanding people to leave that things began to escalate. ctl: I don’t think we’re getting anywhere today, but I don’t think this was resolved. I propose that we table this. jhawk: Nothing is resolved because there’s no goal. What’s to table? ctl: I believe that we do need some assurance that kenta’s behavior will improve. kenta has not admitted any wrongdoing yet. jhawk: Nobody’s going to invite Ken into the machine room. Do you think something’s more is necessary? ctl: Yes. belmonte: Chris makes a good point, but I don’t really see how it’s going to be resolved right here. kenta: Move to end discussion. jbarnold: I would argue that we are making progress. <3-5-9, motion fails> ctl: I agree with jbarnold that there’s still stuff to do here, but letting people sleep on it would help, and it would also help to have geofft here. jbarnold: geofft is going to be out of town for three weeks. quentin: And we probably don’t want to have it during orientation. ctl: I don’t think this is a time-critical thing, I just feel it’s unfinished. aseering: I would be sort of interested to hear a little bit from geofft. jbarnold: Here’s my take on the situation. Because of some history between Tim and Ken, when Ken was asked to leave the machine room along with the other people that were not Linerva maintainers, Ken concluded as he said that he did not think that Tim was being reasonable, and the way that he chose to respond was to utilize his presence in the machine room as a bargaining chip in order to convince Tim that he has no choice but to “listen to reason” and allow Ken to participate in this installation process. I would say that it’s at the point when Ken started making that decision that things first went from a realm of people behaving in a fairly civilized way to people behaving strangely and doing strange things. I do think that was the first and perhaps more significant of the missteps that occured. You should instead have brought this up with the Linerva maintainers later, or the board. People are sympathetic to the idea that SIPB projects should be more inclusive, but that it’s crazy to resolve it in that way. tabbott: I would like to remark that people seemed to think they had questions of geofft. You’re free to send emailed questions if we table this for one week. That would be plenty of time for an email discussion. awozniak: We do need geofft’s input as to why he called the CPs. jhawk: Move to limit debate on Chris’s motion to 3 minutes, 60 seconds maximum per person. tabbott: That’s going to waste a lot of time arguing over that. jhawk: Motions to limit discussion are not debatable for that reason. <3-9-4, motion fails> quentin: I was not part of this really at all, but I was at the office at the time. I can say my recollection of what geofft did and what he said he did. arolfe: I have some serious questions about whether what geofft did was reasonable, and I would rather hear from him. We should table this discussion for now and have it when geofft is here. kenta: You may want to know why geofft acted the way he did, in which case you need him. I will state that Tim’s description of what he did is accurate. kenta: I would like to request a roll call ballot. The idea is that names are attached to the vote. tabbott: Why is this important on the motion to table? kenta: I would like it to be on the record who wants to end the discussion. tabbott: To table? kenta: To table the discussion. ctl: Nobody is discussing ending the discussion. kenta: I withdraw my request for a roll call ballot. jbarnold: geofft requested that I call him if his presence was important. I have done so. [geofft enters] ctl: I withdraw my motion to table discussion. jhawk: It would be nice if people who feel we need to keep this discussion would bound it in some way. It is unclear to me whether this is for fact finding or chastising or what. jbarnold: I would like to hear from Ken whether he agrees in retrospect that choosing to use his presence in the machine room as a bargaining chip or a means of enforcing reason on Tim and the other maintainers, was a step backwards or affected the discourse in a negative way? jbarnold: This was before the CPs were called. kenta: I very strongly disagree with your characterization. I stayed in the machine room because I felt I had the right to be there. jbarnold: Is this a general right that you believe all SIPB members have a right to be in the machine room at any time, or is it more specific to this situation that gave you this right? kenta: The latter. jbarnold: Was it that you think that anyone who has volunteered to participate in a project has the right to be present for any work done on that project, in the machine room or not? That you were added to the debathena-listeners mailing list a few minutes prior? kenta: No to the first and no to the second. jbarnold: I’m guess I’m looking for your explanation as to what gave you the right, and under what circumstances someone would not have the right to be in there. kenta: Modulo security concerns such as setting the root password, I believe that any SIPB member that is interested in working on a project and behaving reasonably has that right. ctl: I think it might serve us well to take the focus off of Ken for a moment, since geofft was so kind to join us. I’d like to ask geofft the same question I asked Ken to start this discussion, which is why do you believe your actions during the course of this fiasco, and not before or after, were acceptable? If you don’t feel they were acceptable, why not? geofft: Why I thought they were acceptable at the time, or now? ctl: I’m not really interested in why you thought they were acceptable at the time, but what you think now. geofft: That’s a hard question. ctl: You can think. If you’d like we can come back to this later. geofft: There were certainly facets of what I did at the time that were not acceptable or less than acceptable, such as not talking to Tim, which is something I realize in retrospect I should have done. That is something I really wish I had done, and I apologize for that. There are other facets of the situation that I think are not so unacceptable that I would have been dissuaded at the time from doing. There were facets that were probably acceptable, such as calling the CPs. I’m still unclear as to whether that was acceptable, because that depends a lot on whether it was necessary, and I’m unsure whether it was necessary. ctl: That’s a reasonable answer. To try to understand that answer, if you had gone to Tim and discussed the idea of calling the CPs with him, let’s assume for the sake of argument that Tim also felt you should call the CPs. Do you think that was the most reasonable course of action at that time, and why? geofft: If Tim gave you consent call the CPs, that would have have made it significantly more acceptable. I think that that only would happen if we determined that it was necessary. geofft: I don’t think the other parties in the room would have given consent if it was not the right thing to do. arolfe: If you could go back and change what you did, would you? geofft: I probably would have done something differently. ctl: I don’t think I have any more questions for anyone. Move to table this again. nelhage: Until what time? ctl: At the discretion of someone other than me. awozniak: If this discussion continues, what course of action would it lead to? Is the goal to create a statute in SIPB bylaws that tries to pre-mediate interactions like this, or what? ctl: I have no goal to create any sort of SIPB anything. My only goal is to have some assurance that Ken will moderate his behavior in general, not specifically associated with the machine room. jhawk: On this topic, geofft, I think when there’s a possibility of taking drastic action, and there are more experienced people around, you should let them deal, when it comes to revocation of membership, or calling the CPs. jhawk: I’m not sure what we are tabling. The motion has no power because we could do it without the motion regardless. ctl: Okay, I withdraw the motion. [more motions, amendments, and withdrawals] ctl: I disagree that we haven’t accomplished anything at this meeting. MIT Computing Report: kenta: It’s been a while since I’ve come to a meeting. The longjobs service was discontinued at MIT about a year ago. I was sad to see it go. Kind of ironically, I think SIPB got some or all of the servers that longjobs was using. quentin: Linerva got upgraded. SIPB Projects Report: tabbott: We had a Linerva outage on Tuesday night for around seven hours to upgrade it to Etch. We’re hoping to move the Debathena repository over soon, but we have various bugs we’d like to fix first. mclamb: I have some questions about tours. Reading through zephyr logs from years past, I was wondering about the list of stops. [discussion of stops] mclamb: There’s talk of this mythical holodeck in CSAIL. Does that really exist? [discussion of scheduling, food, saferide vans, publicity] jbarnold: My impression in general is that postering is completely worthless and a waste of time. Consider running an ad in the Tech. Word of mouth works the best. kenta: I would be willing to create a Facebook event group, unless otherwise directed. mclamb: Can I get bits for /afs/sipb/admin/text/tours? jhawk: Yes. tabbott: SIPB computer stories are going to be on Wednesday, September 5 at 8 PM. We should consider whether we want food to show up at the beginning of the event, or at 9 or 9:30. Office Report: danjared: The toolbox was left in the machine room. jhawk: We might want to think about enabling long-distance access to our phones since it no longer costs money for calls. tabbott: This would be like $120 for four phones? aseering: Seems reasonable. tabbott: This is simple a one-time fee for the convenience of not having to dial a 7-digit number for every long distance phone call. quentin: Move to allocate up to $150 to enable long-distance on our phones. tilia: Do people actually use these phones to call long-distance? quentin: I haven’t only because I don’t know what the access code is. tabbott: I did once. jhawk: I’ve done it several times, but I don’t have a problem dialing the access code. arolfe: I don’t think it’s worth it. kenta: How about if we were to purchace a phone card and dial with it? jhawk: That would be even stupider. jbarnold: This 7-digit string is not hard to look up on Athena. I think that instead we should have a stack of $1 bills, where every time you go to the effort of dialing this code, you take a $1 bill off the stack. It would cost us less money. <3-5-7, motion fails> nelhage: Someone who knows how to look up the number should put a piece of paper on the wall. jbarnold: You stella scripts and it’s the account number. aseering: quiche seems to work now. quiche also has a new graphics card. This is why quiche works. kenta: What graphics card did quiche get? aseering: Radeon X1300. Proprietary drivers. kchen: Athena might be trying to add the fglrx to the install. quentin: Athena currently just uses the vesa driver for everything. kenta: I have been unable to locate the universal power adapter and the CompactFlash cards for the camera. jhawk: The CompactFlash adapter for the oscilloscope showed up, and is installed. Education Report: tabbott: Activities midway is coming up. It would be good if we had a flyer. quentin: August 31, 4–6 PM. tabbott: Also if people think we should have any cool demos at the midway, let me know. Are people available to be at the table on Friday, August 31? It’s poor for only one person to be there. Volunteering for this entitles you to free cokes from CokeComm as you hand them out to freshmen. tabbott: I’ll send email. kenta: Did we ask for electricity and did we get it? tabbott: Yes, yes. tabbott: We’ve discussed having an event in the SIPB office over orientation, similar to the Inode and fsck Block Party. Is anyone interested in helping organize that? tabbott: Sounds like people don’t think we need to do it. quentin: We can try to hold a hackathon then. tabbott: A hackathon is a weird thing for people to show up for. jbarnold: I think something should happen. I don’t want to organize it. tabbott: It could be a hackathon where we give a presentation about what project SIPB is working on. tabbott: Nelson and I will try to figure out a plan for that. tabbott: geofft has volunteered to organize an Ask SIPB column. It would be great if other people could help him out. In particular, looking for inaccuracies in previous columns or things that are no longer true would be a good way to help out. Other: '() Other Other: '() Discussion Report: The meeting was adjourned at 21:00. Minutes taken and submitted by andersk.