Minutes of the SIPB Meeting of 07/21/97 The meeting was called to order at 19:30:00 by kretch. In attendance were Voting members: kretch yonah gutnik emhavens amu svalente katyking danw rjbarbal | sit djib nygren yoav bert Associate members: ghudson rshah yandros mkgray nocturne kcr jemorris mhpower jhawk mcmullan srz jweiss Non-voting members: Prospectives: jfmurphy lgdean sepherke jdaniel jewell kiev | sjens "Martin Peck" Guests: pocky zacheiss Treasurer's Report: kcr: The treasurer was asleep in the zone earlier. Chairman's Report: kretch: The chairman was asleep in the zone earlier. yonah: We still have a broken stapler and a broken chair. We should deal. Office Report: yonah: We have this nice big monitor but it's on a machine with a not quite so good resolution. So it might be nice to switch it with an SGI head, since the SGIs have better resolution. kcr: It might be possible to come up with other monitors like that, but it's a little early to tell. I don't think that screen is wasted with that resolution; it makes it easy to read things without causing eye-strain, like other monitors sometimes do. rshah: It's good that that head is at a hot seat machine. [But we could put it on steve.] gutnik: It might be nice to put it on a machine with more desk space so you can sit farther away. danw: I think mkc's speed is enough of an incentive. jweiss: Moving office hardware around the office generally doesn't require a motion. ghudson: But it does require loose consensus if you know people disagree. kcr: Move to get rid of oliver. It is one of the four last Athenized RS/6000s; it is ludicrously heavy; it is ludicrously large. The display is fairly nice for its era, but it is possible to find much nicer displays in this office. [Second.] jhawk: I'd like to object to people getting rid of machines prior to the start of the term. While I have my usual objections, I would like such motions not to take effect until the term starts. kcr: Friendly amendment to myself to make it not take effect until the term started. jhawk: It might be good to give it some time to settle where hodge is before making a decision. yandros: I don't quite understand the requirement that we always wait until the beginning of term. kretch: The idea is that there are enough people not around that waiting to make big changes may be worthwhile. kcr: How many people here are voting members? [9 voting, 12 associate.] If we have a reasonably representative percentage of the voting members, then I think they reasonably represent the SIPB and can make this decision. danw: I don't think the objection about it only having been there a week is valid, since I don't see how the altitude would help. mhpower: Karl was implying that the presence of voting members is important to deciding whether to make decisions. I would like to object to giving voting members a special role in this respect. kcr: I don't think this is germane to the motion; we can talk about it later. I think the associate membership is less cyclic, so there's less reason to think it would be less well represented. Friendly amendment to remove the not taking effect until the end of the term. [The motion is to dispose of oliver at some reasonable point in the future, probably by auction.] jhawk: I move to amend Karl's motion to add "not prior to the start of the term." [Amendment fails 7-8-5.] srz: Is it MIT property? Then we can't auction it. jweiss: I think we can deactivate it; I volunteer to do so. jhawk: Is there any good reason for it to not remain without its head space? kcr: I believe that there is no particularly good reason not to, but neither is there a particularly good reason to do so, since pickled serves the same purpose better (faster, smaller). jhawk: On the contrary, I think we should keep machines where it's feasible to do so. The only resource it really takes up is space, and we can make room for it. I can't see any justification for this motion. yonah: What's the point of keeping it, though? jhawk: I think it's useful to have a variety of platforms. pickled might fail. srz: Can we call the question? nygren: [Mumble. Stutter.] kcr: Objection. What did Erik want to say? nygren: Had Sam been using oliver? kcr: No, pickled. jhawk: He used them both remotely. Also, pickled isn't configured so that you can run tcpdump on it due to its streams configuration. Of course, oliver had problems too, but maybe oliver has features that pickled does not due to its configuration. mhpower: Within the past year this guy from the AI lab used to wander over to SIPB to check some web server code on an RS/6000. So we might want to see if he'd like to take the machine. yandros: He's gone. nygren: Another question is, since we do tend to accumulate hardware that we want to keep around but not give head space, maybe we should get a rack. kcr: Yes, and I was thinking of making an appropriate motion later. jweiss: I would like to make a friendly amendment to kcr's motion to the effect of making the machine gotten rid of through whatever means the Office Czar in Charge of Special Operations decides. (That would be me.) kcr: Accepted. jhawk: I'd like to note my objection to further enshrining the position of office czar in charge of special operations in any but the most informal operands. kretch: Noted. yandros: I liked Erik's idea; I was thinking of suggesting it myself. I think we might want to decide on it before we decide whether to get rid of this machine. danw: We have another RS/6000. jhawk: So? Then we'll only have one, and if it breaks we're [out of luck]. danw: Good! [Motion is called.] [Motion passes 12-5-5.] ghudson: We have this rack equipment under the couch. jhawk: It's not a very good one. ghudson: We could set it up here if we wanted to. yandros: How big is it? ghudson: It's double-wide, 4-6' by about 1.5'. kcr: I don't know whether Erik was thinking of a single-wide rack, which I was planning to move to obtain. mhpower: We might have more room for a rack after the cokecomm change happens resulting in automatic delivery. jhawk: I'm not sure what the rationale is behind putting a rack in the office rather than putting a rack in the machine room. yandros: My understanding is that politically, putting more machines in the machine room is not a very good idea. jhawk: Moving, say, deathtongue, to a rack might not serve much purpose. If we're looking for higher containment, then maybe we should do that in the machine room. ghudson: Keeping non-server hardware in the machine room is not really acceptable. kiev: I just retired a 7' tall standard single-width aluminum rack which I'd be willing to donate. kcr: Move to accept the donation of kiev's rack. kretch: We don't need to vote to accept it. Will you research this and see if it's reasonable first? yandros: I say we should thank him for donating hardware every week, but not have a motion. [Motion dies.] danw: Karl mentioned that oliver was one of the last four athenized RS/6000s. He didn't mention that snork was *the* last athenized RT. So I move that we get rid of snork. [Seconded.] jhawk: As the last athenized RT, snork holds a special place in our hearts. While some of us may be cold-hearted, I think we should keep snork. yandros: I find myself in the strange position of agreeing with jhawk. danw: I think it hinders the goals of SIPB by holding a special place in our office. mkgray: Would Matt find there to be a serious decrease in the number of useable-to-him heads in the office? mhpower: I think the decrease in heads is an issue for me, but the decrease in the number of distinct processor architectures is a more important issue. jweiss: Has anyone determined if the machine is still usable after it lost its disk and most of its swap space? mhpower: I'm logged into it. danw: It's not just the last athenized RT on campus; it's the last RT on campus. jhawk: Does Dan have a conflict of interest here? danw: Are you suggesting that I get rid of excess platforms so that I don't have to build moira for them? jhawk: Certainly not the former. But ancient platforms do require resources from Athena operations and Dan may have opionions. kcr: He's not part of Athena Operations. kretch: Any further discussion on this? danw: People sometimes compile things for it; it takes up office space; it's not doing anyone any good. We should move on! We should put other things there. srz: I have found snorkelwacker to be one of the loudest machines I've ever seen. I've been hoping to see it gone for about ten years now. jhawk: Isn't it the oldest machine we have? yandros: I think this [the mac] may be the oldest machine we have. sit: Does snork serve any other purpose other than being the most useable head for Matt? jhawk: I log in on it occasionally. mhpower: One time recently I was looking for security problems that happened to be specific to BSD 4.3 and it was useful for that. jweiss: At one point there was a screaming tape drive with snork and I'd like to include it in the motion. [We think it was auctioned off.] [Motion fails 8-10-4.] ghudson: snork survives again! kcr: I think I can find another disk for it. kretch: Good for you. kcr: I move that we dispose of banana-sr. jweiss: I'd like to note that that monitor is technically one I borrowed from DCNS back when DCNS still existed. Before we got rid of it under any such motions, I would be forced to confirm that they did not want it back. I suspect that this would cause them to laugh in my face but I'm willing to put up with that. danw: I thought banana-sr was doing Appleshare file sharing. mkgray: I used it last week. I think that service should continue to exist. nocturne: Friendly amendment to wait until the functionality is moved elsewhere. kcr: Not accepted. It's not clear to me that the functionality is particularly useful. The functionality has been largely replaced by net-dist at this point. yandros: Maybe four people have used the file service in the last six months, and we can do it on the old powermac much better. jhawk: I ask those presenting the motion to justify what they gain by removing the machine. kcr: That area could be used for other stuff. Say, someone's lunch. mhpower: It isn't clear to me that we are keeping gene-simmons, which affects this motion because of file service. kcr: I don't think we will move to get rid of it if we think it's useful as a file server. srz: If we're worried about the service, we have at least two other macs that can do that service, and we're not moving that the service goes away. mhpower: If you're doing something like running Kerberos 5 clients, you don't want network load from file-sharing clients. kcr: But we have a second Mac that probably won't be an office head. jhawk: It is the last 68k mac we have and thus the last of its kind. kcr: Isn't there another 68k motherboard in the office? [It's an '030 NeXT.] jhawk: That's not comparable. srz: I move the question. [Moving the question passes many-1-2.] [Some points of order regarding the wording of the motion; motion is clarified to include the CPU and the two disk cases and their contents but not the keyboard or mouse.] [Motion passes many-3-1.] djib: If the services need to move to another computer, do we need to deal with naming issues? jweiss: We can solve the problem trivially by renaming gene-simmons to banana-sr. kcr: Move to dispose of deathtongue. (There will be a motion to make bart-savagewood a NetBSD/sparc machine if this motion passes.) [Second.] jhawk: deathtongue is our last non-sun4m Sun machine. So it has a unique kernel architecture. I don't think it takes up a significant amount of space and I can't see any justification for removing it. And it's a comparatively new machine. kretch: We don't need to do everything in one day. jhawk: Allowing mass psychology to drive your motions forward does not constitute a justificatin. kiev: Not having an IPC around to use as a test NetBSD machine could be a pain. mhpower: I think it's good to have a machine (bart-savagewood) running a supported Athena release in the back of the office, for people who don't want to log in in the front, and converting it to a NetBSD machine would be a disservice. kcr: [re Kiev] There's no particular difference between a NetBSD/sun4m machine and a NetBSD/sun4c machine except for the kernel. [re Matt] I think we have a sufficient number of suns that it wouldn't be a problem, and I wasn't aware that we had a need for supported Athena machines in the back of the office. jhawk: The only Sun we have that's running the non-beta release is btc. So if we reduce the number of Suns, we'd probably have to reduce the number of beta machines, which I think is a bad thing given the state of bug-finding at MIT. kcr: I disagree that there is insufficient bug-finding or that having one fewer sun would particularly reduce the number of beta machines in the office. This is not strictly germane to the motion. mkgray: Move the question. [Moving the question passes many-2-0.] yandros: We used to ask if anyone had made up their mind instead of moving the question. [Motion fails 2-15-3.] kcr: Sal, what is the status of thornhump. svalente: I use it. ghudson: We have sipb-athena built for it except for MH. svalente: And most lockers don't support it. A lot of people use it for an X terminal. nygren: Given that the machines in the 6.003 and 6.111 labs are HPs, I would rather see more stuff built for it. svalente: The HPs there don't have AFS or Kerberos or anything useful. [Some disagreement.] yonah: The fifth floor machines don't have it; the sixth floor machines do. ghudson: It would be cool to get EECS to run useful stuff, but it's probably a fair amount of work and no one's doing it. mhpower: I think HP has the most worldwide market share of any Unix platform not supported by IS, so this is a valuble machine for general Unix development. yonah: If MH was built for it, someone should look to push the changes back on EECS. kcr: How difficult would it be to update the machine to HPUX 10? Can we get a distribution? jhawk: I'm sure we can. jweiss: I move that if nobody at IS wants the mac monitor back, that we get rid of it like the other stuff we're getting rid of. [Second.] [Motion many-0-2.] sit: I think it's not useful for people to just state that they disagree with things. jhawk: I'd like to suggest that we avoid getting rid of more than one or two machines per meeting, since it's difficult to evaluate space consumption. [Analogy to thermostat.] srz: Given that this process seems to be so painful, multiple things happen at one time to get it over with. People would not look forward to this type of debate every week. kcr: Unless anyone strenuously objects, I'd like to install NetBSD on binkley and make hodge the X display machine. jhawk: I've used binkley as a Vax BSD 4.3 machine for nroff purposes. I doubt I'd be able to do that under NetBSD using binary compatibility. So I would strenuously object. kcr: Would it be reasonable if it could still boot BSD 4.3? jhawk: That might even potentially be reasonable. svalente: It's still nice to have a 4.3BSD vax still in the office. They were really popular once; there are still some out there. And why would we want to put NetBSD on that? kretch: Move this discussion off line. svalente: I move to allocate up to $40 to buy up to $40 worth of 3.5" high-density floppy disks. [Second, white-ballot.] jhawk: I move to allocate up to $50 to purchase all the remaining Bloom-County books. [Second.] ghudson: I support this motion because of label-making. jhawk: Could anyone suggest whether $50 is reasonable for the remaining books? mhpower: Move to amend the motion to have all expenses come out of the auction budget rather than the IS SIPB budget. jhawk: Not accepted as a friendly amendment. [Amendment passes many-2-1.] jhawk: What is our authority to make motions regarding auction money? jweiss: I've collected the auction money on the assumption that the SIPB would eventually tell me what it wants me to do with it. So making a motion is fine. I may not execute the motion, but I'll pay back whoever does so. jemorris: How much is there? jweiss: I think several hundred dollars. [Motion passes many-0-2.] jweiss: Whoever does this should get the receipt and bring it to me. We'll pay for tax. MIT Computing Report: ghudson: 8.1 goes public today. jhawk: I think I'm going to update bill [to 8.0L] tonight. danw: It wouldn't be a complete shock if there were new dialups by the end of the week. Computer Services: mkgray: The web server changeover is likely to take place this week. The first of the interesting new services is a bidirectional web zephyr gateway; it will be available soon. yonah: If you know of things which should be taken care of before the migration, let me know. mhpower: Has the web zephyr gateway been mention to the people maintaining the zephyr servers? mkgray: No. kretch: it has now. kcr: It's possible that it should include a reasonable limitation as to the number of people using it at a given time; say 100. mhpower: We'll be doing occasional updates to the news servers due to problems (including some security problems) in INN. The news software on beacon is now upgraded to INN 1.5.1. amu: Weren't there performance issues? mhpower: Yes. They were worked around. SIPB Backup Report: jweiss: When's the last time an AFS backup completed? kretch: Last Monday. jhawk: Did he fix the problem or just get lucky? kretch: The latter. kretch: We think UFS backups all happened. Publications Report: kretch: amu is supposed to deal with ILinux? amu: Yeah. kcr: Maybe we should distribute 8.1 release notes? jhawk: Have paper 8.1 release notes been handled? danw: I believe they're being frantically finished. ghudson: The content is finalized, I think, but I don't know about the publication. jweiss: We should recycle all of the 8.0 release notes. jhawk: We should send mail to sendpubs. kretch: I will. R/O Report: kretch: amu has begun to R/O suck and is doing a nice job of it. IAP Report: '() CokeComm Report: '() Other: mkgray: The Alta Vista robot chose today to index a large chunk of our web server, including most of /afs/athena.mit.edu/astaff/project, two user home directories, and an old backup of one of Greg Hudson's todo lists. jhawk: Alta Vista has been returning a lower than expected number of hits in the athena AFS cell; maybe this is being fixed. jemorris: xiphmont says he fixed some bugs in the Linux audio player in outland. sit: There should be a linux-dev meeting some time soon. yandros: The AFS clue dump will happen this Wednesday at seven. Other Other: jweiss: It's inadvisable to put three inexperienced canoists in the least stable canoe. [sesamest: We weren't all *completely* inexperienced.] rjbarbal: I've never fallen out of a canoe before. jewell: I've never yet played Ultimate. This should be fixed. danw: I can telnet from my living room to my bedroom now. kcr: Some of us have been able to do this for the last five years. jweiss: There will probably be post-meeting ultimate. The meeting was adjourned at 20:32. Minutes taken and submitted by ghudson.