Interview with Patrick Brady

0:00 All this stuff. Okay, so now it's actually working, and it's the 23rd of March at a quarter to two in the afternoon, and I'm talking with Patrick Wright. So, it's great, now you're immortalized. Yeah. Uh-oh. Yeah, well, of course, unfortunately no one but me is ever going to listen to this. I already knew who you were. Okay, so let me see. I think a good place to start. I know you've worked in a number of different areas, so even though I know most of the story, you probably start off with saying how it comes about that you're... that I'm here? Yeah. Okay. So how it comes about that I'm here. So, right, so you know I did my PhD with Werner Israel in Canada and went to Newcastle in England for two years. From there, I did the standard thing of applying for every single postdoc I could think of and was fortunate enough to get an offer from KIPP, which was something not to be missed. that's how come i'm here yeah that's why i'm here so so but um but you mean interest in gravitational waves or yeah i mean presumably that entailed sort of a shift in right so yeah so for me um it wasn't even clear when i was coming over because it seemed to me that you guys had already done a lot of initial source modeling and um it seemed that kip was changing and moving towards data analysis. And I must admit, when I was coming here, that actually, for me, was a big question, whether or not it was something I wanted to get into. And, of course, I knew nothing about it, so it was hard to make an informed decision. But what's been very obvious, though, I think, to the younger people in gravitational physics, at least it's my impression, is that moving in the direction of LIGO or LIGO-related data now, LIGO-related work is clearly a good thing. It's where there's going to be money in the future, and so it was a sensible move. Plus, having come here, I mean, it's another problem that needs to be solved, and it's pretty neat working on something like that.

2:30 So, I mean, that's been, for me, probably has just been, I just viewed it to some extent as a nice mathematical problem to some extent initially to be worked on. And so that's what I got interested in, certain aspects of that. Of course, once you're here, you can't help but get involved in source modeling and issues of sources like Lee and Ben now discovering these things, these new unstable modes. So yeah, so I guess, I guess really what brought me to it is that it's an open field which have lots of problems. Right. So, well, I guess then the obvious next question is what are you working on at the moment? So, okay, what am I working on at the moment? So there are two data analysis lines being pursued that I'm involved in. The first one is periodic sources, so gravitational radiation from things rapidly rotating neutron stars. And fortunately, we just got a boost have a potential source in this class. So this is pretty good news. And I've been doing that with Teviot, Teviot Creighton. And the status of that is that we've been working on issues related to essentially figuring out what the template bank is. It's essentially equivalent to the coalescent binary problem, except that we have Doppler corrections as are the issue. And also, for example, in sources like the R-mode sources, you've got very rapid spin down, and so the frequency drifts quite rapidly just due to the gravitational radiation reactions. You have to incorporate that somehow. So we've been figuring out exactly how big the parameter space is and exactly how hard that problem is going to be. So that one's still moving ahead, and we're working on hierarchical search issues for it. The interesting, and I'm not sure how much officially I'm allowed to say, but no, we're working on this 40 meter data analysis problem, which I don't know if anyone else has mentioned it to you. So in 94, the interferometer here took about 40 hours of data, out of which about 28, actually more like 30, 35, or so 30 to 35 hours of it, is actually, the interferometer

5:00 is in Locke. And so we've, with Bruce Allen, pretty much everybody in the group here, except Kip, and is there anybody else from Milwaukee? So on the theoretical side, there isn't anybody else from Milwaukee, but we've been basically putting together a whole analysis to look for wave gravitational radiation from coalescing binaries in our galaxy. And so we have the place, we've actually done a first cut through the data and got some results, which are pretty scary in terms of the signal-to-noise ratios we're seeing, from what we think are not coalescing binaries in our galaxy. And we're moving ahead on that. So the goal is really just to have a first cut through the data, see how the data analysis works, and go from there. There we go, and it's off to go. Okay, so where are we? We were on 40-meter data analysis and stuff like that. So, yeah, so the goal... Oh, yeah, that's where we were. The goal is to try and put a bound on the number of events that occurred in our galaxy during these 40 hours. Right. We're hoping it's going to be somewhere around zero, but so far that's not looking too promising, but we'll see how it goes. Because in actual fact, the realistic bound looks like we'll be setting about around one every five hours or something, which is clearly way off. But given that we've only got 40 hours of data to work with, actually 25 good hours of data to work with, that's as good as we're going to be able to do. So how instructed has it been looking at data? Pretty interesting. I wish I had curves. I don't even have any plots that I can show you at the moment. Maybe before you leave, I might have to take them up myself and show you. keep things under control. But if you compare the results to, say, the normal case where you assume you've got stationary Gaussian noise and all these assumptions, you find that at about the 80% level, the false alarm curve deviates from stationary Gaussian noise, which is really bad. I mean, we're talking about if you set a threshold based on based on stationary Gaussian noise which would say around 5

7:30 you're just you're getting something like how does this you're getting something like an order of magnitude more events than you should be in the period of time that you're looking so that's pretty bad but even when you get up to higher signal the noise gets even worse I mean it's like 10 orders of magnitude more events than you expect. So it's pretty nasty. But, of course, this is only just a single interferometer. So the goal, as you know, you've always heard, is two interferometers in coincidence. And so we don't know how well that'll do. We'll see if we can go there and find out. We're planning on splitting the data in half and just sticking it together, like side by side, as if it's from two interferometers, and just see what that does. so many precisely that's about right well it's interesting for you here because one of the things that always strikes you with LIGO related stuff is that theorists are always talking about Gaussian noise so it's interesting yeah this is definitely coming through so we're the first group of theorists I think to have looked at the data. There's certainly been experimenters who've looked at this data before and also other types of data. Actually, and it's not quite true, the Cardiff and Cardiff-Potsdam Glasgow group did look at some data as well, but not through as many, not through a whole template bank of filters, which is really the new thing that we've done. And it's just that that just makes things worse still, because you're, well, you've got, like, we've only got 600 templates, but Ligon's going to have tens of thousands, and so you're going to have a lot of possibilities for false alarms. So it's interesting, definitely. So do you have any sensitive state of any particular cause for most of the false alarms? Oh, well, yeah, so Bruce did a lot of work before we actually did this data run, In the data stream, you can find a lot of different things. And unfortunately, I don't have any of the audio files.

10:00 I could actually play you some of the sounds. You can identify a lot of these things by just listening to the data. And so if you listen to the data, you get through. Some of them are like clicks. Some of them are like pops. And then others are like a shovel being dragged across the concrete. You can identify a lot of them by just looking at time frequency plots. You get to see a particular type of structure that's associated with the fossil arms. Now, we haven't gone through to sort of veto such things, which to some extent is what you'd like to do, but we haven't done that. Just we figured that we'd do a first cut without doing anything like that and figure out what was going on. but my guess is that you could probably you could probably classify about half or so of the false alarms that you have very easily the trouble is that the ones that have signal to noise ratios only around 8 with a match filter could be quite broadband and they might not be easily picked out even in time frequency but we haven't done enough to really know if that's true that's a goal Are there any other ideas about eliminating false alarms other than, say, coincidence between two detectors? Yeah, actually, probably the best example is something Bruce came up with again, which has the easiest way to explain how this works. So, you know how the match filtering basically is sort of this coherent phase addition of the signal. So the thing is, if you take a look at that in the frequency domain and, say, draw a picture of the amplitude, let's say the characteristic amplitude of the signal, if you divide that up in such a way that there's equal power under each sort of subdivision of it, and you do what that's equivalent to in terms of match filtering, so filtering with filters that are really only like one-eighth of the length of each, so that this expected signal-to-noise out of each filter is only one-eighth of the original, you can actually construct a statistic to ask whether or not the power is properly distributed across the frequency band and whether it's accumulated in the correct manner as you come up.

12:30 And so the idea has been to do that. So once you understand the source, you then ask whether or not the power that you just add up is bigger than some number, but whether or not it was correctly distributed across the entire evolution. And so things like these cliques tend to have very short time duration and not the right bandwidth. And so what tends to happen when you look at them is you tend to get rid of a lot of them. And so it still doesn't clean it up perfectly, but it does improve things. And so you probably are not as bad as an order of magnitude then at that stage. It sort of cuts it down by factors of four or five in terms of the number of fossil arms. So this is a pretty good method. The trouble is we're still not completely sure how bad it is in terms of rejecting signals that are close to the threshold you might want that are real signals. So that's, again, something that needs to be explored further. But it definitely works really nicely. I mean, for a reasonable signal-to-noise ratio, you get to see that it just cleans up the data very nicely and does not reject the signals. ratio is around 10 or 15, which is quite loud. So it's about the only one. I don't know of anything else that's been done in that direction at the moment. What's your feeling about how the experimentalists look at noise like that? I mean, are they saying that's something they're going to be able to deal with on their end, or that has to be done in the data analysis? So there are two answers to that question. First one is to sort of quote Ray Weiss, which unfortunately I can't give you the exact quote, but I can tell you roughly what he said, which was it irritates me that you theorists spend so much time worrying about these noise events. It's up to the experimenters to deal with those in the instrument. That's the nice way of putting it. He was a little more abrupt about it. But the on the other side, to some extent, one thing that they do want to know is they would like the theorists to give them some idea of how clean they need to get each one of these interferometers so that when we do do coincidences, like you say, we don't have so many and then have too many coincidences

15:00 even by accident. And so he wants, he would like to see a proper analysis of that. And I think that's the general view, that the experimenters are going to have to work as hard as is necessary, so that with two interferometers, we don't have to worry about coincident noise events. And that's their general feeling on it. Does it look at the present moment as if the data analysis would remain sort of the purview of the theorists in this case, or once the detectors are up and running? I mean, will the experiments just become more involved? They're starting to. So we had a collaboration meeting last weekend, the one I was telling you about. And the working groups on data analysis are split into three different working groups. One is astrophysical source identification and algorithm development for that particular type of, for sources that we can identify. The second one was issues in statistics, non-gassianity of the noise, thresholding issues, things like this. And then the third was what they called, I think they used the term trigger events or trigger information. The idea being to somehow understand the noise well enough, to characterize the noise so that they could trigger on certain events to think that that could be a real gravitational wave signal or not. So the third one is the one that all the experimenters are trying to get involved in. And really, from what I could gather, again, I haven't had a chance to filter my notes yet, but from what I could gather, they're really taking a fairly experimental point of view in the sense that they want to just use the data to understand the instrument. But the astrophysical source identification is very much a theorist's game at the moment. The only person that I would say that's going to be fairly seriously involved, that is not only a theorist, is Tom Prince. So Tom has attached himself to the project now and is going to be pretty much heavily involved in the data analysis. And I wouldn't classify him as a theorist, but he's not a LIGO experimentalist. But everyone else pretty much in that group is actually a theorist.

17:30 And I think that's likely to stay for a while. The experimenters are just too busy trying to get the instrument built. So I don't think they have time to sit down and really think about the stuff they're not as familiar with. That's my impression. Right. The, well, just since you mentioned the, oh, sort of going wildly off the subject, but you mentioned that you felt that this was an area in which a lot of the young people in the field were sort of keen to get into. and I think Ben and Ben was sort of talking about at the PCGM there was some kind of discussion about relevance and their attitudes towards the LIGO and what that means for them so I was wondering what the sort of attitudes to express there so what that was I don't know if you're aware of this topic it's not a topical group but there's this committee that is forming what's every 10 years to look into what are the major scientific directions that, say, some sub-field of physics should undertake. And this particular, I mean, I think they do it over in fairly broad groups. But this is either the first or only the second time it's ever been done in gravity. And Jim Hartle is heading up the committee, and it's quite broad. So he was there, and he wants to get some feeling from the audience of what they thought were the up-and-coming fields that should be focused on and identified as ones that are requiring clear support from the government and NSAF. Unfortunately, the discussion that was had seemed to degenerate into a discussion about how bad funding is nowadays for people who already have faculty positions. And that's, it just, it was a strange, strange discussion from that point of view. A lot of them seem to have forgotten that LIGO's already being built. And therefore, the train has left the station. It's going to happen. Funding is going to get diverted from certain areas into gravitational wave research

20:00 in that direction, and a lot of them were arguing that maybe we shouldn't fund things like LIGO, but should fund more post-docs. It's clearly a silly thing to say at this stage, since the situation is not going to be that way. So I think there's a concern among these people that there's not that their particular subfields like very, very small areas of research on quantum gravity or just slightly, you know, just things that's sort of on the edge of mainstream, that this stuff is just not going to be supported and consequently we're going to lose a lot of the diversity in the gravitational physics field. And there is a serious, I mean, I personally think there's a serious issue there. that this is going to be the case unless somebody points out that what has to happen is the pot of money has to grow rather than it actually just getting diverted away from all of the traditional stuff. And so my personal opinion on it is that as the community grows that are working on LIGO, there are going to be more people asking for the NSF money that's there And if it doesn't increase, then there's just going to be other people cut who are not working on LIGO-related research. And so I think people need to realize this and then identify how to make sure it doesn't get stupid and get to the level where only LIGO research is being done. Because clearly we need some diversity. But the people who are worried, I don't think they quite saw it that way. they're scared that they're not going to get any money anymore and all they're kind of doing is sticking their head in the sand and going if we don't if we pretend it isn't going to happen or if we argue that it shouldn't happen it won't and this is not going to be the case I think they have to work harder and actually make sure that the government realizes that money needs to be increased somehow I don't quite know how but somehow So... Well, yeah, it's certainly unlikely that they're going to have any luck asking for LIGO to be postponed.

22:30 Exactly. That is just not going to happen. And the thing is, like someone like me who's trying at the moment to get myself set up with a faculty position, the selling point is that LIGO's being built. And you go to a faculty that has a small gravity group or it doesn't have any gravity group at all, LIGO's going to be built and I'm going to work in the field and I'm going to try and connect myself to the collaboration and they go oh that's good it's very sensible for people who have an interest in it and who have some sort of interest in astrophysics and or gravitational waves only or whatever that they actually go that direction it would be silly not to I don't know I don't know what people I've been Scott Pink, having spent four years doing nothing but data analysis-related work. They sound like they want to get away a little bit from it, but the point is they'll never be far, and if they want to come back to it, they will be able to. That's sort of my feeling on that one. So, from what you say, it's a little bit like some of the older, more established people in the field working are unlikely, you know, in the face of it, to benefit much from the impact of LIGO feel threatened by the fact that... Yeah, and even, it's not completely that broad, it tends to be, I mean, you know our field, right? Our field has a few top, like, everybody's good, but there's a few really exceptional top-notch people who could do anything, right? They just choose to do whatever it is they do. and but then there's also an intermediate level an intermediate to lower level of that particular group of people that that tend to have things they do it's not necessarily that they couldn't do anything else but they just don't feel like they can change their new their direction at all and it tends to be those people that are worried naturally because they're the first ones going to be cut if they don't change their direction and so i it's hard i mean it's hard for them What's this like, the first year, say, of meeting at the PCGM, that you were aware of this kind of sentiment, or maybe it just came out because of... For me, I think it's the first year I was aware of it.

25:00 Certainly not anymore. Yeah, I... I mean, I knew that astronomers had this kind of feeling, but it's interesting to see now coming in. In the gravity, yeah. I never, I was never aware of it because everybody that I've ever talked to always goes, well, like when I was thinking about coming here and Ian Moss turns around to me and goes, and I said, I said to him, I said, well, you know, I'm not too sure. I'm not sure if it's the way I want to go, blah, blah, blah. And he sort of went, what? You know, this is the most exciting area of relativity for the next while. And he's right. It really is. it's somewhere that's actually experiencing a lot because it's driven by the experiment how close the experiment is we're moving ahead a lot faster than we would normally and so it is a very exciting area to work in and he's a quantum cosmologist admittedly he's not a US quantum cosmologist so he can afford to be a bit more enthusiastic about it but that's always been my impression that people are really enthusiastic about it I can understand why other people who are worried about their grants disappearing tend to have the opposite point of view and it's just this group that Hartle is chairing, this is organized by AIP or NSR? no, so that I think we can be able to give out for you we might be interested in so I forget who exactly is So, it's the Committee on Gravitational Physics, and he doesn't exactly say who's in charge of it, but... I can dig it. I don't even think it's NSF. I think it's some other government agency or general government committee put together to look into it. I think the report goes to the NSF. I can look into it anyway.

27:30 Yeah, you should check it out. There's a deck of waiting for what should be the top priorities in the NSF about that, assuming a level or declining budget. No question, but discussing any other possibilities. Yeah, that's about right. The possibilities. Okay. The, well, let's see, I suppose while we're still on the subject of all these meetings, what were the, what other interesting things did you, that came out of this LSC meeting in Hanford? So, well, there was one interesting thing, which was a discussion of how we would decide whether or not a gravitational wave had been seen. And this discussion, it was fairly brief, it was only an hour, and so not many people get to say a whole lot in an hour of discussing this type of thing. But Ray put up his criteria for deciding that something was a gravitational wave based on what type of source it was. So whether it was a burst source, stochastic background, or periodic source. And the criterion is pretty much what you'd expect. It's where you have a detection in two or more interferometers that the signal-to-noise ratio scales as it should if the detection is in a two- and a four-kilometer interferometer and that there are no correlations or other noise events in other channels that correlate with this thing. And so everybody was pretty happy with that whether or not you would publish a result based on a single event. A single event, even with coincidence. Even with all this coincidence, plus all lack of correlation in channels. And that brought on a lot of discussion. Some people said, well, I think what you should do is, once you've discovered an event, you should discard all that data for the moment, forget about it completely. And go ahead and wait until you see the next one. And when you see the next one, you should make your decision of whether or not you believe it.

30:00 Of course, then one of the objections, which wasn't really voiced openly at the time to that particular case, is that it's not like you're looking for a given new particle, where when you see the track once, you see the track twice, you go, hey, look, we did see it once, and we forgot about it for a while, we've seen it a second time, now we believe it. the trouble is that if it's an astrophysical source the signatures might be quite different you might get 10 solar mass black holes spiraling into each other or you might get 2.2 solar mass boson stars spiraling into each other and so how do you decide that you've got enough events to believe so there was no real resolution I mean stochastic background other people then said of course well using the the two interferometers that LIGO has is really not enough to state a detection of stochastic background because you're using cross-correlations, and so really you should have a third completely independent device in another location, so LIGO or some, or sorry, VIRGO or GEO or something, and use it to cross-correlate with the data stream, and then that way you have really got two independent detectors. and so this was one of the objections to that one and periodic sources seem to be the one which had the most chance of or had the least stringent demands on it because you can go back and in this case it is like saying okay, we've identified a source at this place on the sky with this frequency let's discard that amount of data and go and look at this new piece of data and see if it's still there And so now you can, with a single interferometer, you can almost do this one. And so there seems to be some thought that even a single interferometer might be good enough for periodic sources, although the gravitational wave community was less certain about that than the people who have done radio astronomy. So who knows how that one's going to go down. But it definitely was a hot topic, and lots of people were, everybody had an opinion on it. Was there any open reference to... None. None.

32:30 To the Barr discovery? There was no open reference, actually, to any bad experience with gravitational waves in the past, Because I thought actually that was what was going to be said, but it wasn't. There were a couple of vague hints in that direction, but they were very vague. And most of the concerns seemed to be more about issues related to how if you didn't work fast enough, your data might become public and people then, someone else might make the first discovery. So how do you temper yourself? how do you make sure your approach is not too conservative so that you lose out in the end, and blah, blah, blah. Clearly there's some sort of, there must be Nobel Prizes on people's minds here as well, I'm sure. So, I don't know. But it was an interesting discussion. That was probably the most interesting of an open forum discussion we had. It sounded interesting. Yeah, it was. It was kind of intriguing to see that. What was, let me see, one other open discussion. We did have one other open discussion. I can't remember it right now. Oh, well, yeah, well, there were some issues in publication policies, internal issue, I guess. It's just trying to figure out who gets to publish what and when, blah, blah, blah, and who should be on the papers and who shouldn't, and things like this. So that looks like it's going to take a little while to iron out, but at least I think the community that are at least aware of this and have become involved in LIGO now in a broader sense than just MIT and Caltech are happy that the LIGO project is beginning to allow other people to have an opportunity to be involved. So, I mean, there's now 23 institutions involved in the first collaboration. So it's quite a big collaboration now. There's about 200 people from those institutions involved in it. So I think it's a pretty, it's a good thing.

35:00 while there were certain aspects of the publication policy that people weren't completely happy with. The general consensus was it's pretty good that LIGO has allowed people the opportunity to contribute and participate in this. And so we can solve the problems with the publication policies down the road. And so I think that's going to happen. But that was the only other major open discussion we had. Was that mostly discussing, like, sort of immediate future publications that are ongoing? Or were they actually discussing... This is further down the road. This is, yeah, this is issues. This is right down to the level of the first announcement of gravitational wave detection. And no one, I think, had any objection to that. It was like, you know, well, we'll have a committee that reports to a single spokesperson for the collaboration. and the committee will decide whether or not the paper is publishable and the first announcement of results based on the data will involve everybody's name in alphabetical order. So it's hard to object to it unless you're somebody who wants to claim they've had a more major contribution and that always looks bad. So that aspect is fine. of things like technical results and who's involved in publications on them. And some of the wordings were a little strange on some of those. But I think that things will get worked out if they're not going to cause any problems at the moment, hopefully. So you mean that the decision was that sort of on the discovery paper there would be like all 200? Yeah. Yeah. So, you know, only one of them can get a prize, or what, three of them, I guess, can get another prize? I think three. I think three is the maximum. I think three is the maximum, although someone was telling me the other day that it has been given to organizations, at least the Peace Prize has been given to Red Cross a number of times. so perhaps it could be given to a collaboration but I doubt it I suspect I'll be trying to identify one person and that will be an interesting piece of sociology but that wasn't discussed at all no, no, that wasn't raised at all which was almost surprising but yeah, I don't know how that one's going to go

37:30 but anyway, the mood obviously was upbeat Extremely. I certainly, well, it's fabulous to go there and see four kilometers of vacuum system in a place, built. The vacuum chambers in the center station are all in place. They're already baked down. They're ready to start installing the optics and lasers into them in, at least in the next little while they are. So, you know, this is really going to happen. And it was actually very inspirational to sort of go along and see the whole setup and realize that, hey, these guys are going to do this. So that was really good. And overall, I think it was a very upbeat meeting. I think the data analysis was the first time that LIGO has put together official groups of people that are going to look into different aspects of it. And the group that I was taking part in, which was the astrophysical source identification one, was extremely positive. It's not completely clear exactly what the role of that group is going to be yet, but it's got a strong working group. for people like Kip, Bernie, Bruce Allen, who else? Well, Alan Wiseman, and a bunch of postdocs as well, and Ana Flanagan. So all those people are there and involved in it. So it's clearly a very strong group, and once this sort of program is identified for it, it can move ahead quite well, I'm sure. So it's pretty exciting to be sort of at the center of that shaping up for it. So, I don't know, I kind of like it. Yeah, that sounds great. Yeah, I know it is. It's pretty damn good. So, I just hope that I can maintain my involvement. That's going to be another story, though, we'll see. I'm sure. But, um, well, let's see the next one. Fascinating. Again, jumping back all of a sudden abruptly to something you said earlier, you're working with Tevye on the Pulsar searches. Right. So I was talking to Bernie Schutz about that, and he was telling me that there is still a possibility

40:00 that you may be limited by the available computing. Oh, I think there's very little doubt that this problem is computationally bad. The strategy that myself and Teviot have adopted in how we have tried to figure out what's the best, tried to figure out how to go at the problem is we ask, suppose you fix the amount of computing power you're going to throw at this, how do you set up your parameter space and your search algorithm in order to make the threshold as low as possible so that you the possibility of detecting most sources for the fixed amount of computing power. It's the only sensible way given that this problem is, as far as I can see, it's going to be always computation bound. It's just, if you try and look over the entire sky for gravitational waves that are nearly periodic, it's just formidable. You need integration weeks to even start to get into interesting regions to place bounds, never mind interesting regions to actually see anything. So it is a huge, huge problem. But I think it's still doable, and I'd say the real chance for it is a combination of searches for known, say, pulsars, searches in the direction of fairly dense clusters of objects in the globular clusters or things like that, the center of the galaxy, stuff like that. That's where the best hope is for detection. Not doing the whole sky, but just doing fairly limited regions where the probability is much higher to have something within a detectable distance. So that's the attitude I'm sort of adopting on that. These R modes could change the picture on that. I mean, it depends. Unfortunately, the signal to noise from them seems to fluctuate every 15 minutes, but as soon as it gets pinned down, if it's even at the worst estimates of the moment, I mean, it's still looking quite promising for, say, enhanced LIGO. And so I think they could be extremely interesting if it's real.

42:30 If nothing else, the thing about that is that unless the understanding of the neutron star structure advances much more quickly, I think, than what Lee seems to think it can, LIGO's going to actually be able to say things about this from an experimental point of view. And it's sort of one of the few, other than binaries, for example, that you can identify that LIGO's definitely got a chance of saying something scientific from the time it's turned on almost. And so I think this is good news. Was there any discussion since we're on this particular topic where while LIGO has all this kind of coverage for pulsars, you're actually going to have to pick your spot from the analysis point of view. Was there any discussion at the LSE meeting about the question of to whom the data will be made available? No, none. My guess about that is that what's going to happen is that anyone in the collaboration will be able to get their hands on the data. The trouble is going to be that the detector is omnidirectional, So how do you, how does someone say, oh, we're not going to look towards the center of the galaxy, we're going to look completely away from it. So, but once you have the data, it's going to be hard to resist looking in good places just because you've been told not to. So I'm not sure what's going to happen there. I don't know, that's going to be a hard question. And it certainly hasn't been decided yet. And I think my impression is that some of the more people who are more experienced in data analysis, like Tom Prince, sort of wants to avoid putting the data out too quickly because he doesn't want the problem of other people discovering the sources while LIGO does nothing but grind out tons of data for these people to work with. So you can understand this, but I don't know how that one's going to shake down. It seems likely to me that there'll be some sort of two-year or three-year rule put on it, but I don't know. I mean, it's going to be hard.

45:00 The data analysis, the interferometers will be turned on sometime around the year 2000. That means that data analysis is required, not fully developed because the first science run is not until the end of 2002, but we need actually for a start of 2002 but we need the data analysis software in place so that we can actually play with it for a while by around the year 2000 which means two years down the road and unless the rate of development increases dramatically I'd be surprised if all of the data analysis issues have been addressed by then and so I think have to need for a teacher to go out to other groups. That's my guess. I'm not in the committee who will be deciding this, so it's, yeah. Well, we mentioned it earlier in the day, so I was going to ask you a bit about Wilson and Matthews, since we've been touched briefly on a new and surprising result that everyone why not touch on one that a new and surprising result that nobody believes pretty much so well I I sort of know the beginning of this story from the contact point of view but I've only been following it at a distance since then so I was kind of curious as to how you became involved so that one that one actually which you're probably aware of this already it seems to me that that one really blew up at Aspen about a year and a half ago. So the result had been known for ages, and Wilson and Matthews had been going around and giving talks about it. But they were at Aspen two years ago, and they made some fairly alarming statements about the evolution of coalescing binary neutron stars, which are bread and butter for LIGO, supposedly, that suggested that while LIGO could still detect them the trouble is that they weren't quite as clean a system as was initially thought so even if you ignored spin they didn't just depend on the masses they also depended on the equation of state somehow

47:30 because this change in the rate of in-spiral was associated with winding up the material in the stars And so it was looking like a real nightmare. But it also, at the time, a lot of the experimenters kind of reacted with the comments like, oh, I never did like the idea of this match filtering thing. And so it became clear that this stuff needed to be looked at at least with a critical eye of nothing else. And some attempt made to identify where the numerics might be going wrong. to say it's completely resolved I think would be unfair the evidence is certainly mounting that there's something in numerics that's poorly understood and not so much a problem with the theory the most likely thing seems to be that I forget how much you know about the stars but the fact that the stars in the system to an equilibrium type of configuration, but they're allowed to have internal motion. And so the most likely cause of the problem is that the internal motion is essentially taking energy. So the energy that's emitted in gravitational waves is not causing the orbits to shrink anymore. Actually, let me rephrase that. Let's start over again. So normally when you think about the evolution of the orbit, you think about the energy in gravitational waves the change in size of the orbit. Given what Wilson and Matthews were claiming about the way the orbit was evolving, it looked like this wasn't the case. It looked like somehow there's an additional thing going on there. And the only additional thing that's evident in their code is that the fluid is somehow getting stirred up. And so there's either energy going into that or it's emitting energy or whatever. So it seems very likely, I think, to me, that that's the source of the problem. But the only way to address that is to actually do a full-time evolution of the problem and not this quasi-stationary evolution that they do because, I mean, they allow these things to come into these nice steady-state rotating configurations. And the question is, as I think it was Kurt, Kurt Guddler,

50:00 who pointed this out even at this ASPEN meeting, the question is, you know, what's the timescale for this to happen? going to happen just as quickly as the orbital timescale or the radiation reaction timescale or is it much longer? And Wilson and Matthews can't answer that within the context of their models. So my feeling on it is until somebody else actually does a numerical simulation in which they allow the matter, the fluid matter, to actually move within itself rather than having it rigidly rotate as it goes around in the binary, until that's done, there's still an agony of doubt that they might be correct, but all the theoretical evidence suggests that it's probably just that they haven't done the full-time evolution. That's my opinion at the moment. So what would be the mechanism that you're talking about, that some energy is being artificially put in to cause the... Either, yeah, either artificially put in or somehow the waves that are being emitted are not, so the gravitational radiation that they're sort of putting in by hand, or I should say taking out by hand, is somehow not being used to change the orbital radius, but is actually being used somehow to cause some sort of, let's call it an instability in the star, to drive some sort of fluid motion in the star. and so rather than if you're only looking then at the evolution, the phase evolution of the two bodies and what radiation they produce you see a frequency evolution that's very different than the post-Newtonian one and so that's that seems to be the type of thing that's going on but it's not I mean it's just not obvious the main thing and again this is this is on hearsay so they claim that if they ask how the change of energy is related to the change in angular momentum of the stars, that the relationship is not dE equals omega dJ.

52:30 And so that's what you'd expect, at least for originally rotating bodies. But that suggests clearly that some energy is going somewhere else, or coming from somewhere else. And that's not clear. Okay, yeah, and so as you say, given that they're putting in, or as you say, taking out the gravitational wave energy by hand, the only way to decide whether it's going where it should go would be to do a forward version. Right, right, yeah. I mean, that's, I think, and I think people are going to start working on that. The evidence I've seen is that Joan Cintrella and Charlie Misner, actually, are both thinking about trying to do some work in that direction. And myself and Jolian have been talking to Joan a little bit about trying to get involved. The NCSA group is clearly interested in doing it, although at the moment they're not gearing up, I think, yet to do the time evolution side of it. The Shapiro? Yeah, so Shapiro's group. Because they did all that quasi-stationary stuff when they relaxed the conformal flatness assumption and things like that. There is also this neutron star... A challenge, that's right. Actually, Shapiro may be involved in that, I guess. It's why I'm assuming it's... I think in some way, but he also has his own approach that he wants to do. Yeah, so, yeah, so there's, I mean, there's plenty of interest in the problem, so it's likely that some progress will be made. I won't claim that they're going to solve it within five years, mind you, because, you know, given the grand challenge, I would manage to do their problem within five years, but, let's see, let's see. So you're thinking of getting involved with Johnson & Johnson? yeah if possible the numerical community it's somewhat difficult to get involved in directly the big projects but Joan seems to be quite accessible in terms of working with her so it just makes sense that if she's looking for someone to work with her hey why not the other people are, I'm sure they'd be happy to have many more postdocs much more than that but I don't know yeah, Joan does seem very comfortable yeah, exactly, I mean I talked to her at Los Alamos last fall and she was

55:00 yeah, she was pretty helpful actually and seemed really interested we've been doing some stuff on trying to get sort of other approximations to binary evolution stuff which you'll probably hear us talk about tomorrow if you come to group meeting because it's going to be a big discussion with Jimmy. But it's fairly simple. It's something Kip has been sort of advocating for a long time. It's just going to the co-rotating frame with the system and evolve on the radiation reaction time scale. So basically do an evolution in F dot over F. F dot over F squared. So that you then have something which evolves like 1 over the radiation reaction timescale is your small parameter, and you do an expansion kind of thing in that to figure out how the evolution proceeds. That's a very simplistic way of saying what we're trying to do, but that's sort of what the idea is. And so we're hoping that Joan's going to pick up on that and try it. well as I say you'll hear plenty more about it tomorrow because Jim's Jim's going to be asking us all sorts of questions well that's great thanks no problems yeah I better let you go back to work because look we've been out yeah actually I'm trying to try and grab Thank you. Thank you.