 So, let me begin with a salutation to Thibaut. Dear Thibaut, wish I could be there at IHS in person to recall some fond memories and convey warm regards to you from me and my family on your 70th birthday. Thank you for sharing your deep insights in physics, being there for me always and your deep friendship over the years. I begin with the prologue after a decade at the Raman Institute starting in 1980, working on problems of GR with some potential astrophysical applications, I felt I should reassess and reorient myself of race. Chandrasekhar's strategy of moving on from one area to a different one every decade made a deep impression on me. Influenced by the school at Karjays, the GR meeting at Stockholm and the meeting at Goa, I felt that early universe cosmology, astrakar variables or gravitational waves were possible areas to explore. So I applied in 1989 to seek opportunities in these three areas. Since I met Thibaut at Karjays and Goa, I applied to him for gravitational waves and I was thrilled when it succeeded. Strangely, it had a Chandrasekhar connection via gravitational waves and the post neutron and formalism. My sabbatical with Thibaut Amur was during 1989-90 when he was himself moving from Newton to IHGS. In a sense, I was Thibaut's first postdoc at the IHGS. It started a new phase in my scientific life which lasted for two decades and led post 2009 to my involvement in Indigo, Ligo India and Ligo India scientific collaboration. I began working on problems related to gravitation radiation, the MPMPN formalism of look Blanche and Thibaut Amur. I still remember that when I reached my office in Newton in September 1989, Thibaut was away on that day, but on my desk was the material I had to start on, a bright yellow dark preprint by Blanche Amur on the PN generation of gravitational waves and a copy of look Blanche's thesis. When I met him the next week, the brief first year, extend the multiple expansion method implemented for massless vector fields to massless tensor fields using symmetric trace-free tensors. Investigates linearized gravity by using HTF methods. By Christmas, we had the first cut results. When finalized, we noticed that the corresponding results using tensor spherical harmonics had incorrect coefficients. This illustrated the algorithmic and computational advantage of symmetric trace-free methods. Thibaut then explained how we had the tools to extend the one PN computation of mass movements to one PN current movements and this led to the second project on the one PN current quadruple with compact support. I took my baby steps in this area working on it when I continued my sabbatical after IHGS that year. On my return to India, to keep in touch with the topic, I generalized it to the one PN current octopole and collaborated remotely with Thibaut. I now come to the theme one in which I collaborated with Thibaut and with Look. When LIGO was funded in the early 90s and efforts to construct accurate coalescing compact binary waveforms started, it was soon realized that far higher-order PN accurate waveforms would be needed to approximate gravitational waves in their final stages of inspiration and merger than in binary pulsar work. Numerical relativity was far from mature and the grand challenge program was started toward this goal. Physical insights were essential to simplify the goals and achieve the required waveforms. For instance, the garden variety in spiraling compact binary would have radiated away their eccentricity and be moving in quasi-circular orbits during the latent spiral. Since matched filtering is sensitive to the phase, it's more important to first control phasing rather than the amplitudes. And finally, in spiral can be treated in the adiabatic approximation as a sequence of circular orbits. This allows one to treat separately the conservative equation of motion problem, the radiation problem, and finally the radiation reaction or the phasing. Following the funding of LIGO, issues related to templates for gravitational wave detection were intensely investigated by Kip's group at Caltech. Kip convened an international meeting to brainstorm this issue and highlight the urgency to address this problem. Look Blanche and I were participants in that meeting and in the discussions there, look expressed the view that the MPMPN formalism could be extended to do this. Soon demonstrated the two PN generation of gravitational waves. Soon after I was visiting Thibault at IHES and this led to the collaboration between the three of us and we completed the two PN phasing for inspiring compact binaries. Using the BDI formalism, which goes from the source moments to the radiative moments, while the canonical moments. The new insight required was a treatment of the cubic nonlinearities. The availability of the two PN equation of motion from the binary pulsar work which Piotr referred to was the contribution from Thibault himself. It really facilitated the computation of the two PN phasing by our MPMPN and independently by Will and Weisman using DIAM. So immediately it was clear to me that the multipolar post-Mincosten formalism matching to a post-Newtonian source that may have appeared like an overkill for the binary pulsar analysis. It's a good example of the advantage a complete and mathematically regressed treatment of a problem can bring for more demanding application that could just be around the problem. Though two PN templates seemed adequate for binary neutron stars, it was clear that for binary black holes the three PN approximation would be necessary. The control of the three PN order as explained till now was more formidable because of the limitation of the Hadamard self regularization which was used. It led to undetermined coefficients in the equation of motion and in the mass quadrupole and the energy flux. And then one started this very interesting work in two groups, the Yenna group and Luk's group. First to attack the equation of motion about which you heard in Piotr's talk in nice detail and the study of the equation of motion and harmonic coordinates by Luk Blanchet and his PhD students Guillaume Fay. Luk also worked on working the general multiple expansion, the computation of the tail of tail and the tail square terms. And therefore when we computed the three PN mass quadrupole using the 3.5 PN flux the 3.5 PN phasing could be given with three with five undetermined coefficients which came from the equation of motion and the mass quadrupole. It's interesting that at this particular point the dire approach could not be extended to three PN. So this was the only calculation which gave the 3.5 PN phasing. Then as explained by Piotr the breakthrough really came in the use of the gauge invariant dimensional regularization in the EDM approach to determine the unknown coefficients and really you had the equation of motion with completely determined. This is followed by the use of dimensional regularization by Blanchet, Damouin and Ressouza-Ferreze to look at the equation of motion in the harmonic coordinate approach. But it was sort of realized that in addition to the regularization, renormalization of the world line would also be needed. As the other two approaches to the equation of motion give consistent result at three PN the effective field theory approach and the strong field point particle approach. And therefore to deal with the radiation at three PN we used a hybrid screen where we first use Hadamard regularization for terms where it could be used and then looked at the ultraviolet divergence by dimensional regularization. The use of the dimensional regularization at three PN again involved Blanchet, Damouin, Ressouza-Ferreze and myself and using the same renormalization of the world line the coefficients could be determined. This work also provided explicit expressions for the mass movements and d dimensions. This work was followed by the computation of the gravitational wave polarizations at three PN by students in my group over the years. All this work was supported by the Indo-French collaboration with Look over two periods. I had to visit the Sephora Projects at support and interactions with Ibo at DIHES. The work up till now involved non-objects so the extension includes spin orbit and spin spin terms really was achieved in Look's group after that. Then came a practical limitation that when we used math tensor, when you had higher multiples it was not very efficient. During that time Jose Maria was visiting IAP and Guillaume Faye decided to use math tensor, X tensor to really deal with these set of symmetries. He rewrote from scratch all the codes and we validated it with all the calculations we had done up till three PN. We then had an efficient set of programs to compute the 3.5 PN mass quadrupole, the 3 PN mass octopole and the associated polarizations to 3.5 PN accuracy. The whole program, the whole set of programs were efficient so that today higher order tidal effects could be calculated. And as I would say the tool to the force of the MPMPN method is the calculation of the 4 PN mass quadrupole which appeared on the archive last week and the 3 PN current quadrupole. Again the program is completed because now you also have the current movements in the d dimensions. So to conclude this particular part the MPMPN formalism I would say is currently the most successful since it can deal with all aspects. The gravitational motion, the radiation field at infinity, the nonlinear effects related to tails and it has evolved over the last 25 years into a consistent algorithmic approach to analytical gravity wave computations for inspiring compact binaries. Currently I would also say that the MPMPN and numerical relativity are the scaffolds which underlie the impressive arcs of the effective one body and phenom template families which are used for detection parameter estimation and tests of gravity which you heard in Alessandra's talk yesterday. I now just move to the next theme in which I collaborated with Thibault and this was the application to gravity wave data analysis. In a series of papers Thibault, Satya Prakash and Myself we provided digested version of the gravity wave analytical results for gravity wave data analysis applications in detection and parameter estimation. In these set of papers we developed tools in equivalent PN families, the notions of effectualness and facialness, the use of window functions, how to go beyond the stationary phase approximation to deal concretively with template construction and to understand template characterizing issue in gravity wave data analysis. With Piotr we also looked at the implication of the effective one body in gravity wave data analysis. All this work culminated in the standard LVK reference for these particular aspects, the paper which is called the BIOPS. All this work led to implications of these wave forms for parameter estimation, for tests of gravity, for investigating the implications of the full wave form for astrophysics and cosmology in the Einstein telescope and in Lisa in my particular group. And therefore over the years it really built a kind of expertise and in a sense is the foundation of the current Indian presence in the test of gravity activity in the LIGO Virgo collaboration. The theme, the third theme I would like to refer to is the how one goes beyond PN expansions by resummation techniques. In this paper with Thibaut and Satya, we looked at the use of resummation techniques like par de approximations to go beyond the post neutron approximation. And this, as you heard in Alexander Stokes, led to the effective one body approach. As you know, the effective one body approach, developing to two streams, the one by Alessandro and Thibaut at IHGS and Bruno and co workers at Maryland and AEI. I was fortunate that during one of my visits, I could collaborate with Thibaut and Alessandro to work on the improved resum templates, which uses the multiplicative decomposition of the multipolar wave form. It was interesting for me because it involved a revisit and completion of the 1990 work on current movements with Thibaut to provide a close form expression for the one PN current movement and the associated HLM for L plus M order. This generalized the work of Kiddur, who had provided these corresponding results for the mass movement and the HLM for L plus M even. This work was sort of generalized and by me and Fujita to give the effective one body polarizations and in the test particle limit up to 5.5 PN order. So anytime, you know, I came to IHGS and work on a particular problem, it had interesting implications in my group back in Raman Rituj Institute. I come back to the last theme and this is the theme related to going beyond the quasi circular inspiral, that is how do you discuss the eccentric inspiral of compact binders. Gravitational radiation reaction decreases the orbital eccentricity to negligible values by the time the gravitation radiation enters the bandwidth sensitive bandwidth of the detectors. For an isolated binary, the eccentricity goes down by a factor of 3 when its semi-major axis is halved. This is the famous Peters result. It's important to track eccentricity because it can help distinguish between the two formation channels for coalescent compact binders. The field boundary evolution or the dynamical formulation formation in dense telescope clusters. The non-zero eccentricity detection of it would be a potential smoking gun for the dynamical formation channel. The tight eccentric binary forms via multi-body interaction and is unable to shed the eccentricity before coalescence. Even a very small eccentricity can produce parameter estimation biases as shown in the work of Fawata. But are there scenarios where eccentric binaries really form? And when looks at the recent literature, one finds there are many such scenarios. For example, in globular clusters, the inner binaries of the hierarchical triplets undergoing cosine oscillations can merge under gravitation radiation reaction. A good fraction of such a system will have an eccentricity of the order of 0.1 when the gravitation rays pass through 10 hertz. If you look at isolated triple systems with a binary black hole inner binary, a few percent would also have large eccentricities about 10 hertz. In the dense environments of galactic nuclei, eccentric binary black binary black holes can be formed. Work by old lady shows that there can be high eccentricities as high as 0.9. And at least 10 percent will have eccentricities of the order of slightly more than 0.1 when they enter the LIGO band. Again, work by Samsung et al. this last year basically showed that 75 percent of the binary black holes formed in galactic nuclei via gravitation capture will have eccentricities of the order of 0.1 or more. So all this basically tells us that eccentric scenarios are very likely and therefore it's a class of problems we should investigate. And therefore, over the years, a group, my group at RRI essentially looked at extending the Peters and Matthews results to eccentric binaries up to 3 pn order because the formalism would be generalized to do that. The phasing of eccentric binaries would need in addition to the conserved energy and the far zone energy flux, the conserved angular momentum and the 3 pn flux of angular momentum also. And this essentially was the thesis topic of my students. For example, the energy flux was calculated by Arun and Hussaila and the angular momentum flux by Sena who then used it to calculate the secular evolution of orbital elements for binaries in posse elliptical orbits. When you look at the technical complications for the eccentric orbits, you realize that there are two, the instantaneous terms and the hereditary terms will need to be dealt with in slightly different ways. For the instantaneous terms in the energy flux, explicit closed form expressions can be given in terms of the dynamical variables related to the relative speed and the relative separation. You can then convert it to and average it over an orbit by using the quasi-keplerian representation, which was given by, you know, Natalie Derival and Thibault long back during the study of the binary pulsar. But since we are working at 3 pn, you would require the 3 pn generalized quasi-keplerian representation. But on the other hand, when you look at the hereditary contribution, you can only write down formal analytical expressions as integrals over the past. If you need more explicit expressions in terms of dynamical variables, you would need a model of the binary's orbit to implement the integration over the past history. When you look at the hereditary contributions, when you look at the circular orbits, you realize that the tail integrals could be evaluated using standard integrals for a fixed non-decaying circular orbit. Directly in the time domain. One could show that the remark past contribution of the tail integrals is negligible and the errors due to the neglect of an inspiral would be at least 4 pn. In the elliptic orbit case, the situation is more complicated. Even after using the quasi-keplerian representation, you cannot do the integrals in the time domain since the multiple moments have a more complicated dependence on time. So the integrals are not solvable in simple closed forms and therefore one evaluates this by going to the frequency domain. And as I explained to you, what really goes in as input is the quasi-keplerian representation where the orbit is sort of written in terms of parameters which can themselves be written in terms of the conserved energy and the conserved angular momentum. Inbuilt into this particular formalism is the double periodicity of the motion in terms of the period related to radial motion and the period related to angular motion which is different because of the period stront procession. It is again fortunate that the inner group of Gerard Schaefer essentially looked at this particular problem and generalized the 2 pn quasi-keplerian representation to the third 3 pn order. The expression is more complicated but it is sufficient to really do what we really need to do. And once you have everything in terms of the quasi-keplerian representation, one looks at the orbital elements which are themselves functions of the conserved quantities and then you can evaluate their variation with respect to time by using the flux balance and averaging the fluxes over an orbit. And this is exactly what was basically done, but how do you go beyond the average? So again, to go beyond the adiabatic secular revolution, one has to look at the equation of motion, the perturbed equation of motion. And once again, Thibault's work using the method of variation of constants comes to the rescue. So in this particular method, you can really deal with the three times scales which are there in this particular problem, the orbital period, the period stront procession and the radiation reaction and go beyond the orbital averaging. So what you do is rewrite the variables, rewrite everything in terms of these equations and the constants in these particular solutions are allowed to vary with respect to time. And when you analyze these particular equations, you realize that these coefficients have two timescales in them, a slow radiation reaction timescale which corresponds to the secular drift and a fast orbital timescale periodic oscillation. So therefore, you put these constants and introduce these two, you know, the average, the barred quantity and the triller quantities. And then you look at these particular equations, the average quantities you average over the orbit and when you subtract it from the force, the driving term, then you get the oscillatory term. So this was something which Thibault had developed in connection with the binary pulsar and in a work with him and Gopukumar, we could arrive at this particular thing. So again, you can see here the continuity which exists between the binary pulsar problem and the phasing for gravitational waves. So as you can see here, the quasi-caplarian representations go order by order and Gerard's Riena group really played an important role in taking this particular forward and we could adapt this to look at the binary pulsar problem. And therefore, this essentially provided generalization of the Peters and Matthews formulas and therefore this work could also be basically completed. So as I said that this, when you look at the results coming from here, you can go beyond the adiabatic approximation and this essentially was the systematic work which was explored by my students in their thesis. So after finishing the quasi-circular case, we will look at the quasi elliptic case and once the phasing is available in terms of the variation of the orbital elements, one wants to know what the waveforms look like. And this was next looked at by my last student, Misra, who looked at the instantaneous terms in the general orbits. The tail terms required much more work and it's only in 2019 with colleagues in Switzerland, we basically could complete the tail contributions also. So essentially we have the tail contribution, the post adiabatic corrections and finally we could also calculate all the non-linear contributions for the eccentric terms. So just to conclude, where are we in regard to eccentric binaries? There is a lot of action people are trying to construct Fourier domain and time domain templates both to detect the eccentric binaries to estimate the parameters of the eccentric binaries. And then look at it, what is the implication of these eccentric template families for things like test of gravity. So I would like to sort of say that the LIGO-VARGO collaboration has discovered binary black holes, binary neutron stars, intermediate mass black holes, neutron star black holes, and asymmetric binary black holes where you can see the higher modes of the system. So eccentric binaries could be the next category of gravity events which are waiting to be discovered. And in a sense, I'll categorize them as being the known unknowns. There are scenarios where they are there and we know that they should exist. The question is when do we first see them? And there is no surprise that the eccentric research is really ramping up. If you look at the literature over the last few years, you will find different aspects of eccentric binaries are being studied. With that I come to my last few comments. Thibaut's views on the problem of motion in the proceedings of Leshush, Karje's 300 years of gravitation and those on the status of the equivalence principle have a lyrical clarity as he reminds us. It's not sufficient to transplant in Einstein's theory the technical steps of Newton's theory, but one needs to transmute within Einstein's conceptual framework the ideas that underlie the technical developments. In a way, I think that a similar natural transmutation seems to underlie mathematical techniques when one goes from pulsar timing to gravity wave phasing via the MPMPM framework. What is unique about Thibaut's body of work related to gravitation waves is the deep insight brought by thinking quantum mechanically in a classical problem. Starting with the use of regularization ideas when attacking the Hulse-Taylor binary, which in a sense led to Mark II approximation methods I would say, and adding reason to normalization for modeling the LIGO-Wergo coalescing compact binaries. Thibaut introduced the par de resumption, effective un-body, dimensional regularization, input from the gravitational self-force and gravitational scattering amplitudes to create most recently an eclectic cocktail. You would this be the Mark III approximation methods we have to just wait and see. And finally, my own initiation into gravity wave research was the sabbatical with Thibaut that allowed me to chip in to model the inspiring compact binaries with look and him and post 2009 to lend a shoulder via Indigo to realize LIGO idea. Thanks Thibaut for the initial opportunity and subsequent support that made all of this possible. The best wishes on your 70th birthday, take care. So we have time for maybe one or two questions. How's there any steer? Online questions. Yes, there is one and I don't know how to give the possibility to talk. You can ask on you then. Thank you very much. I was waiting to be on you. Thank you very much for this fantastic summary of the various developments. My question was about the last topic, eccentric systems. It's precisely in those systems, you could also imagine the spins being misaligned and black holes could have in principle lock spins. And so if you combine then eccentricity and spin evolution, it could be even more complicated. I know that there are some efforts, but what do you think the status of the feed should be in constructing faithful templates for eccentric orbits with large spins? Where should it be? You are hit on the hard problem. So you are right. People have started looking at what are the implications. For example, the eccentricity and spin when they have come together, what kind of systems would have problems and what is there a phase space where the problem would be less severe etc. But as you are rightly saying, since I was talking about the eccentric part, I slurred over that particular part. But what you are saying is one more problem we should be addressing. I really don't have a good answer to, the only way I can think of is probably one can look at regions of phase space where the effect of the spin would be less compared to the effect of eccentricity, the eccentricity high or something like that. For even to do that, we probably need very good templates that can handle both large spins and maybe moderate eccentricity. I think that's a development that is immediately needed, otherwise we might be missing systems, potentially missing systems that could exhibit this sort of dynamics. I agree. So does this pick up again?