Fed Bianco (NYU) gave a great talk today about eta Carinae and light echos, mass loss, and the progenitors of stripped supernovae. Amazing stuff, especially the light echos. One thing the whole audience was excited about was the idea that if you had truly high-quality synoptic imaging like LSST will provide, you might be able to do some amazing simultaneous modeling of the dust (the reflector) and time-variable point sources (the radiation field). My intuition that this "blind deconvolution" problem could be solved is based on the facts that (a) the radiation source is a very sparse set of point sources and (b) the dust lies in a (fairly) continuous, (fairly) isotropic distribution, while the light echo geometry is the surface of an expanding anisotropic ellipsoid. I bet that one isn't in the LSST Project Book.
In the afternoon, I spoke with Vakili about a possible huge speed-up in Cramer-Rao-saturating point-source centroiding methods, and I spoke with Huppenkothen about gamma-ray bursts in the Fermi data. On the latter, she is planning on putting in a data analysis proposal which would develop further our "Magnetron" project for modeling bursty bursts.