We fit a flexible double power-law (`αβγ') model to the stellar density fields observed for the Milky Way's known dwarf spheroidal satellite galaxies. We show that where standard criteria for model selection are decisive, the αβγ model is favored over the special case of the Plummer model, and also over the model and its special case, the exponential. The Milky Way's dSph population exhibits a diverse range of stellar density profile shapes, as quantified by the values we infer for outer and inner power-law indices β and γ. Several of the most massive dSphs (e.g., Eridanus II, Fornax, Leo I, Leo II) have steeply-declining outer profiles, with β≳ 8; others (e.g., Sextans, Boötes I) fade slowly, with β≲ 4. The inner profiles of dSphs with stellar mass ≳ 10^5 M_⊙ are consistent with `cores' of uniform stellar density (γ≈ 0). At lower masses the slopes of inner density profiles are poorly constrained, except in a few ultrafaint dSphs (e.g., Hercules, Ursa Major II) where we infer steep stellar cusps, with γ≳ 1.5. Owing to the αβγ model's flexibility, the inferred halflight radii and total stellar masses are significantly more uncertain than previous estimates, with halflight radii larger by up to an order of magnitude in some cases. Finally, we demonstrate that allowing for flexibility in the shape of the stellar density profile is crucial for dSph mass modeling, where systematic errors associated with choice of stellar density profile can outweigh random errors in the observed velocity dispersions.

de Grijs and Bono compiled 211 independent measurements of the distance to galaxy M87 in the Virgo cluster from 15 different tracers and reported the arithmetic mean of a subset of this compilation as the best estimate of the distance. We compute three different central estimates -- the arithmetic mean, weighted mean, and the median -- and corresponding statistical uncertainty for the full data set as well as two sub-compilations. We find that for all three central estimates the error distributions show that the data sets are significantly non-Gaussian. As a result, we conclude that that the median is the most reliable of the three central estimates, as median statistics does not assume Gaussianity. We use median statistics to determine the systematic error on the distance by analyzing the scatter in the 15 tracer subgroup distances. From the 211 distance measurements, we recommend a summary M87 distance modulus of $31.08^{+0.04}_{-0.05}$ (statistical) ${}^{+0.04}_{-0.06}$ (systematic) mag, or combining the two errors in quadrature $31.08^{+0.06}_{-0.08}$ mag, rounded to $16.4^{+0.5}_{-0.6}$ Mpc, all at $68.27\%$ significance.