We consider the dissipation by Fermi acceleration of magnetosonic turbulence in the Reynolds Layer of the interstellar medium. The scale in the cascade at which electron acceleration via stochastic Fermi acceleration (STFA) becomes comparable to further cascade of the turbulence defines the inner scale. For any magnetic turbulent spectra equal to or shallower than Goldreich-Sridhar this turns out to be $\ge 10^{12}$cm, which is much larger than the shortest length scales observed in radio scintillation measurements. While STFA for such spectra then contradict models of scintillation which appeal directly to an extended, continuous turbulent cascade, such a separation of scales is consistent with the recent work of \citet{Boldyrev2} and \citet{Boldyrev3} suggesting that interstellar scintillation may result from the passage of radio waves through the galactic distribution of thin ionized boundary surfaces of HII regions, rather than density variations from cascading turbulence. The presence of STFA dissipation also provides a mechanism for the non-ionizing heat source observed in the Reynolds Layer of the interstellar medium \citep{Reynolds}. STFA accommodates the proper heating power, and the input energy is rapidly thermalized within the low density Reynolds layer plasma.

We propose a new two-stage model for acceleration of electrons in solar flares. In the first stage, electrons are accelerated stochastically in a post-reconnection turbulent downflow. The second stage is the reprocessing of a subset of these electrons as they pass through a weakly compressive fast shock above the apex of the closed flare loop on their way to the chromosphere. We call this the "shock reprocessing" model. The model reproduces the energy dependent arrival time delays observed for both the pulsed and smooth components of impulsive solar flare x-rays with physically reasonable parameters for the downflow region. The model also predicts an emission site above the loop-top, as seen in the Masuda flare. The loop-top source distinguishes the shock reprocessing model from previous models. The model makes testable predictions for the energy dependence of footpoint pulse strengths and the location and spectrum of the loop-top emission, and can account for the observed soft-hard-soft trend in the spectral evolution of footpoint emission. Our model highlights the concept that reconnection is an acceleration environment rather than a single process. Which combination of processes operate may depend on the initial conditions that determine, for example, whether the reconnection downflow is turbulent. The shock reprocessing model comprises one such combination.

We reexamine stochastic Fermi acceleration (STFA) in the low energy (Newtonian) regime in the context of solar flares. The particle energization rate depends a dispersive term and a coherent gain term. The energy dependence of pitch angle scattering is important for determining the electron energy spectrum. For scattering by whistler wave turbulence, STFA produces a quasi-thermal spectrum. A second well-constrained scattering mechanism is needed for STFA to match the observed 10-100keV non-thermal spectrum. We suggest that STFA most plausibly acts as phase one of a two phase particle acceleration engine in impulsive flares: STFA can match the thermal spectrum below 10kev, and possibly the power law spectrum between 10 and 100keV, given the proper pitch angle scattering. However, a second phase, such as shock acceleration at loop tops, is likely required to match the spectrum above the observed knee at 100keV. Understanding this knee, if it survives further observations, is tricky.