deletions | additions       diff --git a/workflow.tex b/workflow.tex  index d9e9144..d976c4f 100644  --- a/workflow.tex  +++ b/workflow.tex 
...
\section{Material Design Workflow}  The workflow of materials design naturally break breaks  apart into in  three steps. The first, and most well-studied, is electronic structure: given a crystal  structure, compute its electronic properties, such as gap size, magnetic  ordering, and superconducting transition temperature. Here, density functional  theory, and its extensions to correlated materials, has been quite successful  in predicting the properties of large classes of materials. In principle,we  can compute  lattice properties can be computed  as well, such as phonon vibrational modes, stress tensors and thermal expansion coefficients, but we simply call this step  ``electronic structure''.  The second step is structure prediction: given a fixed chemical composition,  say Ce$_2$Pd$_2$Sn, predict composition--take Ce$_2$Pd$_2$Sn for example--predict  its ground state crystal structure. Structure  prediction Generically, atoms are placed in a unit cell and a chosen algorithm  is used to efficiently traverse the space of atomic configurations and cell  geometries to arrive at the lowest energy structure. This step  requires having an accurate method for producing the energy of a given configuration of atoms. For weakly correlated materials, DFT has been quite successful~\cite{Fredeman_2011, Gautier_2015} successful at providing  total energies which enable accurate comparisons of the generated structures:  successes include the prediction of a new compounds in the Ce-Ir-In  system~\cite{Fredeman_2011}, as well as predictions of over 50 new 18-valent  ternary semiconductors~\cite{Gautier_2015}. We argue that  a) Structure to Property [ compute Gaps, Tc’s etc.]