deletions | additions       diff --git a/workflow.tex b/workflow.tex  index 06d3aa2..b0ccc4d 100644  --- a/workflow.tex  +++ b/workflow.tex 
...
this intution can then be supplemented by more quantitative calculations. When this understanding is lacking,   one can appeal to analogies, descriptors and machine learning .[ CITE NORMAN'S REVIEW ? arXiv:1601.00709 ].  Designing a proper strategy, thus requires an understanding of the nature of correlations in a given class  of materials. Irrespectively of the degree of correlation, it is natural to divide the workflow of materials design   into three different steps, of course, how to carrry them out in practice will depend on the level of correlation.  Theworkflow of materials design naturally breaks apart in three steps. The  first, and most well-studied, is the calculation of the  electronic structure: structure, namely how to go  from structure to property, i.e.  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 for weakly correlated systems. Sometimes, like  inpredicting  the properties case of  semiconducting materials, where the standard implementations of DFT considerably underestimates the effects  of large classes exchange at low enerties, the first correction in the screened Coulomb interactions, namely the GW self energy, $\Sigma_{GW} - {V^{LDA}}_{XC}$  is required to get good results.   The basic feature  of materials. correlated materials is their electrons cannot be described as non-interacting particles. Often, this occurs because the material contains atoms with partially-filled $d$ or $f$ orbitals. The electrons occupying these orbitals retain a strong atomic-like character to their behavior, while the remaining electrons form bands; their interplay poses special challenges for theory. Consequently current implementations of DFT cannot describe their electronic structure accurately. This led to the development of combinations of DFT and dynamical mean field theory (DMFT) which can treat and the GW approximation.  LDA+U can be viewed as a static limit of LDA+DMFT ( when the impurity solver used is the Hartree Fock approximation) and works in the  presence of magnetic and orbital order.   %  In principle, lattice %lattice  properties can be computed as well, such as phonon vibrational modes, stress %stress  tensors and thermal expansion coefficients, but we simply call this step ``electronic %``electronic  structure''. The second step is structure prediction: given a fixed chemical  composition--take Ce$_2$Pd$_2$Sn for example--predict its ground state crystal