diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..07fc0e5
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1 @@
+gh-build
diff --git a/source/_themes/uwtheme/layout.html b/source/_themes/uwtheme/layout.html
index 71b8721..259a2f1 100644
--- a/source/_themes/uwtheme/layout.html
+++ b/source/_themes/uwtheme/layout.html
@@ -15,6 +15,7 @@
       <ul id="menu-main" class="sf-menu">
         {%- block rootrellink %}
         <li class="menu-item"><a href="{{ pathto(master_doc) }}">{{ shorttitle|e }}</a></li>
+        <li class="menu-item"><a href="DAGMC">DAGMC</a></li>
         {%- endblock %}
         {%- for rellink in rellinks|reverse %}
         <li class="menu-item" {% if loop.first %}style="margin-right: 10px"{% endif %}>
@@ -33,7 +34,7 @@
 {%- macro uwsidebar() %}
 <div id="sidebar">
  <div id="sidebar_section"></div>
- <div id="sidebar_section">{{ toctree() }} </div>
+ <div id="sidebar_section">{{ toc }} </div>
 </div>
 {%- endmacro %}
 
diff --git a/source/_themes/uwtheme/static/default.css_t b/source/_themes/uwtheme/static/default.css_t
index 57fd92c..3f18ed1 100644
--- a/source/_themes/uwtheme/static/default.css_t
+++ b/source/_themes/uwtheme/static/default.css_t
@@ -49,8 +49,6 @@ p { margin-top: 0.75em; }
 
 .highlight-python { padding: 4px; border: 2px solid #000000; margin-left: 2em; }
 .highlight { background: #ffffff; }
-pre { font-family: courier; font-size: 80%;}
-.pre { font-family: courier; font-size: 90%;}
 
 body { line-height: 1; }
 /* ol, ul { list-style: none; }*/
@@ -90,6 +88,9 @@ strong { font-weight: bold; }
 em { font-style: italic; }
 li h3, li p, li a { font-size: 1em; }
 
+.headerlink { color: #ffffff; }
+.headerlink:hover { color: #b70101; }
+
 html {
   color: #000;
   background: #FFFFFF url(../_static/images/background.jpg) repeat-x top;
@@ -97,7 +98,7 @@ html {
 
 body {
   margin: 40px auto 0;
-  width: 960px;
+  width: 1260px;
   color: #8E8979;
   background-color: transparent;
   font: normal 100%/1.5 Verdana, sans-serif;
@@ -147,10 +148,10 @@ body {
   font-family:georgia,Times New Roman,serif;
   font-size: 2em;
   position: absolute;
-  top: 26px;
+  top: 16px;
   left: 87px;
-  width: 343px;
-  height: 34px;
+  width: 500px;
+  height: 72px;
   overflow: hidden;
 }
 .home #siteTitle {
@@ -317,7 +318,7 @@ body {
 /*layout*/
 
 #c_container {
-  width: 958px;
+  width: 1258px;
   border-right: 1px solid #666666;
   border-left: 1px solid #666666;
   border-bottom: 2px solid #666666;
@@ -685,7 +686,7 @@ body {
 }
 .page #content {
   margin: 0 7px 0 0;
-  width: 640px;
+  width: 940px;
   padding-top: 40px;
 }
 .page #sidebar {
@@ -699,11 +700,23 @@ body {
   width: 250px;
 }
 
+.align-right {
+  float: right;
+  padding-left: 10px;
+}
+
 #sidebar_section {
   margin: 0 0 0 10px;
+  position: fixed;
+  width: 245px;
 }
 
-#sidebar ul, #sidebar ol { list-style: none; }
+#sidebar ul, ol { list-style: none; }
+
+#sidebar ul ul { margin: 0 0 0 1em; list-style-type: circle; }
+#sidebar ul ul ul { margin: 0 0 0 1em; list-style-type: square; }
+
+ol { list-style-type: lower-alpha; margin: 0 0 0 3em;}
 
 .page #sidebar .related_posts {
   margin: 16px 16px 32px;
@@ -726,7 +739,20 @@ body {
 
 h1, h2, h3 {margin-top: 0em; color:rgba(50, 04, 3e, 1); font-family:"Myriad Pro",Trebuchet,Arial,sans-serif;}
 
+h2 {font-style: italic; text-decoration: underline; margin-top: 1em;}
+
+h4 {font-weight: bold; margin-top: 1em;}
+
+pre {font-family:Courier; background: #ffefef; border: 1px solid #ffaaaa; margin: 1em;}
+.pre {font-family:Courier; background: #ffefef; border: 1px solid #ffaaaa;}
 
+dl.docutils dt {
+   text-decoration: underline;
+   margin: 1em 0 0 0;
+}
+dl.docutils dd {
+   margin: 0 0 0 1em;
+}
 /*Innovations single*/
 
 .single .post .label {
@@ -750,9 +776,10 @@ table {
 margin-top: 16px;
 margin-bottom: 2em;
 text-align: left;
-border: 1px solid #666;
+border: 0px solid #666;
 border-collapse: collapse;
 font-size: .9em;
+width: 100%;
 }
 caption {
 text-align: left;
diff --git a/source/index.rst b/source/index.rst
index edf541e..b7124bd 100644
--- a/source/index.rst
+++ b/source/index.rst
@@ -7,10 +7,10 @@ the same way, fusion shielding is necessary to protect the system's
 components from the radiation produced by the fusion system they are
 supporting.
 
-This project includes a variety of tools, methodologies and workflows
-that are relevant for the nuclear analysis of complex system,
-including fusion power plant designs, accelerator irradiation
-facilities and critical reactor systems.
+This project includes a variety of `tools, methodologies and workflows
+<map.html>`_ that are relevant for the nuclear analysis of complex system,
+including fusion power plant designs, accelerator irradiation facilities and
+critical reactor systems.
 
 Single Physics Analysis Tools
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
diff --git a/source/map.rst b/source/map.rst
new file mode 100644
index 0000000..e5df521
--- /dev/null
+++ b/source/map.rst
@@ -0,0 +1,286 @@
+Concepts and Tools for Radiation Shielding Analysis
+====================================================
+
+Over recent years, the set of concepts, methodologies and tools that are
+available for the analysis of radiation transport and effects in shielding
+systems has grown dramatically.  Here we attempt to describe the map of these
+capabilities with mention of specific tools that have gained, or are gaining,
+broader application.
+
+Fundamental Concepts
+--------------------
+
+The fundamental analysis need for any radiation shielding problem is
+**radiation transport**, for which two primary approaches are in widespread
+use: Monte Carlo radiation transport and discrete ordinates radiation
+transport.  Many of the effects of radiation can be calculated directly by the
+tools that implement these capabilities, but determining the induced
+radioactivity caused by neutron irradiation requires **activation analysis**
+to be used in conjunction with radiation transport.
+
+Monte Carlo Radiation Transport
+++++++++++++++++++++++++++++++++
+
+Monte Carlo radiation transport directly simulated the stochastic behavior of
+particles that leave a source and travel around the engineered system.  If
+enough particles are simulated using independent sequences of pseudo-random
+numbers, then their mean behavior can be taken as representative of the real
+system.  The single biggest advantage of Monte Carlo radiation transport is
+its ability to model the entire phase space of radiation transport (space,
+direction and energy) as continuous variables.  Continuous modeling of space
+allows for
+arbitrarily complex geometric models with no need for approximation.  Taken
+together with continuous treatment of direction, this also means that
+following particles through complex streaming paths is also possible with
+these methods.  Finally, continuous treatment of energy allows less
+approximation of resonance behaviors for particles that are slowing down
+through that energy region.  This continuous modeling capability is the
+primary reason that Monte Carlo radiation transport is widespread for this
+kind of analysis.
+
+The biggest drawback of Monte Carlo techniques is in the modeling of deep
+penetration problems with substantial shielding.  Since such systems are
+designed to prevent most particles from reaching the far side of such shields,
+and Monte Carlo radiation transport essentially simulates the behavior of
+these particles, then very few simulated particles will penetrate the shield.
+This, in turn, results in very low quality results behind thick shields, since
+the quality of statistical results relies on a substantial number of
+contributions.  While it is possible to simply increase the total number of
+simulated particles, a number of schemes have been introduced, collectively
+known as *variance reduction* techniques to help improve this situation.
+
+The most widely used tool for Monte Carlo radiation transport in neutron-based
+systems (e.g. fission, fusion, some accelerators) is MCNP (Monte Carlo
+N-Particle), versions 5 and 6, written & maintained by Los Alamos National
+Laboratory.  Many other tools exist and some see larger usage in other kinds
+of systems, particularly high energy physics: Fluka (CERN) Geant4 (CERN),
+Monaco (ORNL), SHIFT (ORNL), Tripoli (CEA).
+
+Discrete Ordinates Radiation Transport
++++++++++++++++++++++++++++++++++++++++
+
+While a variety of so-called deterministic methods exist, the discrete
+ordinates method is the most commonly used in radiation shielding analysis.
+This method discretizes space, direction and energy, and then solves a
+difference formulation of the radiation transport equation.  The most
+important advantage of deterministic methods is their ability to find global
+solutions with uniform quality.  Discretization in space introduces some
+approximations that can be largely overcome by refining the spatial
+resolution.  Discretization in direction, however, results in two related
+effects that can undermine the quality of results.  First, for transport in
+large (near) vacuum regions, particles are transported only along a discrete
+set of directions introducing so-called *ray effects*.  Furthermore, for
+arbitrarily oriented streaming paths and/or dog-legs, there may be no discrete
+angle that is aligned with such a path, inhibiting accurate streaming
+solutions.
+
+The biggest benefit of discrete ordinates techniques is that the quality of
+the solutions is independent of the thickness of the shielding region, and
+results can generally be found much more quickly than with Monte Carlo
+radiation transport.
+
+Use of discrete ordinates software is less wide-spread with less
+concentration in any single tool.  Some tools with user communities that
+extend beyond their developer communities include:
+
+* PARTISN (LANL)
+* DORT/TORT (ORNL)
+* Attila (Varian)
+* Exnihilo/Denovo (ORNL)
+
+Activation Analysis
++++++++++++++++++++++
+
+Activation analysis generally combines the results of neutron radiation
+transport with cross-section data to determine the transmutation rates of all
+the nuclides in a material.  The resultant nuclides may also undergo
+transmutation or decay, and so on.  The equations that describe the evolution
+of the isotopic composition are a relatively simple set of first order ODEs,
+often referred to as the Bateman equations.  However, this is generally a very
+stiff set of equations, as governed by the transmutation and decay rates that
+have time constants that span from milliseconds to millions of years.  There a
+variety of mathematical and computational approaches to solving matrix
+exponential that arises from this set of equations.
+
+The majority of tools available for isotopic inventory calculations are
+derived from fission burnup codes, solving the inventory problem at a single
+point in space, that is, a single material with single neutron flux spectrum.
+They generally employ different computational approaches.  Some commonly used
+tools are ORIGEN (ORNL), FISPACT (CCFE), and CINDER (LANL).  ALARA (UW) is a
+special purpose activation tool that solves for multiple points in space
+simultaneously.
+
+Advanced Capabilities
+---------------------
+
+In recent years, attempts to analyze the most challenging radiation shielding problems have resulted in more advanced capabilities beyond the fundamental concepts and tools above.  In some cases, the details of complex shielding goeomtries emerged as the most complex aspect of assessing shielding performance, necessitating support for CAD-based analysis.  The reliance on Monte Carlo modeling for thick shielding problems also cast a focus on variance reduction techniques, and in conjunction with complex gometries, it became necessary to automate the process of generating paramters for variance reduction.
+
+
+Geometry Representations and CAD-based Analysis
++++++++++++++++++++++++++++++++++++++++++++++++
+
+Since the early 2000's, a number of different teams have pursued the ability
+to perform radiation transport and activation on geometries that are based on
+computer-aided designs (CAD).  For Monte Carlo radiation transport, the most
+common approach (Germany/KIT/McCAD, China/FDS/MCAM, USA/Raytheon/TopAct, etc)
+is to simplify the CAD models to include only the same geometric surfaces that
+can be represented in the constructive/combinatorial geometry representations
+that are native to the standard Monte Carlo tools, decompose the CAD regions
+into more simpler regions, and generate an input file in the standard format
+for those tools.  An alternative approach (USA/University of Wisconsin/DAGMC)
+is to modify the Monte Carlo software itself to perform its geometry queries
+directly on CAD-based representation of the geometry.  
+
+For discrete ordinates radiation transport, all tools but Attila rely on
+structured orthogonal grids for their geometry representation.  This requires
+some degree of approximation, either by choosing a dominant material for each
+voxel in that mesh or by mixing the materials in each voxel according to their
+volume fraction in that voxel.  CAD-based geometries can be queried according
+to either of these approaches to populate a structured grid representation of
+the geometry.
+
+Attila implements the discrete ordinates method using a finite element
+approach on a tetrahedral mesh, allowing the mesh to conform to arbitrarily
+complex shapes of the geometry and for each mesh element to contain a single
+material.  Attila is able to directly import CAD-based geometry in order to
+generate mesh for its transport algorithm.
+
+Frequently, the CAD models available for a given system have been generated
+without engineering analysis in mind.  As such, these models include details
+that are unnecessary for nuclear analysis as well as overlaps of different
+components.  The cleaning and repair of CAD-based geometries is currently not
+automatable and benefits greatly from experience working with previous models
+for this purpose.
+
+Automated Variance Reduction Techniques
++++++++++++++++++++++++++++++++++++++++
+
+As indicated above, Monte Carlo radiation transport can be challenged by
+problems that involve deep penetration of thick shields, a characteristic of
+most shielding problems.  A number of techniques, generally known as variance
+reduction techniques, exist to accelerate these calculations for a specific
+response.  However, as the problems become more complex, it becomes difficult
+to manually configure those techniques for maximum benefit.  Over the last
+15-20 years, a number of approaches have been developed to automate the
+configuration of variance reduction.  Some approaches (MCNP weight window
+generator, MAGIC) use iteration, where one round of Monte Carlo simulation is
+used to improve a guess for the variance reduction parameters for the
+successive round.
+
+An alternative approach is to use the results of a deterministic calculation
+to provide the parameters for variance reduction.  This is effective because
+the deterministic calculation is generally much faster than a Monte Carlo
+calculation with uniform quality of results.  ORNL has implemented the CADIS
+and FW-CADIS methodologies in the tool ADVANTG for this purpose, when used in
+conjunction with MCNP.  CADIS will optimize variance reduction parameters for
+a single response while FW-CADIS can be employed to optimize variance
+reduction parameters for many responses, including global solutions.
+
+Types of Analysis
+------------------
+
+These tools are used in various combinations to accomplish different types of
+analysis.  Depending on the combination of tools, different skills and
+expertise are necessary.
+
+Radiation transport, damage, heating, and tritium breeding
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+A basic radiation transport calculation can be configured to provide a variety
+of nuclear responses in addition to the neutron flux distribution.  These
+responses must be linear combinations of the neutron fluxes, typically based
+on cross-sections or similar nuclear data.  It is common to estimate radiation
+damage in terms of displacements per atom and 
+hydrogen (H) and helium (He) gas production, as well
+as heating from neutrons and various secondary particles.  In fusion systems, it is also
+of interest to estimate the number of tritium atoms that are produced in
+breeding regions.
+
+Increasingly, these analyses rely on CAD-based geometries to define the
+original geometry.  In addition, when relying on Monte Carlo radiation transport, 
+most of these analyses benefit from the use
+of automated variance reduction techniques.  Finally, results tabulated on 
+high-fidelity mesh are frequently used as the source term for other physics analysis.
+For example, high fidelity nuclear heating results may be used to drive 
+thermo-mechanical analysis.
+
+Basic expertise: 
+
+* Monte Carlo radiation transport, usually MCNP5/6, or
+* Discrete ordinates radiation transport, such as ATTILA
+
+Additional expertise: 
+
+* CAD-model preparation when using CAD-based approaches.
+* Automated variance reduction when necessary
+
+Shutdown Dose Rate Analysis
+++++++++++++++++++++++++++++
+
+Another important consideration in any system with neutrons is the impact of
+transmutation caused by those neutrons.  The most common consequence is the
+activation of structural components that results in a distributed photon
+source that dominates the radiation environment when the main system stops
+operating.  The shutdown dose rates that arise from these photons can be
+calculated throughout the facility.
+
+The most rigorous approach for this analysis is to first perform neutron
+transport to determine the spatial distribution of the multi-group neutron
+flux throughout the system.  At each point in space for which the neutron flux
+is known, an activation problem is performed to determine the photon source at
+that point.  Finally, the superposition of all photon sources is used for a
+photon transport problem in which the spatial distribution of the photon dose
+is determined.  Because of its separate transport steps for neutrons and
+photons, this approach is often referred to as the "rigorous 2-step approach"
+(R2S). For very large systems, it may be necessary to use spatial
+decomposition of the neutron flux results and photon source results.
+
+In most cases, a large mesh is used for both distributions of neutron flux and
+consequent photon sources, and perhaps a different mesh for the photon dose
+distribution.  This results in up to 1 million separate activation calculation
+points, and thus requires automation to be tractable.  A semi-automated R2S
+approach has been implemented robustly as part of the PyNE toolkit, coupling
+MCNP to ALARA, with support for CAD-based geometries.
+
+Basic expertise: 
+
+* Monte Carlo radiation transport, usually MCNP5/6, with  
+* Activation analysis using FISPACT, ALARA, or ORIGEN, or
+* Discrete ordinates radiation transport with internal transmutation, such as ATTILA.
+
+Additional expertise: 
+
+* CAD-model preparation when using CAD-based approaches.
+* Automated variance reduction may be used for the photon dose phase.
+
+Experimental work is underway (ORNL/UW) to determine how to best use automated
+variance reduction for the neutron transport step.
+
+Activation of Flowing Systems
++++++++++++++++++++++++++++++
+
+In some cases, acitvation of liquids that flow throughout the system becomes a substantial issue for assessing the radiation environment.  Most commonly, those liquids are coolants used to remove heat, but they can fill other roles.  Analysis of these sytems requires tools and capabilities beyond normal activation analysis.  The flowpath of the coolant must be analyzed to determine the residence time in different irradiation environments.  These neutron fluxes and residence times for each control volume must then be combined in an activation analysis tool that permits the modeling of potentially arbitrary irradiation schedules.
+
+Basic expertise:
+
+* Monte Carlo radiation transport, usually MCNP5/6.  
+* Activation analysis using FISPACT, ALARA, or ORIGEN
+
+Additional expertise:
+
+* Activation in complex irradiation histories with varying flux spectrums and
+  irradiation times
+* Fluid dynamics simulation of flow through complex structures
+
+
+Resource Needs
+---------------
+
+As the fidelity and complexity of models has increased the computational
+resource requirements have also grown.  Virtually all of the fundamental
+tools, advanced capabilities, and analysis types requires 1000s or 10,000s of
+cpu-hours for each calculation.  For deterministic radiation transport and
+some applications of Monte Carlo radiation transport, this requires parallel
+computing clusters with high-performance network interconnects.  Other
+applications of Monte Carlo radiation transport, simple require access to a
+large number of computers whether local or cloud resources.