Review changes requested by Parth and Vassil

docs/userDocs/source/user/FloatingPointErrorEstimation.rst

@@ -13,9 +13,9 @@ develop important new lossy compression algorithms.
 How does Automatic Differentiation (AD) fit into this?
 ======================================================
 
-AD helps evaluate the exact derivative of a function. AD applies the chain rule
-of differential calculus throughout the semantics of the original program. In
-the context of FP error estimation, this research uses:
+AD helps evaluate the exact derivative of a function. AD applies the
+differential calculus chain rule throughout the semantics of the original
+program. In the context of FP error estimation, the implementation relies on:
 
 - **Reverse-Mode AD** (as opposed to Forward-Mode AD), since it provides the
   derivative of the function with respect to all intermediate and input
@@ -25,13 +25,13 @@ the context of FP error estimation, this research uses:
   most of the work at compile time. The Clad Framework that is used in this
   research also uses source transformation.
 
-   **Clad** is implemented as a plugin for the Clang compiler. It inspects the
-   internal compiler representation of the target function to generate its
-   derivative. Clad requires little or no code modification, supports a growing
-   subset of C++ constructs, statements and data types, it enables efficient
-   gradient computation for large and complex codebases, and is deeply
-   integrated with the compiler, allowing automatic generation of error
-   estimation code.
+**Clad** is implemented as a plugin for the Clang compiler. It inspects the
+internal compiler representation of the target function to generate its
+derivative. Clad requires little or no code modification, supports a growing
+subset of C++ constructs, statements and data types, it enables efficient
+gradient computation for large and complex codebases, and is deeply
+integrated with the compiler, allowing automatic generation of error
+estimation code.
 
 Where does the FPEE logic reside?
 =================================
@@ -52,15 +52,16 @@ files:
 
 
    Above files include a lot of useful documentation in the form of code
-   comments. 
+   comments. Please view the `Doxygen Documentation`_ and the `Clad Readme`_
+   for more details.
 
 How does the FPEE Logic work?
 =============================
 
-While parsing the code using Clad, if it encounters a floating point
-variable,it needs to be registered into the system (to accumulate relevant
-errors against that variable). Next, the Error Estimation Calcualtion Formula
-(Error Model) needs to be built (using ``EstimationModel.h``).
+While parsing the code using Clad, if it encounters a floating point variable,
+it needs to be tracked (to accumulate relevant errors against that variable).
+Next, the Error Estimation Calculation Formula (Error Model) needs to be built
+(using ``EstimationModel.h``).
 
 ``EstimationModel.h`` contains the information needed to calculate the estimate 
 value of the error. It is highly customizable (e.g., you can plug in your 
@@ -188,4 +189,8 @@ forward block.
 
 .. _FP-EE Proposal: https://compiler-research.org/assets/docs/Garima_Singh_Proposal_2020.pdf
 
-.. _CHEF-FP Examples Repo: https://github.com/grimmmyshini/chef-fp-examples
+.. _CHEF-FP Examples Repo: https://github.com/grimmmyshini/chef-fp-examples
+
+.. _Clad Readme: https://github.com/vgvassilev/clad#floating-point-error-estimation---cladestimate_error
+
+.. _Doxygen Documentation: https://clad.readthedocs.io/en/latest/internalDocs/html/index.html