now we also have a examples/ directory for self-contained examples,

and this is already used to document DynBlock

Eigen/Core/DynBlock.h

@@ -26,6 +26,23 @@
 #ifndef EIGEN_DYNBLOCK_H
 #define EIGEN_DYNBLOCK_H
 
+/** \class DynBlock
+  *
+  * \brief Expression of a dynamic-size block
+  *
+  * This class represents an expression of a dynamic-size block. It is the return
+  * type of MatrixBase::dynBlock() and most of the time this is the only way this
+  * class is used.
+  *
+  * However, if you want to directly maniputate dynamic-size block expressions,
+  * for instance if you want to write a function returning such an expression, you
+  * will need to use this class.
+  *
+  * Here is an example illustrating this:
+  * \include class_DynBlock.cpp
+  * Output:
+  * \verbinclude class_DynBlock.out
+  */
 template<typename MatrixType> class DynBlock
   : public MatrixBase<typename MatrixType::Scalar, DynBlock<MatrixType> >
 {
@@ -84,6 +101,10 @@ template<typename MatrixType> class DynBlock
   *
   * Example:
   * \include MatrixBase_dynBlock.cpp
+  * Output:
+  * \verbinclude MatrixBase_dynBlock.out
+  *
+  * \sa class DynBlock
   */
 template<typename Scalar, typename Derived>
 DynBlock<Derived> MatrixBase<Scalar, Derived>

Eigen/Core/MatrixBase.h

@@ -85,7 +85,7 @@ template<typename Scalar, typename Derived> class MatrixBase
     
     /** This is the "real scalar" type; if the \a Scalar type is already real numbers
       * (e.g. int, float or double) then RealScalar is just the same as \a Scalar. If
-      * \Scalar is \a std::complex<T> then RealScalar is \a T. */
+      * \a Scalar is \a std::complex<T> then RealScalar is \a T. */
     typedef typename NumTraits<Scalar>::Real RealScalar;
     
     /** \returns the number of rows. \sa cols(), RowsAtCompileTime */