Clarify types in README

README.Armadillo.md

@@ -1,4 +1,4 @@
-# Armadillo (blaze-lib) Matrix Market
+# Armadillo Matrix Market
 
 `fast_matrix_market` provides read/write methods for [Armadillo](https://arma.sourceforge.net/) sparse and dense matrices.
 
@@ -14,6 +14,8 @@ arma::Mat<double> A;
 arma::SpMat<double> A;
 ```
 
+Not restricted to `double` matrices. Any type supported by Armadillo that makes sense in Matrix Market is also supported, such as `int64_t`, `float`, `std::complex<double>`, etc.
+
 ### Reading
 ```c++
 std::ifstream f("input.mtx");
@@ -22,7 +24,7 @@ fast_matrix_market::read_matrix_market_arma(f, A);
 ```
 
 **Note:** Armadillo versions older than 10.5 [did not zero-initialize memory](https://arma.sourceforge.net/docs.html#changelog).
-On those versions loading a sparse Matrix Market file into a dense `Mat` is not supported.
+On those versions reading any Matrix Market file into a dense `arma::Mat` is not supported.
 
 
 ### Writing

README.Blaze.md

@@ -17,6 +17,9 @@ blaze::CompressedVector<double> A;
 // or
 blaze::DynamicVector<double> A;
 ```
+
+Not restricted to `double` matrices. Any type supported by Blaze that makes sense in Matrix Market is also supported, such as `int64_t`, `float`, `std::complex<double>`, etc.
+
 ### Reading
 ```c++
 std::ifstream f("input.mtx");

README.CXSparse.md

@@ -15,9 +15,9 @@ cs_dl *A;
 std::ifstream f("input.mtx");
 fast_matrix_market::read_matrix_market_cxsparse(f, &A, cs_dl_spalloc);
 ```
-Note the last argument. It is the `cs_spalloc` routine that matches the type
+Note the last argument. It is the `cs_*_spalloc` routine that matches the type
 of `A`. It must be specified explicitly because it is impractical to autodetect due to the way CXSparse
-implements multiple index and value types.
+implements multiple index and value types. All CXSparse types are supported, such as `cs_dl`, `cs_ci`, `cs_cl`, etc.
 
 `read_matrix_market_cxsparse` creates a triplet version of the matrix structure. For CSC, follow up with
 CXSparse's `cs_compress()`:

README.Eigen.md

@@ -36,4 +36,6 @@ fast_matrix_market::write_matrix_market_eigen(f, mat);
 
 Again, for dense matrices/vectors use `write_matrix_market_eigen_dense`.
 
+Not restricted to `double` matrices. Any type supported by Eigen that makes sense in Matrix Market is also supported, such as `int64_t`, `float`, `std::complex<double>`, etc.
+
 Note that Vectors are written with object type = `matrix` for compatibility reasons. Eigen's `saveMarketVector()` does the same. As of writing Eigen's loader crashes on `vector` files.

README.md

@@ -88,7 +88,7 @@ Matrix composed of row and column index vectors and a value vector. Any vector c
 struct triplet_matrix {
     int64_t nrows = 0, ncols = 0;
     std::vector<int64_t> rows, cols;
-    std::vector<double> vals;
+    std::vector<double> vals;       // or int64_t, float, std::complex<double>, etc.
 } mat;
 
 fast_matrix_market::read_matrix_market_triplet(
@@ -108,7 +108,7 @@ Be mindful of whether your code expects row or column major ordering.
 ```c++
 struct array_matrix {
     int64_t nrows = 0, ncols = 0;
-    std::vector<double> vals;
+    std::vector<double> vals;       // or int64_t, float, std::complex<double>, etc.
 } mat;
 
 fast_matrix_market::read_matrix_market_array(