diff --git a/CMakeLists.txt b/CMakeLists.txt
new file mode 100644
index 0000000..735683a
--- /dev/null
+++ b/CMakeLists.txt
@@ -0,0 +1,19 @@
+cmake_minimum_required(VERSION 3.12)
+project(mapreduce)
+set(CMAKE_CXX_STANDARD 14)
+set(Boost_USE_STATIC_LIBS        ON) # only find static libs
+set(Boost_USE_MULTITHREADED      ON)
+set(Boost_USE_STATIC_RUNTIME    OFF)
+find_package(Boost 1.36 REQUIRED system filesystem iostreams)
+
+MESSAGE(STATUS "find boost ${Boost_INCLUDE_DIRS} ${Boost_LIBRARIES}")
+
+include_directories(${Boost_INCLUDE_DIRS})
+include_directories(include)
+include_directories(include/detail)
+include_directories(include/detail/intermediates)
+include_directories(include/detail/schedule_policy)
+
+add_subdirectory(examples/wordcount)
+add_subdirectory(examples/prime)
+add_subdirectory(examples/friends)
diff --git a/README.md b/README.md
index 2bfa1f5..9096689 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,8 @@
 MapReduce C++ Library
 =
+
+This is a fork of cdmh/mapreduce, this fork aims to help the portabillity of the original code, for any questions please contact the original developer.
+
 The MapReduce C++ Library implements a single-machine platform for programming using the the Google MapReduce idiom. Users specify a map function that processes a key/value pair to generate a set of intermediate key/value pairs, and a reduce function that merges all intermediate values associated with the same intermediate key. Many real world tasks are expressible in this model, as shown in the Google paper.
 
     map (k1,v1) --> list(k2,v2)
diff --git a/examples/friends/CMakeLists.txt b/examples/friends/CMakeLists.txt
new file mode 100644
index 0000000..3ffc202
--- /dev/null
+++ b/examples/friends/CMakeLists.txt
@@ -0,0 +1,3 @@
+project(friends)
+add_executable(friends friends.cpp)
+target_link_libraries(friends ${Boost_LIBRARIES} pthread)
diff --git a/examples/prime/CMakeLists.txt b/examples/prime/CMakeLists.txt
new file mode 100644
index 0000000..551712d
--- /dev/null
+++ b/examples/prime/CMakeLists.txt
@@ -0,0 +1,3 @@
+project(prime)
+add_executable(prime prime.cpp)
+target_link_libraries(prime ${Boost_LIBRARIES} pthread)
diff --git a/examples/wordcount/CMakeLists.txt b/examples/wordcount/CMakeLists.txt
new file mode 100644
index 0000000..c0a05a9
--- /dev/null
+++ b/examples/wordcount/CMakeLists.txt
@@ -0,0 +1,3 @@
+project(wordcount)
+add_executable(wordcount wordcount.cpp)
+target_link_libraries(wordcount ${Boost_LIBRARIES} pthread)
diff --git a/examples/wordcount/wordcount.cpp b/examples/wordcount/wordcount.cpp
index f0e3cee..3542f3e 100644
--- a/examples/wordcount/wordcount.cpp
+++ b/examples/wordcount/wordcount.cpp
@@ -20,6 +20,13 @@
 #include <crtdbg.h>
 #endif
 
+
+#if defined(__GNUC__) || defined(__GNUG__)
+#define STRNICMP    strncasecmp
+#elif defined(_MSC_VER)
+#define STRNICMP    strnicmp
+#endif
+
 #include "wordcount.h"
 #include <iostream>
 
@@ -52,20 +59,8 @@ bool std::less<std::pair<char const *, std::uintmax_t> >::operator()(
          std::pair<char const *, std::uintmax_t> const &first,
          std::pair<char const *, std::uintmax_t> const &second) const
 {
-    return (strnicmp(first.first, second.first, std::min(first.second, second.second)) < 0)
-         || ((first.second < second.second)  &&  (strnicmp(first.first, second.first, std::min(first.second, second.second)) <= 0));
-}
-
-template<>
-constexpr
-bool std::less<std::string>::operator()(
-         std::string const &first,
-         std::string const &second) const
-{
-    return
-        std::less<std::pair<char const *, std::uintmax_t>>()(
-            std::pair<char const *, std::uintmax_t>(first.c_str(), first.length()),
-            std::pair<char const *, std::uintmax_t>(second.c_str(), second.length()));
+    return (STRNICMP(first.first, second.first, std::min(first.second, second.second)) < 0)
+         || ((first.second < second.second)  &&  (STRNICMP(first.first, second.first, std::min(first.second, second.second)) <= 0));
 }
 
 namespace {
@@ -81,7 +76,7 @@ double const sum(T const &durations)
 
 void write_stats(mapreduce::results const &result)
 {
-    if (result.map_times.size() == 0  || result.reduce_times.size() == 0)
+    if (result.map_times.empty()  || result.reduce_times.empty())
         return;
 
     std::cout << std::endl << "\nMapReduce statistics:";
@@ -103,7 +98,7 @@ void write_stats(mapreduce::results const &result)
     std::cout << "\n    Reduce key processing errors            : " << result.counters.reduce_key_errors;
     std::cout << "\n    Number of Reduce Tasks run (in parallel): " << result.counters.actual_reduce_tasks;
     std::cout << "\n    Number of Result Files                  : " << result.counters.num_result_files;
-    if (result.reduce_times.size() > 0)
+    if (!result.reduce_times.empty())
     {
         std::cout << "\n    Fastest Reduce key processed in         : " << std::min_element(result.reduce_times.cbegin(), result.reduce_times.cend())->count() << "s";
         std::cout << "\n    Slowest Reduce key processed in         : " << std::max_element(result.reduce_times.cbegin(), result.reduce_times.cend())->count() << "s";
@@ -174,7 +169,7 @@ void run_wordcount(mapreduce::specification const &spec)
         job.run<mapreduce::schedule_policy::sequential<Job> >(result);
 #else
         std::cout << "\nRunning Parallel WordCount MapReduce...";
-        job.run<mapreduce::schedule_policy::cpu_parallel<Job> >(result);
+        job.template run<mapreduce::schedule_policy::cpu_parallel<Job> >(result);
 #endif
         std::cout << "\nMapReduce Finished.";
 
diff --git a/examples/wordcount/wordcount.h b/examples/wordcount/wordcount.h
index 739f047..4c37ce6 100644
--- a/examples/wordcount/wordcount.h
+++ b/examples/wordcount/wordcount.h
@@ -46,7 +46,7 @@ template<typename KeyType>
 struct reduce_task : public mapreduce::reduce_task<KeyType, unsigned>
 {
     template<typename Runtime, typename It>
-    void operator()(Runtime &runtime, key_type const &key, It it, It const ite) const
+    void operator()(Runtime &runtime, KeyType const &key, It it, It const ite) const
     {
         runtime.emit(key, std::accumulate(it, ite, 0));
     }
diff --git a/include/detail/datasource.hpp b/include/detail/datasource.hpp
index 042c2c3..7b4b43d 100644
--- a/include/detail/datasource.hpp
+++ b/include/detail/datasource.hpp
@@ -4,6 +4,7 @@
 #pragma once
 
 #include <boost/iostreams/device/mapped_file.hpp>
+#include <mutex>
 
 namespace mapreduce {
 
@@ -15,7 +16,7 @@ template<typename Key, typename Value>
 class file_handler : ::mapreduce::detail::noncopyable
 {
   public:
-    file_handler(mapreduce::specification const &spec);
+    explicit file_handler(mapreduce::specification const &spec);
 
     bool const get_data(Key const &key, Value &value)   const;
     bool const setup_key(Key &/*key*/)                  const { return false; }
@@ -176,7 +177,7 @@ template<
 class directory_iterator : mapreduce::detail::noncopyable
 {
   public:
-    directory_iterator(mapreduce::specification const &spec)
+    explicit directory_iterator(mapreduce::specification const &spec)
       : specification_(spec),
         file_handler_(spec)
     {
diff --git a/include/detail/intermediates/in_memory.hpp b/include/detail/intermediates/in_memory.hpp
index d222c51..f6f267c 100644
--- a/include/detail/intermediates/in_memory.hpp
+++ b/include/detail/intermediates/in_memory.hpp
@@ -93,7 +93,7 @@ class in_memory : detail::noncopyable
 
         const_result_iterator &operator=(const_result_iterator const &other);
 
-        void increment(void)
+        void increment()
         {
             ++current_.second;
             if (current_.second == iterators_[current_.first]->second.end())
@@ -114,7 +114,7 @@ class in_memory : detail::noncopyable
             return value_ == other.value_;
         }
 
-        const_result_iterator &begin(void)
+        const_result_iterator &begin()
         {
             for (size_t loop=0; loop<outer_->num_partitions_; ++loop)
                 iterators_[loop] = outer_->intermediates_[loop].cbegin();
@@ -122,7 +122,7 @@ class in_memory : detail::noncopyable
             return *this;
         }
 
-        const_result_iterator &end(void)
+        const_result_iterator &end()
         {
             current_.first = std::numeric_limits<decltype(current_.first)>::max();
             value_ = keyvalue_t();
@@ -130,12 +130,12 @@ class in_memory : detail::noncopyable
             return *this;
         }
 
-        keyvalue_t const &dereference(void) const
+        keyvalue_t const &dereference() const
         {
             return value_;
         }
 
-        void set_current(void)
+        void set_current()
         {
             for (current_.first=0;
                  current_.first<outer_->num_partitions_  &&  iterators_[current_.first] == outer_->intermediates_[current_.first].end();
@@ -190,12 +190,12 @@ class in_memory : detail::noncopyable
         intermediates_.resize(num_partitions_);
     }
 
-    const_result_iterator begin_results(void) const
+    const_result_iterator begin_results() const
     {
         return const_result_iterator(this).begin();
     }
 
-    const_result_iterator end_results(void) const
+    const_result_iterator end_results() const
     {
         return const_result_iterator(this).end();
     }
@@ -263,7 +263,6 @@ class in_memory : detail::noncopyable
     }
 
     // receive final result
-    template<typename StoreResult>
     bool const insert(typename reduce_task_type::key_type   const &key,
                       typename reduce_task_type::value_type const &value,
                       StoreResult &store_result)
@@ -272,7 +271,7 @@ class in_memory : detail::noncopyable
     }
 
     // receive intermediate result
-    bool const insert(typename key_type                     const &key,
+    bool const insert(key_type                     const &key,
                       typename reduce_task_type::value_type const &value)
     {
         size_t const  partition = (num_partitions_ == 1)? 0 : partitioner_(key, num_partitions_);
diff --git a/include/detail/intermediates/local_disk.hpp b/include/detail/intermediates/local_disk.hpp
index 641d556..642a624 100644
--- a/include/detail/intermediates/local_disk.hpp
+++ b/include/detail/intermediates/local_disk.hpp
@@ -4,6 +4,7 @@
 #pragma once
 
 #include <iomanip>      // setw
+#include "../job.hpp"
 #ifdef __GNUC__
 #include <iostream>     // ubuntu linux
 #include <fstream>      // ubuntu linux
@@ -65,7 +66,7 @@ struct file_merger
         using std::vector<std::string>::const_reference;
     };
 
-    void open_files(void)
+    void open_files()
     {
         // open each file and read the first record (line) from each
         while (files.size() > 0)
@@ -247,7 +248,7 @@ class local_disk : detail::noncopyable
             kvlist_.resize(outer_->num_partitions_);
         }
 
-        void increment(void)
+        void increment()
         {
             if (!kvlist_[index_].first->eof())
                 read_record(*kvlist_[index_].first, kvlist_[index_].second.first, kvlist_[index_].second.second);
@@ -262,7 +263,7 @@ class local_disk : detail::noncopyable
                &&  kvlist_[index_].second == other.kvlist_[index_].second);
         }
 
-        const_result_iterator &begin(void)
+        const_result_iterator &begin()
         {
             for (size_t loop=0; loop<outer_->num_partitions_; ++loop)
             {
@@ -287,19 +288,19 @@ class local_disk : detail::noncopyable
             return *this;
         }
 
-        const_result_iterator &end(void)
+        const_result_iterator &end()
         {
             index_ = 0;
             kvlist_.clear();
             return *this;
         }
 
-        keyvalue_t const &dereference(void) const
+        keyvalue_t const &dereference() const
         {
             return kvlist_[index_].second;
         }
 
-        void set_current(void)
+        void set_current()
         {
             index_ = 0;
             while (index_<outer_->num_partitions_  &&  kvlist_[index_].first->eof())
@@ -333,11 +334,9 @@ class local_disk : detail::noncopyable
   private:
     struct intermediate_file_info
     {
-        intermediate_file_info()
-        {
-        }
+        intermediate_file_info() = default;
 
-        intermediate_file_info(std::string fname)
+        explicit intermediate_file_info(std::string const& fname)
           : filename(fname)
         {
         }
@@ -361,7 +360,7 @@ class local_disk : detail::noncopyable
                 std::ofstream::open(filename.c_str(), std::ios_base::binary);
             }
 
-            void close(void)
+            void close()
             {
                 if (is_open())
                 {
@@ -405,7 +404,7 @@ class local_disk : detail::noncopyable
                 return true;
             }
 
-            bool const flush_cache(void)
+            bool const flush_cache()
             {
                 use_cache_ = false;
                 for (auto it  = records_.cbegin(); it != records_.cend(); ++it)
@@ -469,12 +468,12 @@ class local_disk : detail::noncopyable
         }
     }
 
-    const_result_iterator begin_results(void) const
+    const_result_iterator begin_results() const
     {
         return const_result_iterator(this).begin();
     }
 
-    const_result_iterator end_results(void) const
+    const_result_iterator end_results() const
     {
         return const_result_iterator(this).end();
     }
@@ -490,7 +489,7 @@ class local_disk : detail::noncopyable
     }
 
     // receive intermediate result
-    bool const insert(typename key_type                     const &key,
+    bool const insert(key_type                     const &key,
                       typename reduce_task_type::value_type const &value)
     {
         size_t const partition = partitioner_(key, num_partitions_);
@@ -646,7 +645,7 @@ class local_disk : detail::noncopyable
     }
 
   private:
-    void close_files(void)
+    void close_files()
     {
         for (auto it=intermediate_files_.cbegin(); it!=intermediate_files_.cend(); ++it)
             it->second->write_stream.close();
diff --git a/include/detail/job.hpp b/include/detail/job.hpp
index d88e120..db609b4 100644
--- a/include/detail/job.hpp
+++ b/include/detail/job.hpp
@@ -3,6 +3,8 @@
 
 #pragma once
 
+#include "datasource.hpp"
+
 namespace mapreduce {
 
 template<typename T> uintmax_t    const length(T const &str);
@@ -65,7 +67,7 @@ class job : detail::noncopyable
       public:
         typedef ReduceTask reduce_task_type;
 
-        map_task_runner(job &j)
+        explicit map_task_runner(job &j)
           : job_(j),
             intermediate_store_(job_.number_of_partitions())
         {
@@ -91,7 +93,7 @@ class job : detail::noncopyable
             return intermediate_store_.insert(key, value);
         }
 
-        intermediate_store_type &intermediate_store(void)
+        intermediate_store_type &intermediate_store()
         {
             return intermediate_store_;
         }
@@ -117,7 +119,7 @@ class job : detail::noncopyable
         {
         }
 
-        void reduce(void)
+        void reduce()
         {
             intermediate_store_.reduce(partition_, *this);
         }
@@ -151,12 +153,12 @@ class job : detail::noncopyable
      {
      }
 
-    const_result_iterator begin_results(void) const
+    const_result_iterator begin_results() const
     {
         return intermediate_store_.begin_results();
     }
 
-    const_result_iterator end_results(void) const
+    const_result_iterator end_results() const
     {
         return intermediate_store_.end_results();
     }
@@ -170,12 +172,12 @@ class job : detail::noncopyable
         return true;
     }
 
-    size_t const number_of_partitions(void) const
+    size_t const number_of_partitions() const
     {
         return specification_.reduce_tasks;
     }
 
-    size_t const number_of_map_tasks(void) const
+    size_t const number_of_map_tasks() const
     {
         return specification_.map_tasks;
     }
diff --git a/include/mapreduce.hpp b/include/mapreduce.hpp
index fa24996..f5f571f 100644
--- a/include/mapreduce.hpp
+++ b/include/mapreduce.hpp
@@ -53,7 +53,7 @@ class joined_thread_group : public std::vector<std::thread>
         join_all();
     }
 
-    void join_all(void)
+    void join_all()
     {
         for (auto &thread : *this)
         {
@@ -146,7 +146,7 @@ void run(mapreduce::specification &spec, mapreduce::results &result)
 {
     typename Job::datasource_type datasource(spec);
     Job job(datasource, spec);
-    job.run<mapreduce::schedule_policy::cpu_parallel<Job> >(result);
+    job.template run<mapreduce::schedule_policy::cpu_parallel<Job> >(result);
 }
 
 }   // namespace mapreduce