Add difficulty rating/mentors list to dynd ideas list

Author: mwiebe

2016/ideas-list-dynd.md

@@ -18,6 +18,10 @@
 
 Complete DyND's string type implementation and develop the supporting string library. Explore how string processing and NumPy-style array programming interact.
 
+| **Intensity** | **Involves**  | **Mentors**        |
+| ------------- | --------------|------------------- |
+| Moderate | | Irwin Zaid, Mark Wiebe, Ian Henriksen |
+
 ### Technical Details
 
 DyND's string type and supporting library are incomplete. A project could involve completing the implementation of the basic string type, fleshing out the supporting string function API, connecting DyND strings to the Numba JIT, and exploring how string API choices interact with the array programming ideas and typical data ingest and analysis workflows.
@@ -44,6 +48,10 @@ C\++ development experience, particularly with C\++11 or C\++14, is necessary. F
 
 Implement binary serialization for DyND array data, and corresponding data hashing that's useful for building Merkle DAG-based systems.
 
+| **Intensity** | **Involves**  | **Mentors**        |
+| ------------- | --------------|------------------- |
+| Moderate | | Irwin Zaid, Mark Wiebe, Ian Henriksen |
+
 ### Technical Details
 
 DyND's nd::array object can store many forms of data in memory. Being able to generically serialize these objects into blobs of bytes for saving on disk or transporting across a network is needed for many use cases. DyND's design aims to use an obvious raw-style format for the serialized data, to be usable as a building block for more sophisticated systems.
@@ -74,6 +82,10 @@ A project tackling this functionality could begin with implementing the serializ
 
 Implement an associative array type for DyND as a hash map.
 
+| **Intensity** | **Involves**  | **Mentors**        |
+| ------------- | --------------|------------------- |
+| Hard     | | Irwin Zaid, Mark Wiebe, Ian Henriksen |
+
 ### Technical Details
 
 This idea overlaps slightly with [Serialization and Data Hashing](#serialization), but focuses on the implementation of a hash map type instead of the hashing its built on. The idea would be for the type to be named like `map[S, T]`, and be a hash table with keys of type `S` and values of type `T`.
@@ -102,6 +114,10 @@ C\++ development experience, particularly with C\++11 or C\++14, is necessary. A
 
 Improve DyND's support for missing data via its `option[T]` type, targetting DyND's ability to be used by projects like [pandas](http://pandas.pydata.org/).
 
+| **Intensity** | **Involves**  | **Mentors**        |
+| ------------- | --------------|------------------- |
+| Hard     | | Irwin Zaid, Mark Wiebe, Ian Henriksen |
+
 ### Technical Details
 
 Having a representation for missing data and ways to meaningfully manipulate them are crucial for data analytics applications. DyND has preliminary support for this, but it is far from complete.
@@ -134,6 +150,10 @@ C\++ development experience, particularly with C\++11 or C\++14, is necessary. E
 
 Integrate robust conversions of float point values to/from strings.
 
+| **Intensity** | **Involves**  | **Mentors**        |
+| ------------- | --------------|------------------- |
+| Hard     | | Irwin Zaid, Mark Wiebe, Ian Henriksen |
+
 ### Technical Details
 
 As part of being a library to facilitate interoperability of data, DyND needs to include good quality routines for conversion of floating point values to and from strings. This basic functionality is used by I/O such as for JSON and CSV, and in displaying output for the user or interpreting user input.