Re: [IMPORTANT] more ideas by today

[Per Bothner]

Attached are some addition to Kawa.  You can write:

Mentors: Per Bothner and Jamison Hope

Then after the existing:

   Kawa maintains a list of ideas here: 
http://www.gnu.org/software/kawa/Ideas-and-tasks.html.

Add:
  Thee following seem most suitable for GSoC:
and the text from the attachment.
--
  
	--Per Bothner
address@hidden   http://per.bothner.com/

• Run interactive process in separate Java Virtual Machine: When developing and testing it is useful for the REPL to support hot-swapping (replacing functions on-the-fly) and debugging. The main goal being able to smoothly reload changed modules (files or functions), and have other modules not break. Debugging (such as setting breakpoints) would not be a priority for this project, but could be a follow-on project. Skills: Should be experienced with Java, and interested in learning about JVM TI and similar low-level parts of the platform. Difficulty: Challenging, but you can study how Java-9's new jshell uses the JVM TI.
• Typed/optimized arithmetic with units: Kawa support units (such as cm^2 for square centimeters) and quantities (such as 4cm^2). The idea is define a syntax for type for a quantity with a specific unit, for example quantity[cm^2]. The compiler would type-check use of such types, and use them to generate primitive arithmetic. Skills: Need good Java experience, and somewhat familiar with the Java Virtual Machine. You will need to become comfortable reading javap output. Difficulty: Modest.
• Easier access to native code using JNI: The traditional way to access native (C/C++) functions is using JNI, but it's very awkward. JNA and JNR are much easier to use. This project would design and implement an easy-to-use Kawa wrapper for for JNR. You should study existing JNR wrappers, such as that for JRuby. Difficulty: Medium. Need to study existing wrappers and "foreign function interfaces" (in multiple languages) and design one suitable for Kawa. Some Scheme (Kawa) experience would be helpful.
• Improve IDE support: There is some IDE support for Kawa, but it is old, incomplete, and/or poorly documented. One possible project would focus on research and documentation: evaluating options (using NetBeans, Eclipse, IntelliJ, or an Emacs package); seeing what if anything works (or can be made to work with modest effort); and documenting it. Alternatively, if you have a favorite IDE/Editor and want to learn how to extend it, you can focus on that. Skills: Solid experience with Java and at least one IDE for the Java platform. Difficulty: Modest. This can be very variable.
• Update XQuery implementation: Kawa includes a mostly complete implementation of the 1.0 version of XQuery, a functional language for working with XML and similar data. However, XQuery has added major feature since then, including grouping and updates. Adding some of these features to the Kawa-XQuery implementation would be desirable. Skills: Comfortable with both Java and XML technologies. Interested in learning more about the latter. Difficulty: Depends on what features you want to tackle.
• Use MethodHandles and invokedynamic Java-7's invokedynamic and MethodHandles were intended to improve the performance of dynamic languages. See JSR 292 and the Da Vinci Machine Project. Kawa makes very limited use of these, focusing on compile-time optimizations, but could do more. For example we can start by optimizing arithmetic when the types are unknown at compile-time. They could make implementing generic functions (multimethods) more efficient. Skills: Very solid Java knowledge, and willing to learn and use complex APIs. Difficulty: High.