Hi Ralf, and all
My name is Wensheng Dou, and I am a first-year Ph.D. student from
Institute of Software, Chinese Academy of Sciences in China. I am
interested in parallel and concurrent programming including programming
models, tool support, especially, refactoring for concurrency in Java.
And I want to get some practical experience about concurrency-aware
refactoring.
So I want to get involved in the GSoC2010 and contribute to open source
project, and I am interested in “Refactor lock engine” project. I think
this project could make me gain practical experience, and make Castor
more scalable, readable and maintainable, and gain good performance.
I have downloaded the Castor code (version 1.3.1, because the main
version does not work now, because of a maven build error), and I find
that Castor doesn’t use the java.util.concurrent library to improve the
performance (it uses some code from EDU.oswego.cs.dl.util.concurrent
library). The java.util.concurrent library provides flexible locking
constructs, more concurrent mechanism, and many lock-free and
thread-safe atomic data type. The java.util.concurrent library is very
useful to construct concurrent programs. I think Castor will benefit
from java.util.concurrent library.
I've read some documentations and some of the code of Castor and as per
my understanding, the outcomes of the project consists of the followings:
1. Remove the EDU.oswego.cs.dl.util.concurrent library, and use the
java.util.concurrent library to improve concurrency.
2. Find the shared data from the whole Castor project, and use atomic
data type or correct locks to protect them, and make the program
behavior-preserving, scalable.
3. Find the concurrency bugs in Castor JIRA, and repair these bugs.
After this project is done, Castor will use the java.util.concurrent
library to provide good performance, and I’ll gain more practice, which
will help me do my research.
I’ll investigate this project more deeply, and provide a more detailed
proposal for this project later.
Any comments and suggestions are welcome.
Thanks in advance for your feedback.
Best Regards,
Wensheng Dou
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
