On Tuesday, 16 February 2016 at 15:03:36 UTC, Jakob Jenkov wrote:
I cannot speak on behalf of the D community. In my opinion I
don't think that it is D that needs a big data strategy. It is
the users of D that need that strategy.
I am originally a Java developer. Java devs. create all kinds
of crazy tools all the time. Lots fail, but some survive and
grow big, like Spark.
D devs need to do the same. Just jump into it. Have it be your
hobby project in D. Then see where it takes you.
Good attitude. Nevertheless, I think there is a much larger
population of people who would want to use D for normal data
analysis if packages could replicate much of what people do in
R/Python.
If the OP really wants to contribute to big data projects in D,
he might want to start with things that will more easily allow D
to interact with existing libraries.
For instance, Google's MR4C allows C code to be run in a Hadoop
instance. Maybe adding support for D might be do-able?
http://google-opensource.blogspot.com/2015/02/mapreduce-for-c-run-native-code-in.html
There is likely value in writing bindings to machine learning
libraries. I did a quick search of machine learning libraries and
much of it looked like it was in C++. I don't have much expertise
with writing bindings to C++ libraries.
