[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [recommenders-dev] Question regarding code searching: GSOC application

Hi,

just a quick addition: We already have two submissions for code search this year - mostly dealing with structural issues and code summarization techniques. They have a clear scope and we know what to expect. In the case of recommending similar methods or code blocks I have the feeling that a GSCO would be very open-ended and looks more like a research (Bsc/Msc) thesis to me than a GSOC.

What I can imagine to work well as GSOC would be a "definition recommender", or improving on Subwords Completion Engine (I don't have the bug number at hand, but it's in Bugzilla and reported by Wim Jongman). A definition recommender or frequent call chains recommenders could be implemented on top of the current code search index - or just based on your own AST indexer that creates recommendations models from your local workspace.

Anyways, you have to hurry up to bring your proposal in :-) (and I'm travelling today and tomorrow with almost no internet)

Best,
Marcel


On May 2, 2013, at 9:48 AM, Andreas Sewe <andreas.sewe@xxxxxxxxxxxxxx> wrote:

> Hi,
> 
>> I am a graduate student in the department of Computer Science at
>> University of Saskatchewan.
> 
> funny, I used to study there myself for one year. Is Mike Horsch still
> the machine learning guy in town?
> 
>> I am particularly interested to the Code
>> Searching problem. Is this proposal also considering searching for
>> similar methods (like method level code completion) or blocks of code?.
> 
> There is a whole MSc thesis [1] and prototype implementation [2] for you
> to look at.
> 
> But the answer is yes, definitely; you want to be able to find other
> code in your current workspace that is similar to the one you are
> currently working on (uses the same types, calls the same methods, etc.)
> 
>> Can I consider that in my application. Could you please also point me
>> anything that can help me to understand the idea? (such as the Masters
>> thesis talked in the proposal). I knew about the GSOC at the last moment
>> from my friend and I found this a great place to blend research idea and
>> implementation. Any suggestion/comment would really help me to complete
>> my application.
> 
> Have a look at the thesis certainly helps. Also, the Bugzilla issue I
> linked to contains much of the discussion on this topic so far. (For the
> rest, search the mailing list archives.)
> 
> Hope that helps.
> 
> Andreas
> 
> [1] <https://bugs.eclipse.org/bugs/show_bug.cgi?id=401843>
> [2]
> <https://git.eclipse.org/c/recommenders.incubator/org.eclipse.recommenders.codesearch.git/>
> 
> -- 
> Codetrails UG (haftungsbeschränkt)
> The knowledge transfer company
> 
> Robert-Bosch-Str. 7, 64293 Darmstadt
> Mobile: +49-170-811-3791
> http://www.codetrails.com/
> 
> Managing Director: Dr. Marcel Bruch
> Handelsregister: Darmstadt HRB 91940
> _______________________________________________
> recommenders-dev mailing list
> recommenders-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/recommenders-dev

-- 
Codetrails UG (haftungsbeschränkt)
The knowledge transfer company

Robert-Bosch-Str. 7, 64293 Darmstadt
Mobile: 0179 131 77 21
http://www.codetrails.com

Managing Director: Dr. Marcel Bruch
Handelsregister: Darmstadt HRB 91940