Hi Lewis, I think an integration between the two projects is more than welcome. In particular because it will give Giraph a bigger and stable user base to provide insightful feedback about improvements, new features, and bugs. As I said, I currently do not see big blockers for the integration on our side, so I can only say we welcome the effort and we are open to help the student.
Best of luck! Claudio On Sat, Mar 23, 2013 at 12:15 AM, Lewis John Mcgibbney < [email protected]> wrote: > Hi All, > > I have a huge apology to make on this one. > I thought that there was no interest in this proposal and therefore > dropped it for the time being. > I reach out specifically to Claudio and Eli respectively here and > apologise entirely for not getting back to you guys. > > So the questions (respectively) were the following > > 1. I do not see a direct connection between giraph and nutch, except > if you want to run ranking/PageRank on the indexed stuff. But in that case, > the integration is quite trivial and boils down to the inputformat. Could > you develop your ideas further please? > 2. Like an internship project? I might know some people. Or did you > have someone in mind? > > My answers are as follows > > 1. We anticipate the delegation of our LinkRank (PageRank > implementation) mechanism a graph library like Apache Giraph. This could > remove a bit of code from Nutch and would hopefully be more efficient. I am > well aware of your contributions to Nutch Claudio :0) I think that your > input here would be extremely helpful in helping us get this off of the > ground. You can read a bit more about the justification behind this here. > [0] > 2. The Google Summer of Code project runs every year and I have been > getting interested and involved in it these last few years. I was looking > for the following > > > - Someone who is a student as of this May > - Someone who is interested in working with Giraph for graph > processing (ideally an existing member of the Giraph community however this > is not a MUST). > - Someone who is interested in a Page Rank implementation within > Giraph which could be utilised within Nutch (ideally also familiar with > graph structures produced by a crawler such as Nutch... however again not a > MUST). > > Honestly if you have any potential student candidates in mind, please > reach out to them. > > Additionally any feedback on the above would be excellent. I appreciate > that this reply is well, well overdue but I think the project would be a > great one to get off the ground and could be the beginning of forming a > nice bridge between out communities. > > Thank you very much in advance. > > Lewis > > [0] http://www.infoq.com/articles/nioche-apache-nutch2 > -- > *Lewis* > -- Claudio Martella [email protected]
