I feel we should plan out some concrete things to be implemented under Angrapa so that other people/communities start using it.
For example: I am in Mahout Dev community and we have had discussions of using Hama but we always ended in downplaying the idea because Hama does not have a concrete library and road map. Tasks like http://issues.apache.org/jira/browse/HAMA-163 can be tackled first to have a basic library up soon. Can we also discuss interface for Input/Output formatter? I Remain, Varun Gupta On Wed, Sep 30, 2009 at 12:26 PM, Edward J. Yoon <[email protected]>wrote: > Hello, > > > I can take up work on Angrapa and therefore would like know if you guys > have > > explored existing Graph libraries to have a threshold to build on. > > Not yet. BTW, Since it's the graph computing framework, which is based > on graph traversal as a M/R program based on sequential scan, It would > be nice to have a basic library for graph data structures such as, > Input/Output formatter, Hbase graph storage, .., etc. (At this stage, > some library for graph algorithms is not our plan, except some > examples) > > > If we are up to creating our own base classes (basic graph class and > likes), > > I will start by creating JIRA tasks and and then submitting code. > > Great. Please feel free to contribute. > > Plus, In google code project, there is our prototype implementation -- > http://code.google.com/p/throb/source/checkout > > On Wed, Sep 30, 2009 at 3:13 PM, Varun Gupta <[email protected]> > wrote: > > I can take up work on Angrapa and therefore would like know if you guys > have > > explored existing Graph libraries to have a threshold to build on. > > > > If we are up to creating our own base classes (basic graph class and > likes), > > I will start by creating JIRA tasks and and then submitting code. > > > > I Remain, > > Varun Gupta > > > > > > -- > Best Regards, Edward J. Yoon @ NHN, corp. > [email protected] > http://blog.udanax.org >
