Re: [foaf-dev] Tinkerpop

On Tue, May 10, 2011 at 4:26 PM, Paul Houle <paul-y/[email protected]> wrote:

>
>     Not all algorithms on graphs will parallelize,  but I'd be much
> more impressed with some kind of system that can do parallel traversals
> that are spread out on a set of database nodes.  Without something like
> that,  Tinkerpop is syntactic sugar for algorithms that aren't that hard
> to code in straight Java,  PHP or whatever you do.
>

well.. for mortals like me, it's always a bless not having to spend time
reimplementing and optimizing certain traversals.

but couldn't agree more.  you are always going to hit the ceiling, no matter
how much iron you put on it.

the paper about pregel is really exciting:
http://googleresearch.blogspot.com/2009/06/large-scale-graph-computing-at-google.html

if your first reaction is a show me da code, bad luck.
instead of blaming the monster, you can play a little with phoebus.

http://nosql.mypopescu.com/post/1292881817/phoebus-erlang-based-implementation-of-googles-pregel

for sure there must be some black magic before getting a sensible
dispatcher, but at least big part of the data from real world will be
already clustered to some extent.

cheers

_______________________________________________
foaf-dev mailing list
foaf-dev-RyYwo1q5J+qsOXdr9/[email protected]
http://lists.foaf-project.org/mailman/listinfo/foaf-dev