Idea for Tutorial: Centralities

[Russell Jurney]

I really wish there was more intermediate documentation for Gremlin. Marko suggested I share some ideas for tutorials, so here goes:

* Centralities: starting with eigenvector centrality (which is a one-liner), proceed to also calculate closeness and betweenness centralities, as well as pagerank. If it helps illustrate principles of crafting graph walks, add other types of centrality. This will give users practice in implementing algorithms.

What is needed is help learning to code in graph traversals, so using any example that can teach people the breadth of operators available and how to use them is going to help.

--

Russell Jurney twitter.com/rjurney russell...@gmail.com relato.io

[Marko Rodriguez]

Hi Russell,

Over the years, we have really failed on the tutorial front. Fortunately, with TinkerPop 3.1.0, Stephen released a much needed "Getting Started" tutorial.

http://tinkerpop.apache.org/docs/3.1.0-incubating/tutorials-getting-started.html

Moreover, Stephen and Daniel Kuppitz have positioned us nicely for others to easily write and publish tutorials. With that, I think I will do the following tutorials:

https://issues.apache.org/jira/browse/TINKERPOP3-1007

https://issues.apache.org/jira/browse/TINKERPOP3-1008

Russell (others), please feel free to comment on these with specific desires and directions…

Thanks,

Marko.

http://markorodriguez.com

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/CANSvDjppfp6z%2BANz6OxEGYoQ14UT66k-He%2BTN3iWyE_aTZ1n5w%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[rjurney]

What is needed is an O'Reilly book titled 'Learning Gremlin.' That would totally address the documentation problem. 

The Getting Started tutorial is very helpful, but only gets you so far (I've worked through it). What is needed are... lets say you pick 10 or 20 complex queries that do a good job of demonstrating what is possible with gremlin. Common problems people need solved. And then walk the user stepwise through how to compose and compute those metrics step-by-step, learning a new technique each step, until the complex query to achieve the task is completed. This would constitute the book I mention.

This would dramatically increase the adoption of Tinkerpop. Right now it feels like less than ten people can actually use Gremlin to its full potential, and you dole out little bits of insight into its workings on the mailing list. The rest of us are not able to achieve general fluency in Gremlin, even after a long time trying. I wish Datastax would commission Learning Gremlin (or Learning Tinkerpop) with O'Reilly.