search

darobin / gr

Personal playground for graph-related things
MIT License
0 stars 0 forks source link

Dataflow rather than dataviz? #1

Closed eric-brechemier closed 9 years ago

eric-brechemier commented 9 years ago

I first read "graph" as "chart" in your project description, and expected to find drawing routines in the code, due to my current bias with data visualization.

Are you exploring graphs in the sense of dataflows?

In this case, I bring to your attention this book: A Practical Handbook on Developing Dataflow Programming Systems; I am only halfway through but I can already recommend it as a good introduction to dataflow programming, packed with experience, written in an accessible style.

You may also be interested in a discussion I had on this topic with @dthevenin a while back and several forms of API that we considered at the time.

darobin commented 9 years ago

No, this is not at all about charting. Otherwise I'd have called it that :)

But it is not about dataflow either. I am playing with graph-theoretic notions, notably complex graphs and random graphs, network motifs, etc. The books are more http://www.amazon.fr/Complex-Networks-Algorithmic-Kayhan-Erciyes-ebook/dp/B00NFODK7S/ref=sr_1_2?ie=UTF8&qid=1418727770&sr=8-2&keywords=complex+networks and http://www.amazon.fr/Structure-Dynamics-Networks-Mark-Newman-ebook/dp/B00BMTSCHQ/ref=sr_1_2?ie=UTF8&qid=1418727818&sr=8-2&keywords=networks+structure+dynamics.

Thanks for the link though, it is also interesting even if unrelated :) I might read that book later, I've certainly been curious about NoFlo.

I am interested about computation in graphs, but unlike DataFlow systems I don't care about designing it but rather about finding the conditions under which it emerges. The core idea (of the first step) is to see in which classes of random boolean networks universal computation can be shown to necessarily emerge.

eric-brechemier commented 9 years ago

All right, thanks for the clarification :)