The Dask community is broad and diverse, mixing a volunteer open source community, academics and government workers, and for-profit corporations. These sub-communities understandably have different needs and expectations around communication, which can sometimes conflict with each other.
For official Dask communications we try to strike a balance that is exciting while genuine, and informative while light-hearted. In particular …
We strive to promote positive news about Dask, either its technical features:
Did you know that Dask can integrate well with Scikit-Learn? Read more here!
or interesting and important ways in which the project has been used
We’re excited to see Dask used in the fight against Malaria. Read more in this article.
We avoid overt self-praise, especially if it is not very well proven.
Dask is 100x faster than Spark!
We temper positive results with ways in which we can improve, and highlight opportunities for others to engage.
We’re excited to announce Dask integration with Cassandra!
But there is still plenty to do.
We don’t linger long on negative results, but we do make sure to acknowledge them in order to maintain an honest reputation.
While running this benchmark we actually found that Dask performed worse for one part of it. That lead us to fix a bug (see PR #1234) and now things are running smoothly again.
We endeavor to highlight the work of others, both among our own contributors, and among the project that are adjacent to Dask
Next, we apply the new dask.dataframe.foo operation (recently contributed by Alice Chen
For non-official Dask communications, either by individuals or associated corporations we can make no expectations of tone. So while speaking about Dask with your own voice, or the voice of your company, please do not feel burdened by the expectations of this document. We trust that you know best how to speak with your audience.