Category structure

Site Feedback
1

Right now the categories are based around Dask technologies, but I wonder if this is developer-focused. This is how we break down the project, but users might feel differently.

For example, I wonder if the following categories might be more aligned with how they’re coming into the site:

  • Help!
  • Showcase
  • Welcome
  • Events
  • Jobs
  • ML
  • Data engineering

Things like dataframes and distributed seem more like tags to me. I also expect that we’re going to want to label discussions with both dataframes and distributed at some point.

2

I basically agree with this. I was trying to avoid very developer-centric categories like breaking it up by GitHub repository, but it’s still fairly tech-focused.

FWIW, the Jupyter Discourse roughly categorizes things by tech, and the Julia and Pangeo versions are at a bit of a higher level.

3

  1. “Help!” is too broad to be a single category. TBH I had assumed the help part was implied here:

    • Dask Array = help with Dask arrays
    • Dask Dataframe = help with Dask dataframes
    • … etc.

  2. Should we add an “Internals and Design” category for developer discussion? The JuliaLang discourse has this, which seems useful.

  3. We might also want a tag (a tag, not a category) for “performance”. There are lots of user questions in the issue tracker that go: I’m using Dask arrays/dataframes/delayed to do $foo, but the performance isn’t good, how can I improve it?
    Not sure if question askers will know this tag exists and think to add it (can other people add tags to questions after the fact?)

In general, I think keeping the number of categories small-ish is a good plan. I wouldn’t want us to add many more.

4

Also, I saw “ML” in Matt’s bullet point list, which reminds me that people might want to ask questions about dask-ml (or other dask projects). How do we want to handle that?