NBToPackage

Software goals, questions and answers

1. What do we care about in terms of reproducibility? How would we remake the identical plots in two years if needed?

To be honest, reproducibility is kinda aspirational, but yeah, it would be cool to have it be really easy to remake exactly the same plots.

2. Related, but slightly different question, what do we care about in terms of “provenance tracking”, i.e., knowing which data and software version were used to make a particular plot?

Being able to know exactly what went input a plot in terms of both data and software would be super useful. Worth putting a bit of effort into.

3. What do we care about in terms of configurability in the plots? Are we going to be trying out a bunch of different versions of the plots? What are we going to be changing?

Probably we are going to be changing the apperence a lot while we write the paper. Probably we are going to be changing the data and possibly cuts a lot while we do the analysis. In any case, let’s keep our options open.

4. How much do we care about consistency in things like axis ranges, axis labels, data selection criteria, etc…?

Personally I care a lot, it takes forever getting all this stuff right when you are submitting a paper. Making it easier would be great.

5. Plots that are formatted for papers tend to not look great in talks, do we want a simple switch to make the same plots, but formatted for talks?

That would be pretty cool.

This site is open source. Improve this page.