Reproducibility Practices Guide

[iamciera]

I really love the Issue 22 thread and this thread mentioned by @cboettig. I propose creating a guide that incorporates much of what has been discussed + other knowledge mentioned elsewhere. Not to mention all the expertise of everyone involved with the hackathon event will be in one place.

Include in guide: tool options, workflow options (with pretty diagrams), problems to consider, data sharing options, best readme file practices, metadata to include, ect ect.

I think having this information in one place could benefit the scientific community immensely, especially scientists new to coding practices. Not sure where this would be housed on the site, but making it contribution friendly is a plus. Possibly part of the R OpenSci Cookbook project (Issue 11)?

Just an idea. Admittedly, my R skills are pale in comparison to many of you guys, but this is a project I could undertake.

[karthik]

Fantastic idea, @iamciera 💯
Totally on board with this. It could be a repo with examples built in that others could clone and immediately use in their efforts. We can, of course, make it a community effort that would allow anyone to send a pull request with new ideas and features.

[karthik]

Another quick note, I think this would more general than (#11) so let's start a new one for this.
@iamciera Would you mind editing the projects page on the wiki and adding it there? We should start moving things over from the issues.

[jhollist]

Love this too, @iamciera and @karthik!

Not sure how much I'll be able to contribute given my remote attendance (and modest development skills) but I would be willing to help out with this effort however I can.

[karthik]

Your just being modest @jhollist
We'll try to make remote collaboration work as smooth as we can. I'll invite everyone to our chatroom this weekend. You'll also be able to interact over GitHub issues and pull requests. If need be, we can also Skype you in.

[szeitlin]

Happy to work on this with you. Might also help us to identify where
data analysis/visualization/annotation functionality is missing (an area
I'd like to propose as a topic, kind of surprised not to see any threads
about it?).

sam

iamciera wrote:

I really love the Issue 22 thread
#22 and this thread
mozillascience/code-research-object#2
mentioned by @cboettig https://github.com/cboettig. I propose
creating a guide that incorporates much of what has been discussed +
other knowledge mentioned elsewhere. Not to mention all the expertise
of everyone involved with the hackathon event will be in one place.

Include in guide: tool options, workflow options (with pretty
diagrams), problems to consider, data sharing options, best readme
file practices, metadata to include, ect ect.

I think having this information in one place could benefit the
scientific community immensely, especially scientists new to coding
practices. Not sure where this would be housed on the site, but making
it contribution friendly is a plus. Possibly part of the R OpenSci
Cookbook project (Issue 11)?
#11

Just an idea. Admittedly, my R skills are pale in comparison to many
of you guys, but this is a project I could undertake.

—
Reply to this email directly or view it on GitHub
#25.

[iamciera]

Thanks for help offers @szeitlin and @jhollist!

Sure I will send this to the projects wiki. Just to clarify, do we all think we want it to be part of the rOpenSci Cookbook? Like a chapter within the book?

Also, @karthik, I'm a little confused by your suggestion to make it a repo with examples. Make the reproducibility guide development project a repo and examples would be topics people can choose to write about/elaborate on? If that is what you are thinking, I think that would work. We can compile everything at the end to make it a chapter in the book...if that is what we want.

What do you guys think?

[eduardszoecs]

👍
I already outlined my workflow in the other issue and would be happy to (remotely) join in / contribute to this issue!

[karthik]

Just to clarify, do we all think we want it to be part of the rOpenSci Cookbook? Like a chapter within the book?

@iamciera If I understand correctly, the book (not quite a book in this case really) is meant to be a compendium of current documentation (more of a how to guide) for all the rOpenSci packages. As it stands right now, all packages have their own versions of a README.md, some generated automatically and some not. Some have vignettes while others don't. So when packages change, there is no consistent way to automatically update the docs across the board. Hence that idea.

There is a separate rOpenSci book in the works (although it is not specific to rOpenSci, just open science with R). See public repo

Based on what you described above, this could be a separate repo to serve as a community source. Much like markdown for science, which started here and is now forked over here.

Make the reproducibility guide development project a repo and examples would be topics people can choose to write about/elaborate on?

or rather, it can encompass everything you listed earlier Include in guide: tool options, workflow options (with pretty diagrams), problems to consider, data sharing options, best readme file practices, metadata to include, ect ect. but also have full examples. Like an example of a good README. Then people can clone that repo, copy that example to a different project and use it. Does that make sense?

So what I'd suggest is that I'll create a new repo under rOpenSci and assign you all to a team. We can push to that repo and allow the larger community to contribute to it.

@Edild: Can you send me an email? I need to get some info over to you and Jeff.

[iamciera]

@karthik Super clear now.

I think the template example files are an excellent idea! I will start compiling what we should include along with references, then at the Hackathon we can prioritize and attack.

[sckott]

great idea @iamciera !

[jhollist]

@iamciera Sounds great.

Just so you can plan, I am joining in remotely and on the east coast. Add that to my commitments on Monday and I will only be available for 1-2 hours on Monday AM. Tuesday is open for me and I am going to try and keep it that way. I may have some time later on Monday to check back in.

Again thanks for the taking the reigns on this. I like the idea!

[karthik]

@iamciera ok, a repo is ready to go. Everyone who expressed interest has access to it (although you might want to work with pull requests to avoid a melt down ;)

Repo: https://github.com/ropensci/reproducibility-guide

[AnyaB]

I'd like to join on this team late Monday & Tuesday mornings.