Update Software references page

[naushinthomson]

• [x] Update page to include guidance about software references when authors don't have information about authors of the software they're citing (make language more flexible)
• [x] Becky approved page
• [x] Melissa approved page
• [x] Fred approved page
• [x] James approved page
• [x] Naushin approved page
• [x] Claudia approved page
• [x] Share page with Exeter and invite feedback
• [x] Revise with Exeter's suggestions

[naushinthomson]

@JGilbert-eLife @FAtherden-eLife @Melissa37 @bcollins14 @griffithsc This page is now ready to review again! Please let me know if this all makes sense. I've also tried to make the author query more flexible so that if authors don't know the information about the software they don't panic but let me know if that can be improved!

[JGilbert-eLife]

Didn't pick up on this before - I think "These need to be added as a full DOI links instead of just the DOI." should be "These need to be added as a full URL links (e.g. http://doi.org/10.6084/m9.figshare.4742866.v1) instead of just the DOI."

[JGilbert-eLife]

Also I think we could just say "These need to be added as software references in accordance with the FAIR principles."

[JGilbert-eLife]

It's "figshare" rather than "Figshare". I think, anyway.

[JGilbert-eLife]

I'm not sure if reusing "eLife's policy is to include full software reference details in accordance . . ." as the query for missing details in e.g. a GitHub reference the best idea. We'll know what's missing, surely? Can't we just as for the version number or whatever else is require?

[JGilbert-eLife]

I think the instructions for "software-replacement-character-presence" need to be clearer - isn't the task to identify what character should be used in place of the broken one?

[JGilbert-eLife]

software-doi-test-2's action doesn't seem to match the guidance provided for software-doi-test-1 - aren't these the same task just on different platforms?

[JGilbert-eLife]

Again, for ref-software-test-5 and ref-software-test-6, I'm not sure we need to bother with the full evocation of FAIR and could just ask for the missing information like we would with any other reference.

[JGilbert-eLife]

What's the table at the end of the XML structure checks for?

[JGilbert-eLife]

Also, were we cutting out Curator entirely? There are a couple of tests that still refer to it.

[JGilbert-eLife]

Otherwise looks great - thanks @naushinthomson !

[Melissa37]

Software version

It is not required, but can we indicate that if it's provided it should not be deleted OR merged into the title but tagged properly?

[Melissa37]

There are interesting discussions on the JATS4R working group on software citations that may well influence us. Too early to implement, but I think this will get updated at some point reflecting those!

[Melissa37]

Website

I am surprised this is not mandatory?

[Melissa37]

If the authors cannot provide this information, the text should be left as is.

I don't understand this as the screenshot and text directly below contradicts it - All software here should be added as software references and cited properly

[Melissa37]

Author - for GitHub references, you can just enter the first author's name. Sometimes the repository will include information about the authors - in this case, they should all be added.

Do we mean you can use the author's GitHUb User ID if their name is not anywhere?

[Melissa37]

Year - the year of the latest commit when the repository is forked.

This is not relevant anymore as we don't fork?

[Melissa37]

R has an RRID, should we be consistent about packages like these? CRAN and biomod2 do too :-)

https://scicrunch.org/resources/Any/record/nlx_144509-1/SCR_001905/resolver?q=*&l=#:~:text=R%20Project%20for%20Statistical%20Computing%20(RRID%3ASCR_001905)&text=Can%20be%20extended%20via%20packages,are%20available%20through%20CRAN%20family.

[fred-atherden]

Website

I am surprised this is not mandatory?

It's a good point - this will have been a legacy mulberry test. Not sure what the original requirement was.

I can't think of a reason why we shouldn't mandate one. If we agree, then I'll add a test

[fred-atherden]

Also, were we cutting out Curator entirely? There are a couple of tests that still refer to it.

It's not ever been used! I'll remove it.

[fred-atherden]

Nothing to add that hasn't already been said. Thanks @naushinthomson.

[griffithsc]

This looks great to me. Perhaps this bit could be clearer 'Year - the year of the latest commit...' - would be helpful to clarify what commit means in brackets

[naushinthomson]

If the authors cannot provide this information, the text should be left as is.

I don't understand this as the screenshot and text directly below contradicts it - All software here should be added as software references and cited properly

I was assuming that if the authors don't know the information we can just leave it as a link as we've tried our best but let me know if we should be doing something more to get the details we need! I've updated the caption for the screenshot to 'All software here should be added as software references and cited properly as far as possible.'

[naushinthomson]

What's the table at the end of the XML structure checks for?

Oops I missed that - it should be deleted as it's covered elsewhere! Done now.

[bcollins14]

Looks good to me!

[naushinthomson]

@FAtherden-eLife @JGilbert-eLife @Melissa37 @griffithsc @bcollins14 I've just updated the page again after our discussions last week! I've changed the section about R software and the 'When to add software references' bit. Let me know if everything is now consistent and ok!