search

research-compendium / how-to-read-a-research-compendium

"How to Read a Research Compendium" - based on S. Keshav's "How to Read a Paper"
https://arxiv.org/abs/1806.09525
Other
30 stars 3 forks source link

Create communities on data repositories/platforms to collect research compendia under one tag/community/group/topic #9

Open nuest opened 6 years ago

nuest commented 6 years ago

https://www.re3data.org/search?query=&contentTypes%5B%5D=Plain%20text&contentTypes%5B%5D=Raw%20data&contentTypes%5B%5D=Source%20code&contentTypes%5B%5D=other

nuest commented 6 years ago

Ideally these tags would have a common landing page... research-compendia.science ?

researchcompendia.science exists (Stodden an colleagues, stale since 2014)

Design/content inspired by http://www.theoj.org/, just a two-liner for what a research compendium is, and then the tags and links of the different platforms.

image

nuest commented 6 years ago

https://nuest.github.io/research-compendium.github.io/

image

benmarwick commented 6 years ago

Looks good!

nuest commented 6 years ago

Now online at https://research-compendium.github.io/

Repositorian commented 6 years ago

Dear Daniel, Ben and Carl,

Thank you for this significant contribution! From the research library perspective -- and one engaged in teaching open authorship practices to produce research compendia (or in our parlance, research papers of the future)- we are also thinking through the necessary components that enable this genre to enter the scholarly record. Findability, long term access/persistence, citeability, rights management are more complex with a multipart object than a simple 'paper' with supplements. Perhaps these concern will impact the archiving of research compendia in repositories more so than just choosing which platform to submit to.

Some particulars:

(1) Can DOI registration agencies add a new resource type for 'research compendium' to signal that the 'parent' DOI is for the aggregation, and that separate DOI's exist for the component parts. Right now the resource types available from CrossRef and Datacite, e.g, limit us to 'article', 'conference proceeding', 'dissertation', and recently 'data paper'. Will they offer a machine actionable way to navigate between objects within the compendium?

(2) Citation practice -- if one uses a dataset or code from the compendium, is the entire compendium to be cited or just the dataset or code, a la Force11 Data Citation or Software Citation principles? Some of this question may be tied to the following, due to license attribution requirements....

(3) Legal/Rights - How are research compendia to be handled in terms of rights ownership and rights notices? US Copyright law would likely give protection to the compendium itself based on the selection, arrangement, and coordination of the constituent components held within, but each constituent part may have its own rights status and license. How to interoperate the rights to remove barriers of access, reuse, and redistribution? This is a concern being looked at by Research Data Alliance's Legal Interoperability of Data Group, but we haven't addressed the specific use case of research compendia. Perhaps we will need to!

Looking forward to seeing how this discussion progresses!

Repositorian commented 6 years ago

Also forgot to mention the upcoming AuthorCarpentry Course at Force11 Scholarly Communication on this topic: https://www.force11.org/fsci/2018/course-abstracts#AM2. We'll likely include your paper and would love to discuss it with you during the course! (Would you be available to join us via web conference?)

nuest commented 6 years ago

Thanks @Repositorian for your feedback!

Re resource type: Excellent idea, though it might be premature given the still moving definition of what a research compendium is. Do you think the minimal one (cata + code + text) would suffice for the agencies? Do you know the process of suggestion such a new resource type?

Re citation: There is not way for "partial citations" that I am aware of, so I think this strongly depends on the success of research compendia and how/where they are deposited. Do you think a compendium should have more than one identifier?

Re legal: You're right in pointing out the complexity here, both in terms of legal systems (the German copyright law is quite different from US) and a compendium being by definition a collection of parts. Will you bring up the challenge of a compendim in the RDA group?

Re the Force11 course: a great programme you have there! The course is running 5 days, July 30 to August 3 - correct? There is some time difference but I can probably make it work since the courses are in the morning, so let me (and @benmarwick and @cboettig) know which time and date precisely would work for you and we can try to make that happen.

Repositorian commented 6 years ago

It seems that this genre is so promising more for its dynamic and reusable structure than solely because of the content. Text, code, and data can be bound together in static modes (eg PDF with embedded multimedia) and links to supplementary files on disparate servers.

But the compendium is a more tightly integrated super-object that can be reproduced with interactivity allowing users to inspect, interrogate, and replicate the work. It is bound by container technology that makes it a living work readily forked and adapted by its readers. It is fixed for the purposes of entering the scholarly record as a stand-alone output ( for social purposes like citation and rights management) So it is a first class scholarly output.