15

Proposals: Open Source and Open Science

As far as I can tell, there is a significant overlap between Open Source and Open Science. Where exactly does this overlap lie and will this be a problem?

As of writing Open Science is much further along than Open Source is and I can imagine that followers of Open Science are not willing to merge with Open Source. To followers of Open Science: is this assumption correct?

12

There is a degree of overlap in which questions will be on topic for each of the two sites. This is not necessarily a reason to combine them into one site. There are many SE sites that have overlap, or even where one could reasonably be a strict subset of the other.

What distinguishes a topic as being suitable for a separate SE site is its community. While Open Source and Open Science have a lot of goals, ideals and attitudes in common, they are two different communities, which benefit from having two different sites. Those people who are both scientists and developers can join both sites. Those people who are only one or the other can join just one site, and see questions and answers that relate directly to their community.

  • 2
    I completely agree, we in graphicdesign SE have some overlap with UX and superuser, but the communities are completely different. – Yisela Feb 3 '15 at 11:43
  • 1
    I agree, Open Science is primarily meant as a research style and initiative and has many challenges that differ from Open Source issues, although Open Source implementation may be often recommended for Open Science workflows. – puslet88 Feb 18 '15 at 8:44
7

IMHO this is 2 separate but related questions masquerading as one, and these questions are so related (by the nature of open science) that I'll add a third (since it's already been asked):

Q1. How do "open source" products (e.g., content (e.g., analyses, images, publications), data, software) relate to the practice of open science?

Open science is fundamentally (though not exclusively!) about increasing or optimizing reproducibility. (Reproducible science has several benefits, including

  • verification: reproducible results are reasonably considered more reliable
  • training: reproducing results of increasingly non-trivial studies is (or should be) a key part of scientific apprenticeship
  • extension: partial reproduction, with intentional variation, is key to the accumulation of "normal science"

) In the best, most "open-science" case, any competent practitioner should be able to access the inputs, procedures, and outputs of a study (e.g., experiment, model run), and verify that applying the procedures to the inputs produces significantly-similar outputs. Obviously (IMHO--ICBW) a scientific study can be most generally reproducible, and (I claim) therefore most open, when

  • its inputs are open data
  • its procedures are FLOSS where computational and open content where not (e.g., descriptive text)
  • its outputs are open content (e.g., images, text) and open data

Hence I assert the practice of open science will require (et al!) skills in using and generating open content, data, and software, in a manner similar to that which developing FLOSS draws upon more general informatic skills. Furthermore, "doing open science" will typically require specialization in

  • generation of particular content, i.e., scientific publications
  • use of particular infrastructure, e.g., data repositories, computational clusters

Q2. How would StackExchanges={Open Source, Open Science} overlap and will this be a problem?

Q3. How would StackExchanges={Open Data, Open Science} overlap and will this be a problem?

My answers to both are the same: IIUC an Open Science SE would often point back to (and, probably all too often, mark as duplicates) issues specific to SE=Open Data or (should it happen) SE=Open Source. But SE=Open Science will (I suspect)

  1. often involve "emergent" or "higher level" compositional and infrastructural issues. E.g., consider the way that IP networking has a vast, specialized problem domain conceptually distinct from the codes and OSes that enable it (i.e., that are the "platforms" for IP networking) and from the data transported by IP networks.
  2. involve or target a distinct (scientific) community (as pointed out by trichoplax here).
  • Open Data is wildly different than open source however I see the common theme you described here. – Gram Jul 22 '15 at 15:29
6

I don't see how it could work. Those have two separate audiences.

Open Source is the community of developers, who create software together, independent of companies and government. They are not scientists. They do not conduct any researches. Some of them may be scientists, but they may be also travellers, historicians, linguists etc. It's like boat programming. If a programmer likes boat swimming, it doesn't make boat questions on topic on SO.

Open Science is the community of researchers, who conduct research together, independent of companies and government, and make all their researches available to public. Some of them may be developers etc.

I feel the difference between them is like between the Workplace and the Academia. There's a strong common point (working for someone) but the issues are too different, so they are complete separate sites.

  • Surely Open Science will not have many questions directly about science or research. Those would belong on sites like physics.SE, chemistry.SE, etc. and academia.SE. As I said, I'm not the most knowledgeable in this area, but what's left seems like it could fit on Open Source. I'm not going to vote on this answer though, I'll leave that to people who are more qualified to judge this than I am. – overactor Jan 21 '15 at 10:29
  • 1
    Don't you think that the suitability of different open source licenses for doing open science may overlap? In my opinion such overlap exist but it's not significant enough to be a problem. – ncasas Jan 25 '15 at 16:49
  • Open science isn't by definition "independent of companies and government." Governments, many foundations, and some companies specifically fund open science (e.g., open access publication, open licenses on data, transparent research processes, etc.). – Thomas Jan 25 '15 at 20:13

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .