Sunday, 8 June 2014


New Models of Scholarly Communication for Digital Scholarship

University of Pittsburgh, School of Information Science


OVERVIEW: Contemporary research and scholarship increasingly uses large-scale datasets and computationally intensive processing.  Cultural shifts in the scholarly community challenge long-standing of academic institutions and call into question the efficacy and fairness of traditional models of scholarly communication. Scholars are also calling for greater authority in the publication of their works and rights management.  Agreement is growing on how best to manage and share massive amounts of diverse and complex information objects.  Open standards and technologies allow interoperability across institutional repositories.  Content level interoperability based on semantic web and linked open data standards is becoming more common.   Information research objects are increasingly thought of as social as well as data objects - promoting knowledge creation and sharing and possessing qualities that promote new forms of scholarly arrangements and collaboration. This talk will present alternative paths for expanding the scope and reach of digital scholarship and robust models of scholarly communication necessary for full reporting.  The overall goals are to increase research productivity and impact, and to give scholars a new type of intellectual freedom of expression.

READINGS:
    Griffin, S. (2013) Scholarly Communication: New Models for Digital Scholarship Workflows Coalition for Networked Information, Spring 2013 Meeting
    Griffin, S. et al (2014) The Denton Declaration: An Open Data Manifesto 
    Borgman, C.L. (2013) Digital Scholarship and Digital Libraries: Past, Present, and Future Theory and Practice of Digital Libraries Conference, September 2013

    Calhoun, K (2014) Exploring Digital Libraries: Foundations, practice, prospects Facet Publishing London, UK

35 comments:

  1. You mentioned a swell of "intellectual philanthropy" after the technological changes of 1980-2010. What do you think were some of the motivations and catalysts for such a shift in attitudes? Why was there such a swell of generosity?

    ReplyDelete
    Replies
    1. This is a question that I have been considering for some time... If the same were true for other aspects of human activities, the world would be a much better place for all.

      There are probably a number of contributing factors. One important factor is that the internet was not "owned" by a corporate entity or government organization. It was essentially a pure community based resource. Had the internet been commandeered by a large corporation, usage would have been dramatically altered. Another contributing factor was the adoption of the TCP-IP protocol. In this case, the network did not monitor (or censure) the data flowing through it - that responsibility (or lack of) was placed on the users directly. Finally, an open source operating system software (UNIX) was demonstrating clear advantage as compared to proprietary software.

      But these are technological considerations. There was also a shift in the culture and thinking of users, who found personal reward and satisfaction in contributing to a global resource. I cannot think of another set of external circumstances to compare. Having grown up in the 1960s, I thought I could discern echos of the idealism of that time in the attitudes of internet, web and repository developments in the 1990s.

      Delete
  2. Ces nouvelles méthodes, recherches et éruditions initiées par la technologie vont-elles finir par éliminer l’enseignant? Pouvons-nous nous passer d’humain pour transmettre la connaissance?

    ReplyDelete
    Replies
    1. Les livres analogues n'ont pas éliminé l'enseignant, donc pourquoi est-ce que les données numériques le feront? Mais c'est sûr que ce n'est plus nécessaire qu'un enseignant prononce un cours oralement plus qu'une fois (à part des mises-à-jour). Le reste sera de l'interaction, soit orale, soit en-ligne (comme ceci).

      Delete
    2. Translation of Clélia's question:

      Will these technical innovations eliminate the teacher?

      Delete
    3. No, I don't think so. Certainly, it will expand the realm of possibilities for teachers and students to incorporate new materials and means of analyses as well as options for local/non-local learning. The most important thing to me is that any course have a good deal of "creative space" built into it - I often shift the direction of my teaching according to the personalities and talents of the students and what they are eager to explore.

      Delete
    4. I think online teaching can improve learning. When the printing press became commonplace students no longer had to sit and listen to the 'chair' read the only copy of a book. Now, instead of listening to a lecture live, we can listen to it online and arrive to class with questions.

      However, if one mega-company takes over the online learning industry we could be in trouble. Say Google-University started. If everyone learns from the same source, we might loose knowledge diversity (this is less of a concern for maths or basic physics compared to neuroscience or the humanities).

      Delete
  3. I like Griffin's recommendation for a new model of scholarly communication which allows for modularity. Research with publishable sections decreases wait time for other researchers to use the findings, thus encouraging recombination, reuse, and collective usage. It reminds me of Heylighen's term stigmergy, which he defines as "a mechanism of self-organizing coordination between independent agents." It seems like the sciences have started to adopt open access/crowd-sourced research, but I have yet to see it much in the arts and humanities.

    ReplyDelete
    Replies
    1. You have hit on a key issue here. Scholarly communication in the sciences differs distinctly from that in the humanities, reflecting the different modes of inquiry. For the sciences, workflow follows successive stages of the scholarly research methods and these are clearly distinguished and have been universally adopted. Because the conduct of scientific research is agreed upon, modularity of reporting makes sense - as you say, it decreases wait time and accelerates enhancement and reuse.

      In the arts and humanities the tasks of establishing factual relationships, substantiating premises and building strong arguments leading to deeper understanding do not necessarily proceed in a standard methodological fashion. Scholars in the “digital humanities” often build new toolsets tailored for unique purposes, but are reluctant to publish these before they can thoroughly test them. It has been the case though that many tools created by humanists are modifications of those created for certain scientific domain applications.

      Delete
  4. I like the idea of releasing publications in stages. First, methods, then raw data, results, and finally interpretation. This manner of publication might encourage different research groups to perform concurrent experiments using the same methods.

    This may improve issues surrounding reproducibility. Rather than one group publishing an article, then other researchers failing to replicate it years later, this way researchers might output their results closer in time. If this is the case, researchers could easily compare and evaluate articles which use the same method. Otherwise we may accept the results of one experiment only to later find out it is not reproducible.

    ReplyDelete
    Replies
    1. It certainly helps motivating students—resonnating on Prof. Griffin's comment on digital scholarship breaking the solitude of the Humanities' grad student. I don't know if the environment of Humanities gives itself to reproducing results, though. I don't see how I could sell such a project to philosophy professors, who are only starting to accept that computer-assisted conceptual analysis is conceptual analysis.

      Delete
    2. Good points Louis. I'm coming from a neuroimaging background and wasn't thinking beyond the natural sciences.

      Delete
    3. I agree, I like how this model encourages a greater degree of transparency and accountability in research practice. If the process of verification of findings and critique of research methods becomes faster and more responsive through the ongoing publication of the research process, negligent research practices would be quickly exposed. A greater degree of rigour would be enforced.

      Delete
    4. Thank you for these comments. You have helped me to think about new and related issues.

      Delete
  5. M. Stephen Griffin a indiqué que des documents pourrait être créés à partir de bouts de d’autres documents, mais qu’il serait difficile d’associer ce nouveau document à un auteur. C'est le concept de modularité.

    Si on définit un document comme un composant/module, et que ce document a été écrit par un auteur spécifique, comment peut-on le décomposer? Qui s’occupe de le décomposer? Est-ce que la décomposition sera fait automatiquement?

    ReplyDelete
    Replies
    1. Translation:

      How can a document be decomposed into authored/credited components? Can this decomposition be automated?

      Delete
    2. This is still being considered. Ideally the document components should be linked directly to workflow stages. In the case of the data collection, examination and analysis, it would be feasible to produce document modules. Also the stage of determining how best to represent and present findings (information visualization, et al.). I do not see how the process can be automated except in the case of certain types research that proceed in silico throughout ...

      Delete
  6. I am very enthused about the potential role of libraries in future Open Data models of research workflow, notably by playing a role in publishing and digital stewardship. This strikes me as a powerful way of using already existing resources to subvert the monopoly of big publishing houses. My questions for Professor Griffin is: What are the main difficulties and obstacles to libraries playing this new role in research? Will this require a significant change in academic culture?

    ReplyDelete
    Replies
    1. I agree that libraries can and should play a major role in supporting research workflow, publishing, digital stewardship and many of scholarly activities. As you say, the big publishing houses journal subscription practices result in severely constraining libraries' discretionary budgets - preventing them from making new investments in more valuable assets and services. Several European countries have developed new arrangements with certain major publishers that ameliorate the financial impact to a point. But in the USA, so far little progress has been made. Any change in academic culture at the department level that decreasing the level of importance placed on publishing in a for-profit subscription journal and increases the reward for publishing in an open access journal or repository would simultaneously not only benefit the faculty, but reduce the pressure on libraries to subscribe to expensive journals. So my hope would be for enlightened and progressive strategies in university departments that encourage and reward faculty for open access publication of papers, data, tools, etc.

      I would also like to see more university libraries create centers for digital scholarship as part of their overall organization.

      Delete
    2. Thank you for the detailed answer, Professor Griffin!

      Delete
  7. I'm ready to start working on a sort of "research IDE". Here's a rough draft idea written down in 5 minutes. Is anybody interested in working on this?

    We need a software that allows to track progress of a search project
    - References
    - Data storage
    Integrating MATLAB, Excel, etc. data, formulas, scripts
    - Document writing and editing
    Including metadata
    Collaborative work
    - Annotations at all time
    - Branching versions

    ReplyDelete
  8. la perspective de digitaliser tout le 'Workflow' d'une recherche est très intéressante pour permettre la reproduction des expériences et des analyses.Toutefois, il est un peu difficile d'ouvrir l'accès aux données recueillis pour d'autres chercheurs qui peuvent les exploiter pour d'autres questions de recherche sont avoir eu a débrousser aucun dollars dans ce travail de recherche ...est-ce équitable !!!

    ReplyDelete
    Replies
    1. Il va probablement falloir une période d'embargo où il n'y a que les chercheurs qui ont généré les données qui auront accès pour les analyser. Sans ça, comme vous dites, on laissera les autres s'amuser à faire la cueillette des données et ensuite moi je l'exploite...

      Delete
    2. Translation of Eltaani's question:

      Open data is a good idea but is it fair to open access to data that the researchers have spent time and money gathering?

      [I replied that they deserved first-exploitation rights during an embargo period, depending on the field and project]

      Delete
    3. I agree with Stevan here. It should be noted that many funders require researchers to make all data created or gathered using grant funds available to the public. If the researchers have spent a great deal of time in bringing raw data to usable form, the question becomes more difficult. But if it is publicly funded research, the answer would still be yes, it is fair.

      Delete
  9. Digitize all the 'Workflow' of research is very interesting perspective to allow the reproduction of experiments and analysis. However, it is a little bit difficult to open access to the data collected for other researchers (who can use this for other research questions) without spending any dollar in this research ... What we can do against this risk.

    ReplyDelete
    Replies
    1. Good point: the researchers who gathered the data will probably need to be a period of exclusive access so they can analyze their own data. Otherwise I can wait for you to do the work of gathering the data, and then I will analyze it... (But this will depend on fields: in the human genome project, the data-gathering was all, until the collective database was completed, so they made it accessible immediately.)

      Delete
  10. As just discussed, you might want to have a look at the following platforms:

    http://orcid.org
    http://www.researcherid.com
    http://datadryad.org
    http://figshare.com

    They don't go as far as your ideas (yet), but might give you some ideas!

    Best,

    Stefanie

    ReplyDelete
    Replies
    1. Thanks! I'll be looking into them (datadryad looks great!)

      Delete
  11. I think it's a good idea for the whole workflow to be just as important as the completed analysis. This involves a shift in the way we think about research, where we value the process just as much as the final product. In the same way that open courseware and online lectures have become popular due to the efforts of some initial actors, perhaps this shift could be instigated by a few labs who start operating and publishing results in this way.

    ReplyDelete
    Replies
    1. Thank you for this good way of stating it and yes, if a couple of labs and/or university research groups lead the way, others will follow.

      Delete
  12. Thanks Stephen. I love it your big question at the last slide. I think we need to become more realistic. It is a relevant question for the web science.

    ReplyDelete
    Replies
    1. Thank you Albert and please post or send along to me any other thoughts you might have.

      Delete
  13. Dear Stephen, Thank you very much for your very interesting presentation ! I have one question : The web and others permit a lot of collaboration between researchers with different background. But understanding between multidisciplinary teams is still a challenge where simplicity is a common value. What's the ideal equilibrium to define rules and thresholds to permit simplicity and progress?

    ReplyDelete