2066 and all that


Published on

Talk at Joint ORCID - Dryad Symposium on Research Attribution, St Anne's College Oxford, 23 May 2013

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

2066 and all that

  1. 1. David De Roure2066 and all that
  2. 2. http://www.scilogs.com/eresearch/pages-of-history/ David De Roure
  3. 3. A revolutionary idea…Open Science!
  4. 4. More peopleMoremachinesThis is a Fourth Quadrant TalkBig DataBig ComputeConventionalComputationThe Future!SocialNetworkinge-infrastructureonlineR&DThe FourthQuadrant
  5. 5. F i r s tBioEssays,,26(1):99–105,January2004http://research.microsoft.com/en-us/collaboration/fourthparadigm/
  7. 7. Information CircuitsCommunitySoftwareExecutionDigital MusicCommunityannotationLinked DataRepositoriesWorkflowsGeneratePaperConference
  8. 8. Notifications and automatic re-runsMachines are users tooAutonomicCurationSelf-repairNew research?
  9. 9. Nigel Shadbolt et al
  10. 10. The challenge is to foster the co-constituted socio-technicalsystem on the right i.e. a computationally-enabled sense-making network of expertise, data, models and narratives.This requires a “social machines” perspective from the outsetas well as humanistic input. The Web, and with it WebScience, are an important exemplar.Big data elephant versus sense-making network?
  11. 11. datamethod
  12. 12. http://www.myexperiment.org/
  13. 13. Research ObjectsComputationalResearch ObjectsEvolving the myExperiment Social MachineWorkflowsPacksOAIOREW3CPROV
  14. 14. Reusable. The key tenet of ResearchObjects is to support the sharing andreuse of data, methods andprocesses.Repurposeable. Reuse may alsoinvolve the reuse of constituentparts of the Research Object.Repeatable. There should besufficient information in a ResearchObject to be able to repeat thestudy, perhaps years later.Reproducible. A third party canstart with the same inputs andmethods and see if a prior result canbe confirmed.Replayable. Studies might involvesingle investigations that happen inmilliseconds or protracted processesthat take years.Referenceable. If research objectsare to augment or replace traditionalpublication methods, then they mustbe referenceable or citeable.Revealable. Third parties must beable to audit the steps performed inthe research in order to be convincedof the validity of results.Respectful. Explicit representationsof the provenance, lineage and flowof intellectual property.The R dimensionsReplacing the Paper: The Twelve Rs of the e-Research Record” on http://blogs.nature.com/eresearch/
  15. 15. www.researchobject.org
  16. 16. Join the W3C Community Group www.w3.org/community/roscwww.researchobject.org
  17. 17. Real life is and must be full of all kinds ofsocial constraint – the very processesfrom which society arises. Computerscan help if we use them to createabstract social machines on the Web:processes in which the people do thecreative work and the machine does theadministration… The stage is set for anevolutionary growth of new socialengines. Berners-Lee, Weaving the Web, 1999The Order of Social Machines
  18. 18. Some Social Machines
  19. 19. ScholarlyMachinesEcosystem
  20. 20. Trajectories... distinguished by purposeTrajectories through Social Machines https://sites.google.com/site/bwebobs13/
  21. 21. http://www.scilogs.com/eresearch/pages-of-history/DavidDeRoure
  22. 22. 1. It was no longer possible to include theevidence in the paper – container failure!“A PDF explodedtoday when ascientist tried topaste in thetwitterfirehose…”
  23. 23. 2. It was no longer possible to reconstruct ascientific experiment based on a paper alone
  24. 24. 3. Writing for increasingly specialist audiencesrestricted essential multidisciplinary re-useGrand Challenge Areas:• Energy• Living with Environmental Change• Global Uncertainties• Lifelong Health and Wellbeing• Digital Economy• Nanoscience• Food Security• Connected Communities• Resilient Economy
  25. 25. 4. Research records needed to be readable bycomputer to support automation and curationA computationally-enabledsense-making network ofexpertise, data, models andnarratives.
  26. 26. 5. Single authorship gave way to casts ofthousands
  27. 27. 6. Quality control models scaled poorly withthe increasing volume
  28. 28. 7. Alternative reporting necessary forcompliance with regulations
  29. 29. 8. Research funders frustrated by inefficienciesin scholarly communicationAn investment is only worthwhile if• Outputs are discoverable• Outputs are reusable• Outputs accrue value
  30. 30. Discussion points1. Citation in tomorrow’s sense-makingnetwork of humans and machines:– What are the artefacts / social objects?– How and why are they cited?2. Think about an ecosystem of interactingScholarly Social Machines3. Science as Social Computation?
  31. 31. Thanks to Jun Zhao, Kevin Page, my{Experiment,Grid}, wf4everdavid.deroure@oerc.ox.ac.ukwww.oerc.ox.ac.uk/people/dderhttp://www.scilogs.com/eresearch@dderwww.digitalsocialresearch.netdigital.humanities.ox.ac.ukwww.researchobject.orgwww.software.ac.ukwebscience.org