Telecon 08.22.13 materials
- 'data scince' vs. 'data management'
Chris: Try 1st rather than 2nd term,... 2nd overly restrictive, looking to past rather than future the 1st is a piece of the 2nd, but broader ranges of issues that must be studied in a rigourous, scientific way e.g when C mangd ornl daac - q was, how much md (metadata) is needed? user group wanted most rigorous, but C felt not sustainable, was md to enable antoher sci to reproduce, or just understand methodology and ...
similar to sci sw - need enough doc to reuse, rerun, etc. a sci research q: what does sci need? reproducability, or enough to understand
'data science' would encompass broader issues someone from same field should be able to verify, ... but reproduce???
open q: what is sufficient, researchable consider unstructured data, map reduce, These kinds of questions would be part of the survey.
ensure credibility, ... coming from data how often are methods really reviewed???
some people mean something else in 'data science' eg data analysis, controversial phrase - more like an application science nrc - analyasis of massive data had to determine diff btwn 2 phrases q1: are we looking at all apps incl stats, vis, etc? c: scoping wkshop would help to determine boundaries but a labe; fpr sys study of data related issues, including data mgmt, ... etc eg library sci, info sci, why not data sci so meaning must be defined s: but must be careful - ... r: too broad a term,
describe some things that are *in* scope: eg verifiable, integrity
a few examples of types of forward looking questions
projecting how people will be using data in 20 years,
c: what is meant by md? esp in world of unstructured data and inferential, inductive science? these kinds of questions
b: how to ensure future use? esp w/ unstructued data eg map reduce db, will they be able to use it?
- 2nd draft is on line, what next?
inject some of the above questions into article, assertions to try back up eg insufficiency of current efforts, show bob and chris to mark where evidence is needed
also make nec logical leaps/steps
Anne to enquire about EOS limits, and time expectations a feature may not have limits -
probably shouldn't reference meeting, remove time bound stuff - 'This will be talked about ..."
ahopeful to converge on draft in 1 -2 weeks some non writers offered to review include panelists as reviewers - include as co authors also
- Getting a figure into the doc, grrr
Reaching out for funding to hold the workshop
- Mentzel, Moore Foundation: positive, will hear back on 8/26
- Zanzerkia, NSF: positive, need to provide her with some add'l material after figure arrives from Moore
earliest funding could arrive would be mid november, with some work needed before c: she was astute re: Survey sruving as unifying forum, longer term coord
- contact Martha???
P: eva indicated she would contact Martha once we provide amended write up, if she doesn't copy us on contact to M, we should follow up say this in cover note
- What roles must be filled?
- Eva was very pleased with Red Hat involvement in panel, wants private interests represented
b: make sense to have funders? role: funders? yes! also cross agency orgs eg USGCRP P: fgdc, ogc, might be sources fo ID private sector players chris greer NIST was head of IWGDD at OSTP, at NIST now, an interagency perspective current head of NITRD (OSTP) RDA?? - ask them for contacts to people other domains bio, health sci p: meeting with natl lib of med and other NIH people chris greer, 2nd vote by p
also strong RDA champion
funders, cross organizational people, private, academia & experts, people familiar with decadal surveys
number of participatins depends on output of wkshop dist btwn Earth sci focus vs other interested parties
others might take focus away from earth science
so careful how we use them
e.g. their case studies,
work w/ bill to set up wkshop drawing on his expertise, then have him facilitate rather than participate
agenda??? how will we organize it? break outs more than plenaries work w/ Bill to break down problem to determ q's to get right answers
goal: strong recommendataion in favor of survey, instr/words for how conduct generate broad comm support
r: EOS paper should give good set of q's to ask
c: meet in two weeks? yes.