HCLSIG/SWANSIOC/Meetings/2010-4-19 Conference Call

From W3C Wiki

attending: 

Alex Garcia: Uni Bremen
Anita DeWaard: Elsevier
Alex Passant: DERI
Eric Prud'hommeaux; W3C
Joanne Luciano
Jodi
Patti Davis: MIND
Paolo Ciccarese: MIND
Tim Clark: MIND
Tudor Groza: DERI

Note: This call was affected by extensive echoes on the concall line

Minutes: Preliminary Discussion 

Anita:  What this call is about, previous presentations reviewed
Anita: Has reviewed a large set of biomedical journals for coarse grained structure, all follow IMRAD
ANITA: IMRAD = Introduction, Methods, Results, And Discussion

Anita's spreadsheet as PDF: File:AnitasReview.pdf

Tim: For more complete semantics, should include the <journal, title, authors> and <references> at front and end
Anita: This is called the Heads and Tails in the publishing industry
Tim: Suggest we proceed by making a very coarse grained model compatible with IMRAD, SALT & ABCDE, but not assuming more than IMRAD
Anita:  But shouldn't we consider NLM DTD as well, this is a fairly universal model of biomedical publications?

Further Discussion

  • [10:32] <jodi> So this is the link, I think: http://dtd.nlm.nih.gov/
  • [10:33] <jodi> that makes sense to me
  • [10:33] <Tim> what are the models you are looking at?
  • [10:33] <Tim> would that be ABCDE & SALT & ...?
  • [10:34] <jodi> Use cases are linked here: Tim: the
  • [10:34] <jodi> err, here: http://esw.w3.org/HCLSIG/SWANSIOC/Actions/RhetoricalStructure
  • [10:35] <jodi> So IMRAD, NLM tag set, and what is the 3rd model?
  • [10:35] <agarcia2> http://esw.w3.org/HCLSIG/SWANSIOC/Actions/RhetoricalStructure/UseCases/3
  • [10:35] <Anita> medium-grained models; SALT, ABCDE, SWAN
  • [10:35] <Anita> http://esw.w3.org/HCLSIG/SWANSIOC/Actions/RhetoricalStructure/UseCases/1
  • [10:35] <Anita> http://esw.w3.org/HCLSIG/SWANSIOC/Actions/RhetoricalStructure/UseCases/2 - less relevant for HCLS
  • [10:35] <jodi> thanks Anita!
  • [10:36] <Anita> Tim: can we establish a coarse-grained model that all these could be compatible with the NLM DTD
  • [10:36] <Anita> Compatible with IMRaD and NLM tag set and medium-grained model
  • [10:36] <jodi> coarse-grained model that works with IMRAD, NLM tag set, and compatible with ABCDE
  • [10:36] <Anita> Anita: that makes a lot of sense!
  • [10:37] <Anita> Head, Body (IMRaD), Tail: generate off of existing documents -
  • [10:37] <Anita> REflect.ws is tool that Tim is discussing
  • [10:38] <Anita> Coarse rhetorical structure would tell Reflect what section is in
  • [10:38] <Anita> Simpler than exposing the entire DTD _ but yes, make a Linked Data representation of Head, IMRaD and Tail
  • [10:39] <Tudor> Not sure what is the difference between the Linked Data
  • [10:39] <Anita> So start with coarse-grained thing, then go more fine-grained with a) IMRaD within body
  • [10:39] <Tudor> and what we did until now
  • [10:39] <Tudor> for example with SALT?
  • [10:39] <Anita> b) CiTO within tails
  • [10:39] <Tudor> (w.r.t. coarse-grained structure)
  • [10:40] <Anita> Paolo: compatibility - not sure if we can do this, or just decide on following one model?
  • [10:41] <Anita> Compatibility: no problem - SALT, abcde, IMRaD are roughly the same.
  • [10:41] <Anita> Linked data: what do you mean by this representation?
  • [10:42] <Anita> in KonnEX KonneX-SALT: First Steps towards a Semantic Claim Federation Infrastructure - semanticweb.org http://bit.ly/dgCB31
  • [10:42] <Anita> Next time: discuss Annotation Ontology we (Tim) submitted
  • [10:43] <Anita> Semantic is metadata, linked-data representation of the Metadata - structure the publication
  • [10:44] <Anita> A joke running around: the Europeans asked Iceland for Cash, not ASH!
  • [10:44] <Anita> Manual exposure of structure, how does this happen?
  • [10:45] <Anita> Tim: once we have this specified, we map our DTDs to this structure - expose XML directly.
  • [10:45] <Anita> Coarse-grained structure - finer grained Elsevier structure
  • [10:46] <Anita> HOw is structure applied - on papers that are already out there?
  • [10:46] <Anita> Paolo: we are working on an annotation framework, can apply different annotation types, some are manual, some are automated.
  • [10:46] <agarcia2> sur
  • [10:47] <agarcia2> sure, there are manual annotations -those comming from users. there are automatic annotations, those comming from annotation pipelines like WhatIzit
  • [10:47] <Anita> Semi-automatically recognize sections - Tim: can use it on any biology paper. Recognize every section on an IMRaD paper
  • [10:47] <agarcia2> reflect is an automatic annotation pipeline
  • [10:48] <agarcia2> I am working on exposing a paper in link data format
  • [10:48] <Anita> http://bit.ly/aiusPc
  • [10:49] <Anita> http://bit.ly/bnHMq2/ - German National Library
  • [10:49] <jodi> HIVE project might be relevant: https://www.nescent.org/sites/hive/Main_Page
  • [10:50] <Anita> What would it take to make a first pass of a semantic model of the type we are talking about? Basic agreement at some point!
  • [10:51] <Anita> Tudor?
  • [10:51] <Anita> Can you take what was being done for KonneX and expand it, perhaps?
  • [10:51] <Anita> Tudor: sure, don't see a problem! yay!!!
  • [10:52] <Anita> Tm -> Paolo - present AO? Yes, let's do that next meeting - wait till the paper is actually accepted.
  • [10:52] <jodi> VIAF, the virtual international authority file, is an effort to harmonize author information among different national libraries. Described at * http://www.oclc.org/research/activities/viaf/ home for it is http://viaf.org/ ; OCLC is doing it, LC, German National Library, a number of other libraries are involved
  • [10:54] <Anita> Let's not support claims in this model - but be compatible with (those weirdos :-) who support claims)
  • [10:54] <Anita> Konnex supports coarse-grain, rhetorical blocks and fine-grained markup
  • [10:55] <Anita> Exposed via KonneX - coarse-grained structure can be exposed -
  • [10:55] <Anita> Tim: would love to see something that we can work with
  • [10:55] <Anita> Schema is OWL, published as linked data in RDF!
  • [10:56] <Anita> Tudor: a repository, or some overlaid form of structure
  • [10:56] <Anita> Tim: our model would be an overlaid structure - does not require control of the document
  • [10:56] <Anita> Tudor: but then you don't have access to the IMRaD structure! Translation is needed...
  • [10:58] <Anita> Come up with something as a strawman - repository or overlay, either is great! AO will allow you to do it as an overlay
  • [10:58] <Anita> Tudor's task but Anita will provide input!
  • [10:59] <Anita> Arrange a call on this in detail

[11:00] <Anita> We will take a collection of documents and model them in this way!

Retrieved from "https://www.w3.org/wiki/index.php?title=HCLSIG/SWANSIOC/Meetings/2010-4-19_Conference_Call&oldid=44011"