ppy/EarthScienceOntolog-s01_chat-transcript_edited_20120823b.txt --------- Chat transcript from room: ontolog_20120823 2012-08-23 GMT-08:00 --------- [09:23] Welcome to the = EarthScienceOntolog: Panel Session-01 - Thu 2012-08-23 = Mini-Series Theme: An Earth Science Ontology Dialog ("EarthScienceOntolog") Session Topic: Value Proposition of Ontology and Semantic Technology for the Earth Science Community Session Co-chairs: Dr. LeoObrst and Professor KrishnaSinha Panel-Briefings: * Professor KrishnaSinha - "How Can Semantics Change Data Practices of the EarthCube Geoscience Community?" * Professor KrzysztofJanowicz - "The Value Proposition of Semantic Technologies and Ontologies for the Earth Sciences" * Dr. DaliaVaranka - "Rethinking the Meaning of Data for Integrated Science Problem Solving" * Professor PascalHitzler - "On the uptake of Semantic Web Technologies" * Mr. MikeDean - "Semantic Web for Earth Science" Logistics: * Refer to details on session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_08_23 * (if you haven't already done so) please click on "settings" (top center) and morph from "anonymous" to your RealName * Mute control: *7 to un-mute ... *6 to mute * Can't find Skype Dial pad? ** for Windows Skype users: Can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad" ** for Linux Skype users: please stay with (or downgrade to) Skype version 2.x for now (as a Dial pad seems to be missing on Linux-based Skype v4.x for skype-calls.) Attendees: AmitDeokar, AnneThessen, BobSmith, BobbinTeegarden, ChuckWard, DeanaPennington, DavidValentine, DeborahNichols, DeborahMcGuinness, DicksonLukose, DougFoxvog, FrankChum, GaryBergCross, GenhanChen, JoanneLuciano, JoelCarbonera, JohnGraybeal, KenBaclawski, KrzysztofJanowicz, LeoObrst, Mara Abel, MarcioFaerman, MarshallXMa, MikeBennett, MikeDean, NaicongLi, NancyWiegand, PascalHitzler, PatriceSeyed, PatrickVirden, PavithraKenjige, PeterYim, ScottHills, ScottPeckham, SiriJodhaKhalsa, ToddSchneider, TomTinsley, YannisRoussakis, YingjieHu, AlexShkotin, KrishnaSinha, mark, Uma, == Proceedings: == [08:34] anonymous morphed into DicksonLukose [08:58] anonymous morphed into MarcioFaerman [09:25] KrzysztofJanowicz: So far I cannot connect to the VNC server [09:25] PascalHitzler: yep cannot connect either yet :) [09:26] PeterYim: @Krzysztof & Pascal - don't worry, shared-screen (vnc) service is only optional [09:33] YingjieHu: I can see the vnc, but it doesn't allow me to input the password [09:35] KrzysztofJanowicz: same for me, you can download the slides at http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_08_23 [09:35] YingjieHu: OK, thank you [09:29] anonymous4 morphed into mark [09:30] anonymous2 morphed into GenhanChen [09:30] anonymous3 morphed into DeborahNichols [09:30] anonymous1 morphed into YingjieHu [09:30] anonymous2 morphed into NaicongLi [09:30] anonymous3 morphed into JoelCarbonera [09:31] anonymous morphed into TomTinsley [09:31] anonymous1 morphed into BobSmith [09:31] anonymous1 morphed into AnneThessen [09:32] anonymous3 morphed into PatrickVirden [09:32] anonymous2 morphed into ScottHills [09:32] anonymous4 morphed into ChuckWard [09:32] anonymous1 morphed into Cybershare - DeanaPennington [09:35] anonymous1 morphed into AlexShkotin [09:37] anonymous1 morphed into DavidValentine [09:39] anonymous1 morphed into Uma [09:40] anonymous1 morphed into BobbinTeegarden [09:41] anonymous1 morphed into MaraAbel [09:43] anonymous1 morphed into MarshallXMa [09:43] anonymous morphed into PatriceSeyed [09:49] anonymous morphed into YannisRoussakis [09:37] PeterYim: == LeoObrst started the session with the introductory slides [09:40] ToddSchneider: Leo, what does 'multi-level' ontology mean? [09:47] LeoObrst: "Multi-level" can mean 2 things: 1) the typical foundational, midlevel/super domain, and domain levels of ontological architecture, but also 2) multiple levels of granularity and contexts/perspectives in specific domains such as Earth Science. [09:51] PeterYim: @Leo - please prompt those on the call to log into the chat-room again (we have 45 people on the call now) [09:43] PeterYim: == KrishnaSinha presenting ... [09:56] ScottHills: Regarding discovery of data: This does not just apply to that held by individuals, but to that maintained in "Data Centers" as well. How many Data Centers might contain data of interest, and how many of those do you know about, and know how to use (each often has a different UI)? [10:14] KrishnaSinha: Hi Scott, The data centers are very visible and their data content is well known; an ontology framework that can map the data held in different centers would be welcome [10:15] PascalHitzler: @Krishna - what do you mean with "framework" here? [10:18] KrishnaSinha: @Pascal- an ontology based infrastructure that points to the data content of data centers [10:15] AnneThessen: This is true for many data centers, but there are some, such as Dryad that takes files with little to no guidance as to structure or content [10:16] GenhanChen: To Krishna: could you provide an example to explain the ontology framework on your slide #11? [10:19] ScottHills: Krishna, we may have different definitions for visibility. They may be accessible (your sense of "visible?"), but I question how many people know *where* to access them. For example, if I ask for the data centers that contain geochemical data, or seismic data, or well cores (take your pick), how many people could offer a comprehensive list? [10:26] KrishnaSinha: @ Scott-- a high level ontology that can capture the data types held at Centers would enable a user to find which centers hold what types of data [10:49] DougFoxvog: ScottHills asks above "how many people could offer a comprehensive list [of data centers that contain various types of geochemical data]". This seems to me to actually be a call for a knowledge base of data centers containing information about what types of data are covered in the various data centers. With linked data, such a knowledge base need not be centrally located, merely a vocabulary for making the statements needs to be defined, each data base could specify what types of data it covers, and the linked data needs to be made widely available. [10:41] ScottHills: Krishna, our thinking is starting to converge. The point I was working toward: I believe a good argument can be made for the need for ontologies that enable discovery of data across the landscape of proliferating Data Centers. I suspect the same ontologies would help enable discovery of data published by individuals. That said, I suspect the "level" of such ontologies need to go beyond data type, or their utility will be very limited. [10:44] KrzysztofJanowicz: Scott, yes ontologies should always be 'more' than just data models [10:47] KrishnaSinha: @Scott--I agree, and I also think most users would be comfortable with high to medium level granularity [10:19] AmitDeokar: @Krishna - We notice a lot of emphasis on managing data. What are your views on how important it is to manage associated computational models, share them, annotate them, and so forth so that they are accessible like data, avoiding problems of reinventing the wheel? [10:22] KrishnaSinha: @ Amit- i support an ontology framework that organizes services that would include existing models and other computational tools..its the service ontology on slide 11 [10:41] JohnGraybeal: @Krishna: Re your endorsement of data citation: Why should this community particularly endorse data citation, more than it should endorse all the data practices that benefit from ontologies: description, provenance, versioning, unique identification, ...? All seem noble causes. [10:58] LeoObrst: @Krishna: I think one very important issue you raised is the incentive for individuals (and organizations) for sharing data: in the research community, what is the incentive for a researcher to expose his/her data and its meaning? What supports scholarly acknowledgment of the contribution? [11:02] KrishnaSinha: @Leo--There are two ways to expose data..regular publications for which there is a citation index ( often used for promotions etc) ..the critical issue is what incentives we can support to have people share data that is not formally published..data citation index is one way to go [09:57] anonymous1 morphed into SiriJodhaKhalsa [09:58] anonymous morphed into PavithraKenjige [09:59] anonymous1 morphed into DeborahMcGuinness [10:02] anonymous morphed into KrishnaSinha [10:00] PeterYim: == KrzysztofJanowicz presenting ... [10:00] PavithraKenjige: Did he say that we have such a semantic infrastructure already available?? [10:01] ChuckWard: Have you also considered community expectations such as data archiving requirements with journal articles? e.g. evolutionary biology [10:09] SiriJodhaKhalsa: what is the advantage of creating or using an ontology for encoding provenance/quality/appropriateness-of-use information beyond creating metadata according to a well-defined standards like ISO 191**? [10:37] JohnGraybeal: To Siria Jodha (Hi!): The advantages of using ontologies for encoding any content, *including* provenance/quality/appropriateness-of-use information: (1) Readers of the metadata know exactly what you mean ('exactly', relatively speaking), (2) there is better mapping/understanding across communities for these concepts. (While I appreciate ISO 191** provides vocabularies, I often find them underdefined, incomplete, and locally inconsistent. And often ISO 191** doesn't provide a vocabulary, so the content is wide open.) [10:09] PeterYim (added subsequently): @SiriJodhaKhalsa - you might want to take a look at the two OntologySummit communiques - Ontology for Big Systems: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_Communique and Making the Case for Ontology: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2011_Communique - where some of the benefits were enumerated by the community [10:23] LeoObrst: @Krzysztof: you cited some research concerning the last bullet on translation between conceptual models on slide 5. Can you provide that reference? [10:27] KrzysztofJanowicz: Yes Leo, for instance the work of Mark Gahegan. This paper may be a good starting point and overview: A Semantic Web Map Mediation Service: Interactive Redesign and Sharing of Map Legends by Mark Gahegan, Will Smart, Sina Masoud-Ansari, and Brandon Whitehead. [10:52] PeterYim: whoever said, "Krzysztof, fully agree with your view. Understanding formal semantic and reusing well fundamented ontologies. " ... please note that the space to type your message is next to the "send" button (at the bottom) and not the box next to the "hand" button [10:15] anonymous morphed into DougFoxvog [10:15] anonymous morphed into ScottPeckham [10:20] PeterYim: == DaliaVaranka presenting ... [10:29] PeterYim: == PascalHitzler presenting ... [10:30] KrzysztofJanowicz: "Did he say that we have such a semantic infrastructure already available??" --> I would argue so [10:52] ScottHills: Pascal, would you agree that whether an ontology is high-quality depends on how well it satisfies the needs for which it was engineered? [10:55] PascalHitzler: @Scott - this looks reasonable, however I would think that more can be said about the "quality" issue. In fact, there is quite a bit of work on "Evaluation of Ontologies", and there is also a established workshop series about this topic. [10:57] BobbinTeegarden: @Pascal URL for the workshop? [11:01] PascalHitzler: The workshop series - see e.g. http://km.aifb.kit.edu/ws/eon2007/ . [10:58] PascalHitzler: I suggest to start with DennyVrandecic's PhD thesis: http://www.aifb.kit.edu/images/b/b5/OntologyEvaluation.pdf [10:48] PeterYim: == MikeDean presenting ... [10:51] anonymous morphed into MarshallXMa [11:18] LeoObrst: @Mike: do you have a url for the LOD Framework initiative? [11:02] PeterYim: == open discussion commences ... [11:03] PascalHitzler: re. "Linked Data Quality", have a look e.g. at http://blog.semantic-web.at/2012/08/09/whats-wrong-with-linked-data/ [11:05] PascalHitzler: Leo / Krishna: The Semantic Web journal, for example, recently established "Linked Dataset Descriptions" as regular paper type. The purpose of this introduction is to generate "academic incentive" for the production of high-quality Linked Datasets. [11:07] MarshallXMa: @Pascal: Another journal (Geoscience Data Journal) with similar ideas: http://onlinelibrary.wiley.com/journal/10.1002/%28ISSN%292049-6060 [11:03] MarshallXMa: @LeoObrst: Such as, a stable repository to upload and describe (with a metadata form) the dataset, and easy way to assign a DOI for it. [11:11] ToddSchneider: John, part of the needed infrastructure is the Open Ontology Repository. [11:12] JohnGraybeal: Yes (the question was, "What has to happen on the ground for these advantages to be obtained?" With particular attention to the practical realities of the practicing data managers. [11:13] JohnGraybeal: @Todd: As provider of an OOR equivalent (MMI's ORR), I want to go farther. Let's assume there's a perfect repository in existence. And even, really good vocabularies from communities. How do they integrate this into their developed data systems? [11:14] ToddSchneider: John, education. It would appear that many people working in the area of 'semantic technologies' don't know enough. Dalia addressed this. [11:16] KrzysztofJanowicz: John, IMHO it needs a earth science specific tutorial and a set of reusable building blocks [11:14] ScottHills: Pascal, thanks for the URL. I'll look through the material (up through 2007?). Any chance I'll find suggestions to answers like, what applications are best served using SKOS vs. OWL DL with domain-specific properties? [11:18] PascalHitzler: @Scott - I would guess that answers to such specific questions can probably not be found easily. "Best practices" in applying semantic technologies are not (yet) written up in any concise form, afaik. Note that it is only two years or so that we got textbooks which cover the bare basics of the discipline. Currently, I guess the best approach to get such questions answered is to consult a specialist. [11:15] MarcioFaerman: My question referred on how to justify and support, throughout the community, @KrzysztofJanowicz "Final Thought" slide - "... It seems that we hope to arrive at semantic interoperability by standardization instead of investing into research on alignment and semantic translation to reduced incompatibility. This may turn out to be a fundamental misconception. I believe that standardization is the more difficult of both approaches..." [11:18] MarcioFaerman: Then I followed up with the question about conciliating multi-disciplinary research and knowledge discoverability through semantic translation. [11:16] MikeBennett: Something we are finding is the distinction between one ontology for one application, versus overall domain ontology (standard) which is use-case neutral. Both are vital. And different. [11:19] KrzysztofJanowicz: MikeBennett, yes -- I am not arguing against higher level ontologies. It needs both. [11:19] MikeBennett: @Krzystof agreed. Work to be done on the heuristics for extracting a relevant sub-set of domain ontology for a given use case. [11:22] JohnGraybeal: @Mike: In addition to single ontology vs overarching ontology, there are typically multiple authoritative (sic) vocabularies within each domain, and maybe one or two ontologies as well. Making integrative use of that information is, for now, impossible without a lot of human investment. [11:24] MikeBennett: @John very much so. Needs to be done across industries. One chance to get it right I think. [11:25] KrzysztofJanowicz: IMHO, this makes GeoSPARQL so strong [11:27] ScottHills: Pascal, thanks for that info (13:18). Good to know, since I haven't been able to find such material to date. [11:29] JoanneLuciano: Thanks to the organizers and presenters! [11:29] FrankChum: W3C Oil, Gas & Chemical Business Group: http://www.w3.org/community/oilgaschem/ is developing a use case to collaborate on ontologies for drilling automation. Involving multiple Oil & Gas standards organizations. [11:29] PeterYim: == Final remarks by KrishnaSinha ... [11:31] FrankChum: Thank you. [11:31] PeterYim: great session! [11:31] KrzysztofJanowicz: thanks, bye bye [11:31] LeoObrst: Thanks, All! [11:31] PeterYim: -- session ended: 11:31am PDT -- ---------