|
|
|
|
|
||
|
|
|
Present:
Agenda
Our potential role take the lead in standards development would look for additional money - 1 year time frame, in parallel with rest of our tasks Walter, GBIF, Frank Bisby are all interested Bob - concept based system that many institutions could use. institution that would serve and use concepts. Standards alone are necessary but not enough. We need something to test in one year Jim - we need to be application driven. But prototype tests to interact with other systems Vegbank, prometheus, etc. Needs to be funded Jessie - must show that it works, not just a seek model Martin - risk management - what is the risk to seek goals if we get involved with TDWG standards. Jim
Bob would get ecologists to markup right Martin GBIF has different immed. goals that are not nec. same as seek. Jessie - do it our way in broader context, then hope others like . can't guarantee that others will buy in. we must do it anyway Jim - others may drop out, what is our risk to spend time and money. Could be major disagreements bw different perspectives Susan can make work together with wrappers even if somewhat different Jim - need all of: schema, repr methods, documentation of schema, transfer protocol, to prove our stuff, others will want software, we are committing to having stuff for others, formalizing schema is only additional work. we needed to do this anyway Jessie - we need our own schema which is subset of federation schema. We will need to be able to anticipate others' needs for interfaces - no it's others' problem Bob - must ask for more money to interface with others Jessie - extra work - must really study/communicate with other systems to collaborate Jim - MOU with GBIF to say how far we will go - whoever is materially involved. Everyone should understand, umbrella of TDWG, but they have no material investment. understanding with TDWG, not vested interests of it. They must have buy-in. Self-nomination for standards - if no stake in it, then just noise, must have money, people, something involved. Bob - supplemental request 20% of SEEK taxon money, need to submit to stinger in 30 days (over 100K). Walter who is involved in management, if we get financial/in-kind investments, what is our commitment to them Susan - we are to develop union of existing schema and how it maps to existing datasets Bob - need buy in from datasources Susan - no just need a portal Jessie - we will contribute % of existing work to broader effort - what will others contribute - only commit year by year. Need $ for travel, indiv. meetings Jim - must meet with Frank, Walter, GBIF - we are prepared to do it, see if they will contribute materially Jessie - meet with people with models, try to extend our model to workable schema, produce something, discuss proposal with everyone, meet with everyone Need money for Jessie and one hire to come with, learn on job, write up. Discuss in Edinburgh. Susan - in year 2 or 3, money for wrapper, generic tool for dataset merging, Thau:
SMS, OWL, Taxon
Context - why does OWL make sense in Taxon and other working groups. SMS group is using OWL and these tools, (Genna - Geon too) This is the language of the semantic web Jessie - nobody else is doing things in OWL at this scale. Taxonomists do not want computers to infer anything. Internal representations of our classes in OWL would be death. Susan - could dump out small pieces into owl, then play and reason with them Jim - would inferences be more important with other stuff? Nico - could represent with indentation - nobody looks at 1000s at a time, so not a problem Dave - when someone registers datasets and enter tax info, SMS finding datasources, could use language to constrain the way that ecologists enter info. Language for computers to talk to each other Jessie - OWL/ ontologies to represent other data (did I get this point?) don't know how we can represent everything in OWL Susan - could spit out small piece in owl then SMS can do other work to compare pieces How do you compare models to each other - how do you represent the models OWL is good: standard for describing ontologies
OWL is bad:
Next:
Other possibilities:
Dave- Useful aspects of OWL/ontologies that you can't get from XML schema? Jessie - taxonomic classification =? ontology (or is it very narrow subset). If we allow others to build sub-ontologies, OWL might be good tool to create then store in another way. other - need to reason b/w other hierarchies and tie in with SMS could just give SMS unique id's, use RDF RDF - subject --> property --> value, standard format for unique id, resource G - ontology is not a distribution mechanisms, Dave - yes, just makes it easy/convenient RDF schema - fundamental to semantic web Nico: TODO: GET SLIDES FROM NICO
Need to work on YET another model Aim - 3 dim past "Genospace" / "Phenospace" soup --> Evolution --> present non-homogenouse distribution of clusters The (typically imperfect) stability of clusters of properties in space and time (kinds) is a precondition for taxonomic names. Discover properties of unique names aim of systematics is to discover and name Kinds The continuity of refernce in absence of an exact relation beween predicate (name) and referent (species) is possible because of a causal chain among the communicators (taxonomists) that stretches from the initial naming event to the present. Must model frequency or stability of use. Susan - would be good to see differences between versions of taxonomies Martin - revisions can take small portions based on groups one is familiar with - there are ways to decide what makes up the subset of revision, could be based on character data Bob - find something as example of best practice Martin - cannot do cladistics in a vacuum, no way to tell whether your patent is better/worse than another, just different Jessie - Nico should come up with what is wrong with our model Susan - give us an alternate proposal Jessie - keep in mind difference between individual and species. species is an idea. This is an individual taxonomists view of what they are trying to do. Not taxonomists as a whole. This is what the individual thinks he has achieved. Jim - How does this fit in with Tree of life , cladistic model, intermediate nodes without names, could just have diagnoses, characteristics Martin - only need to name terminal nodes, internal nodes (clades?) only need internal reference. cladogram is not a formal classification with??? Jim - Most phylogeneticists are not doing this post-classification. Can our system handle this in 10 years when ??? Martin - Concepts are in terminal taxa (cladistic classificiation) Susan - store the links between the nodes not a problem ??? can handle them as long as there is a link to something that we recognize Jim - most systemestists do not go back and name all the leaf nodes Nico - ??? clarification of our mission Jim - formal classification is fading away - breaking down on the edges, we decide whether we can deal with it with our model
Jessie - get info from ecologist what field guide is used Susan - get a 2nd hierarchy, then Implications of new taxonomic schema Must get a data source with real concepts to fill new schema - better test
datasets - Birds of Mexico does Town have bird concepts too? Bob will export his data in XML Jessie will give us input on mapping of fields Dave will write XSLT translator to go to our latest schema January meeting: Jim- How can we interact more effectively between meetings Use cases - haven't done that
Susan -
Have a mission for each meeting
Maybe we should look at a tool that takes a dataset with taxa names and produce EML (taxonomic coverage only) Jessie - Do we ever record just genus or family? If so, we should never expand it (in the EML) Susan - need great data on German mosses and datasets with great EmL detailing tax coverage of German mosses, good If we have another taxonomy - it is fine if we have Use case - ITIS has full tree of weak concepts have another tree of full concepts, small tree tools for importing conceptual data map concepts between trees Dave tasks: look at use cases?
Jessie - will work on visualization tools to do the mapping between trees (for someone who knows the data) Nico - Treemap - maps topologies - hypothetical extensions, doubling of trees, minimizes steps between trees We should all email more often
Commitment for others within management process (decision making):
|
| This material is based upon work supported by the National Science Foundation under award 0225676. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF). Copyright 2004 Partnership for Biodiversity Informatics, University of New Mexico, The Regents of the University of California, and University of Kansas |
