In my understanding, TCS assumes that addition of a new lower taxon to a taxon changes the taxon concept. For example, if one found a new species and assign it to genus Eucalyptus, then we have yet another entry of genus Eucalyptus. Is it right understanding?

TrevorPaterson Notes: I don't see why you would necessarily need or want to create a new concept for Eucalyptus. It would depend largely on the definition of Eucalyptus that you were happy with. You can create a concept for your new species taxon according to 'Ytow2004', which includes as part of its definition the relationship - "is a child of the concept Eucalyptus according to 'L.1758'". The original definition of Eucalyptus is not alterred as this is a later relationship to it, not part of, its definition. In a circumscription centric world, you might want to derive a new definition of the Eucalyptus taxon that includes all of its (in your modern opinion) member species - however - you could just treat this as a 'run-time' taxon - ie just the result of the query "get me anything that any one has ever included in L.1758's concept of Eucalyptus" - alternatively, a Taxonomist as part of a revision might wish to create a new concept/definition for Eucalyptus specifically including its member species and now being according to 'another new guy'. However, apparently a taxonomist working at one level typically does not want, or feel comfortable with making new taxon definitions outwith his expertise. A new concept definition for Eucalyptus could point at the earlier one ('includes E. L.1758') and add the new information ('has child' newspecies Ytow 2004)

A search terms query of Eucalyptus at IPNI site returns 3707 entries, so we would have the same name details of Eucalyptus 3707 times because each new species, new infrageneric classification change TaxonConcept (3707 came from both APNI and IK so it contains redundant records, but it can happen if we get TCS data through such as GBIF TaxonCoceputPortal. It would be overestimation because a pulblication may add two or more new species at once.). I don't think it is good way for data exchange in sense of bandwidth.

TrevorPaterson Notes: I dont see what problem that you are alluding to - the search returns and details all the concepts that have Eucalyptus in their name - it only becomes redundant if your query searches the tree backwards from each of these reuslts to get parent details and make a tree etc, and addition of some simple logic rules can control sending out duplicate records for the trees returned. Admittedly each species record in the set would have the same Genus name recorded (<firstEpithet>Eucalyptus</firstEpithet>), and point (possibly indirectly) to the same parent (<Relationship relationshipType="is child of"><ToTaxonConcept? ref="xyz"/></Relationship>)- but relative to general bandwidth horrors of uncompressed XML is that a worry....?

Isn't it better to provide an element for name strings (not name here) just under root node, and to replance Name element of TaxonConcept with a reference to the name string element? It also resolves the ambiregnal issue, the multiple valid names in single Code issue (e.g. Fungi), the vanacular names issue, and the stranges term "vernacular concept" or "vernacular taxon". It may be necessary to retain NameSimple as part of TaxonConcept even if we make the above modification. -- JamesYtow, 20 Oct. 2004/ammend. 02 Nov