• Phylorefernecing 'test cases' - (1) Apply definitions used in Cantino et al. to later published plant phylogenies. (2) Apply phyloreferences in Campanulaceae; start with a manageable group.
• Linnaeus dissociated meaning from names through binomial nomenclature - has this been discussed in the literature?
• The use of binomial nomenclature enabled biologists to publish in their own languages, rather than Latin.
• Since when did biologists/taxonomists start having a notion that there is a natural order of organisms? Before Darwin?
• Phyloreference data models or computational tools need to be biologically meaningful and practically useful.
• Guanyang needs to articulate what uses or desirable capabilities that biologists would like to have and that we could help them realize? This will go outside the theoretical framework of phylocode. Need to do this in writing.

GY: Currently experiencing some writer's block. Concept taxonomy. At least four attempts were made to develop or implement concept taxonomy in databases - Berendshon, Zhong, Kennedy and Franz, but it just has not gained traction in the taxonomic or informatics communities. The question is naturally why, why concept taxonomy didn't get widely recognized or used. Gaurav: NF talked about the challenges of concept taxonomy. What would it take for people to use it. Killer application. Lotus123, first spreadsheet application. It is so useful that users will make investment to learn to use it. TDWG discussions often brought up this issue. For example, R. Page would also talk about why we need to integrate data and how it will facilitate science. GY: Is an killer application analogous to paradigm shift? GV: Not really. Added value of learning a tool/method needs to be obvious and substantial. For example, a pretty ggplot2 graph will easily convince a student to invest two months to learn it. GV: What's the big story [of concept taxonomy]? J. Cracraft: We do not care - if we aren't sure, we just go check the specimen. There is little doubt biologists agree that names are ambiguous, but the perception of ambiguity is not homogenous -- most would feel negative about, but some may argue ambiguity is essential and even beneficial. Reconcilable ambiguity.

Research idea: Can we model or document the history of taxonomic changes ever since Linnaeus? Look up Sangster (2009, 2011) who wrote about taxonomic change and found that diagnosability is what taxonomists really cared about.

Thought experiment: If we were to pitch concept taxonomy to Carl Linnaeus, how would he think of it at his time? How about phyloreferencing (despite that Linnaeus didn't know about phylogenetics). A corollary question that came from the thought experiment was that what kinds of taxonomic changes existed at Linnaeus' time? Name change (e.g., generic transfer)?

Phyloreference is a machine-readable formulation of a phylogenetic clade definition. In its simplest form, it's just a set of specifiers (that define a phylogenetic clade).

Curation tool. Open Tree curation tool is a good example to follow. Should be fairly intuitive to somebody who does not understand what a phyloreference is.

Guanyang feels that relying a human to verify if a computer has correctly resolved a node is problematic. How do we know if the human has done the verification correctly? Gaurav explained that the verification is only needed to make sure the code works and will not be part of the actual curation workflow.

Gaurav says curation tool is best for small or hypothetical phylogenies.

A tool that facilitates reconciliation of taxonomic knowledge and phylogenetic trees/insights is what taxonomists really need.