Assertion is already used by Rich Pyle in his publication under different meaning as he pointed out: see http://www.ecoinformatics.org/pipermail/seek-taxon/2004-August/000301.html . I suggest to use Annotation, because it is used to call a data structure equivalent to Assertion sensu TCS in Nomencurator paper. No, not advertisement. Only a recommendation to re-use terms same to published terms for the equivalent structure, mainly from my experience in mapping between data models, such as http://www.nomencurator.org/TDWG.html presented in past TDWG meeting. -- JamesYtow, 13 Aug. 2004 (UTC)

We appreciate that there is the possibility of confusion and consider having a different element name. Can anyone suggest alternatives that emphasise the fact that these relationships are created/stated at a later stage (and by someone else) than the relationships that are part of the concept definition. --RobertKukla 13/8/04

Doesn't annotation have such emphasis, or still insufficient? -- JamesYtow, 13 Aug. 2004 (UTC)

Annotation to me implies that the original data is being commented on - ie not suitable to mean a third party opinion. TrevorPaterson 13 Aug

I didn't pick this your sense from OED at least. How my annotation on say Wittegenstein can't be a third party opinion, even though he died prior to my birth? -- JamesYtow, 19 Aug. 2004

Annotation is generally used to mean adding comments or explanatory notes to something. Asserting a relationship between two concepts is not 'commenting on' those concepts, but creating something new. Therefore annotation does not seem to me to be a strong enough word.... TrevorPaterson 19 Aug

There are revison works comparing views in previous publications and support one of them. Are this kind of relationships, annotation rather than assertion becuase no creation, to be ignored by TCS? -- JamesYtow, 19 Aug. 2004

If the revision work really didnt create anything new - just said "I agree with everything author Bloggs said" .. I dont think this is a new assertion or concept.
However, a revision is more likely to 're-use' some concepts from one author and some from another - making a new taxonomy according to the revision author. If for example Smith accepted genus Aus as defined by Jones and genus Bus as defined by Evans - but Smith put them together in his new familly Novasmithiae, Novasmithiae would be a new concept which as part of its definition included the concepts of Aus and Bus as previously defined by Jones and Evans. This would be relatively simple if GUIDs were being used, and would be a possible mechanism for integrating/aggregating the work of many different authors into one big taxonomy. TrevorPaterson Aug 19

(It might be better to create new page) Even though Smith stated that she/he accepted Aus by Jones, Aus sensu Smith can be different from Aus sensu Jones for Smith's misinterpretation of Jones. Later revision work on Novasimithiae, say by White, may find this difference. To whom should Aus sensu Smith be attributed in this case? Unrealised creator Smith or finder White? How should it be recorded/expressed in TCS? Nomencurator uses triplet to capture this situation, i.e. (Aus, pointer to Aus entry sec. Smith, sec. White) for Aus sensu Smith in White, and (Aus, pointer to Aus entry sec. Jones, sec. Smith) for Aus in Smith. The original Aus by Jones is represented as (Aus, pointer to itself, sec. Jones). The middle element of the triplet, i.e. the pointer to the other TaxonConcept, could be mapped to Relationship in TCS. Is it right understanding/mapping? Do these type entires pollute TCS? Should relationship between Aus in Smith and Aus in Jones be affected by existance of White's work? JamesYtow, 19 Aug. 2004

It is a basic premise of the model that once a concept C1 is created by author A1 - it can't be modified by a later author A2. If a author A2 asserts something in relation to the prexisting concept C1, or reuses it in his taxonomy, it remains C1. If A2 wants to alter C1 - he can't - he makes a new concept C2, which can have a relaltionship to C1 in its definition. If author A3 then thinks A2 has made a mistake - whatever that means....A3 can still use the original C1 concept, or he can use C2 as defined by A2. You seem to be want a later author A3 to assert that A2 had made a mistake and shouldnt have used C1 but really meant to make a new concept C2, but then C1 wants to use this 'non-existent' C2 and reference it according to A2. I dont think this is possible, A3 should make the concept as he sees it , C3, not guess what other people might or should have done. I think it gets difficult to 'correct' previous peoples work, or suggest that they made a new concept 'subconciously' and then represent this. This seems to be getting into the realm of controversial taxonomy - and maybe it is not appropriate for a general transfer schema for concepts to cope with every taxonomic argument - author A3 can still talk about any existing concepts as defined without chains of 'accordingtos'. A clearer model of taxonomy with immutable and shareable concepts should facilitate a less ambiguity about what is meant by a concept - it is exactly as defined by the originating author. Again, we are not taxonomists so might not be able to represent their views fully, maybe we can represent the views of the users of taxonomic concepts who need to have a clearer model and exchange schema for resolving the ambiguities and concflicts created by taxonomists! --TrevorPaterson 20 Aug

I agree with you that later author can't affect on previous concept itself; it is also an assumption in our model. In this sense we shouldn't create Aus sensu Smith as part of White's work. I agree that we should record C3 for A3 as you say. However, if we record C3 in this way, then why we don't create record C2 for A2? If we exchange name usage instead of taxon concept, we should record C2, whatever it is becuase the name is used by A2 with some concept. I'd like to have an exchange schema, not only scrutinised taxon concept itself but also facts which showing the taxon concept was used. NameUsage, instead of taxon concept, enables us to exchange these facts. Of course we need another way to record wheter it is explicitly sated as new concept, via by either a flag or pointer to itself as origine of the concept. Current (rather psychological) model behind TCS requires pre-scrutinise to data providers. It is too hard for some taxa including protozoa. I think broder coverage is better for purpose of GBIF at least, without reducing quality already available. I mean broder coverage with choice of quality. I'd like to discuss on it in different page as part of mapping issue. See TaxonConceptMapping. -- JamesYtow 21 Aug. 2004

I am not sure annotation captures the secondary nature of information. It could be annotation (= free-form text) by the creator of the definition. Following the use in xml Schema, UBIF/SDD use "Annotation" for free-form text content developer comments in the data. In UBIF it is part of the EnablingGroup? used for design extensions. At the moment UBIF could be changed, but I think the XML schema use is a good precedence. -- GregorHagedorn, 13. August 2004

Suggested alternative: I believe "Interpretation" is a good general concept to express secondary opinion. We use it in the DiversityWorkbench information models for this purpose. At the moment it seems adequate for what I believe is expected to be expressed in your Assertions. -- GregorHagedorn, 13. August 2004

Interpretation sounds also suitable. The issue is that the word Annotation is already used as term to represent the structure in the TCS, and whether we need to introdude another teminology to represent the same thing in similer context, i.e. taxon-name scope. Isn't it better to use Interpretation for UBIF, in sense of preoccupation? -- JamesYtows?, 13 Aug. 2004 (UTC)

Should we start a list of suggestions here?

We originally changed the name from Relationships to Assertions in order to distinguish the usage from that in the Taxon Concept definition - but maybe they can have the same name or maybe we should rename those Relationships now to something else. TrevorPaterson 13 Aug

JY adds

I don't see any reason to avoid Annotation because it is used to represent the same structure. I'd like to be conservative, or, I'm too lazy to introduce new name. -- JamesYtow, 13 Aug. 2004 (UTC) ammend. 15 Aug. 2004 (UTC)

To clarify again - within the current version of the schema there are two places where relationships between concepts can be recorded. The first is as a sub-element of the TaxonConcept element. Relationships are provided for here to allow us to represent the fact the some authors when defining a concept will include relationships from their concept to other concepts and therefore are partly defining their concept using these relationships. The second place that relationships between concepts can be recorded is in what we currently call Assertions and which seems to have caused concern amongst some people. As we said we are not bothered about these labels but more concerned with the content and semantics so that any user can represent their deinition of concepts. We chose assertion as the label because they are simply assertions representing some relationship between 2 existing concepts - i.e. relationships which were not stated during the process of defining the concept, but recorded at some later date. Therefore another proposal for the label for this element could be:

--JessieKennedy 19/8/2004

I'm not confident wheter I understand the distinction between Assertion and Relationship. Suppose publications P1, P2, P3 published in this squence. If a relationshp in P3 links TaxonConcept in P3 and another in other than but previous to P3 then it is an Relationship of the TaxonConcept, while it links TaxonConcepts in P1 and P2 but never in P3 then it is an Assertion. Is it right understanding? -- JamesYtow 19 Aug. 2004

Reply: I think that you do understand correctly ( using the labels Relationship and Assertion as in v0.70).
The definition of concepts in P3 can include relationships with those already published in P1 and or P2.
However, any relationships in P3 talking only about prexisting concepts in P1 and P2, i.e. not part of a concept in P3, are represented as 'Assertions' ( According to the author of P3).
It is this label 'Assertions'/'Assertion' that we are now proposing to rename as AssertedRelationships or whatever --TrevorPaterson 20 Aug

Thanks for clarification. My concern is in misinterpretaiton of author of P3 on P1 and/or P2. I'm not a taxonomist so I can't judge whthere Px made a misinterpretation. In other words, I can be free from any taxon concept becaues I'm ignorant. Paradoxical? But it is what a system can do. -- JamesYtow 20 Aug. 2004

in reply to above comments of JY...
This part of the schema is for taxon concept to taxon concept relationships of any type (i.e. it is general) -the original definition of a concept can also include relationships to other concepts - i.e. first party (or primary) in that they are made by the author of the concept at the time that it is defined. So the relationships we are discussing here are third party or later or secondary......TrevorPaterson 14 Aug

If it is any type relationshps, why does the schema has TaxonConceptRelatioships as an element of TaxonConcept? It should be any type but something captured by TaxonConceptRelationships. -- JamesYtow, 15 Aug. 2004 (UTC)

I have taken the liberty to move Nico's comments to a seperate page (NicosGrouping) as they slightly get off-topic for this one. --RobertKukla 1/9/04