- tdwg-content - lists.tdwg.org

Re: (RQT) Character and item hierarchy
by P.Hovenkamp 02 Dec '99

02 Dec '99

At 04:26 PM 01-12-99 +0100, Gregor Hagedorn wrote: >> There is a containment hierarchy (e. g. a flower contains petals) >which >> raises no problems in XML. > >I would like to challenge this assumption, which seems to be >generally accepted in the discussion here. (...) >Thus: there are multiple concurrent or competing hierarchies, which >may overlap. (...) >*** Call for more examples: Can anybody come up with more good >examples (perhaps from animals) as to why a fixed hierarchy in the >form of a feature path may be unpractical? Here's my example: In ferns, leaves are often highly divided, into pinnae, pinnules, segments and even lower order elements. In most descriptions, a clear hierarchy is implied: pinnae are part of leaf, and inherit the properties, pinnules are part of pinnae and inherit... Programmers may be tempted to hard-code this hierarchy. However, for many characters, an opposite hierarchy can be devised, which may be much more informative: ultimate segments have margin, veins, and properties of those, which are inherited in the penultimate elements, and then onwards to the leaf. Especially in species where degree of dissection varies strongly (they occur), this may be a much better way of descibing a fern leaf. Hard-coding the usual hierarchy of leaf, pinnae, pinnules may therefore stand in the way of understanding and describing fern morphology. Is that clear enough for non-fern people? Best wishes, Peter P. Hovenkamp Rijksherbarium, Leiden The Netherlands hovenkamp(a)rhbcml.leidenuniv.nl http://rulrhb.leidenuniv.nl/

1 0

Re: (RQT) Character and item hierarchy
by Robert A. (Bob) Morris 02 Dec '99

02 Dec '99

P.Hovenkamp writes: >[...] > > Programmers may be tempted to hard-code this hierarchy. Programmers who hard-code /anything/ should be sent back to school for re-education. Not only should software itself behave according to the principle of least amazement, but we know that this is best accomplished if the programmers never build in assumptions about how the clients will use the software or present the data. The biggest advantage of XML for this group is its impending ubiquity, not its formal structure, and (in my opinion) the discussion here should return to discussing problems, not solutions because that ubiquity will make the community want to use it whether it is the best fit or not (just as has happened with HTML as a presentation vehicle). I am reminded of a talk I heard from a high-level member of the technical staff at Adobe a few years ago about an aborted three-year language development project. By his cleverly crafted presentation it became clearer and clearer to the audience that this new language had exactly the design features of another, just recently released language, he finally ended his talk by saying: "As we prepared to take the cloak off this project, we stepped into the street, looked up and saw that we were in the path of a huge, oncoming train. Its name was Java. We scrapped the project."

1 0

Re: (XML)Vocabularies getting out of hand
by Stuart G. Poss 01 Dec '99

01 Dec '99

: > Jean-Marc writes: > > >Just say for the moment that we'll have several vocabularies by large > >domains: > To which Kevin Thiele responds: > > Getting messy - I can't imagine that ... > I think this whole idea of a universalist vocabulary (even several > universes) is becoming silly. I believe the point was that we need to plan for a mechanism to handle consistently and with reasonable efficiency a wide variety of vocabularies in an extensible format, not that we would actually mandate any universal format. Any recommendation once implemented will, however, necessarily require us to forego some possible schemes. Lets just be sure that useful ideas are not lost as we collectively thrash things out. > Leigh comments: > > Even a core set for mammals, insects, nematodes, sponges... would probably > be a null set, except for cellular characters. > The fact that not every pair of taxa can be compared for a particular basis of comparison doesn't mean establishing a number of "vocabularies/name spaces" for particular levels of generality is a useless approach. Considering that such a common set or "core or intersection" might actually be several thousands of "protein" characters alone might suggest that this might be more fruitful than one might first suspect. I would agree, however, that it will likely take a potentially large number of "namespaces" to tease out all the interesting/potentially informative features in comparing widely disparate taxa, especially since one would like to characterize features along more than one basis of comparison (eg biochemical as well as morphological or ecological). We need not settle on a "final solution" as to name useage here, but only develop a mechanism that can be modified through time, according to conventional useage, so as to encompass a wide array of complex associations, some not yet discovered. If I have at least partially correctly digested the references on the RDF syntax correctly, we seem to need a mechanism to handle a wide variety of subject,predicate,object relations relating to both taxa and characters. The extent of these statments will likely arrange themselves according to the context (taxa) in question, at least to a large degree. Consequently it seems to me that we need some mechanism to specify taxa as a qualifier, or at least the "taxonomic range" (extent) over which the basis of comparison (character) implies. Likewise, we also need some mechanism (set of name spaces?) for communicating the basis(bases) of comparison implict for a given character (ie shape, molecular structure, fine structure, gross morphology, scale of comparison, physiology, etc.). >>>From my perspective, for statements of the form Taxon "A" HAS character "X", it would seem we need distinct "name spaces" (if my present conception of this term is the one being used by XML programmers) for both the taxa (subject) and character (object) part of the dialog. It remains unclear to me if this implies whether we may also need a potentially separate set of "name-spaces" for the reverse mappings (of the form character "X" can be found in/identified in/is representative of Taxon "A"). My suspicion is that we do, since such relations will usually not be isomorphic. However, this may depend upon the discipline (context). Likewise, the nature of the qualifiers for the predicates themselves (such as "is found in", "has", "has many of", "is a protein of", "has the shape of", "is a synapomorphy for" etc.) are even more unclear to me, as is whether there are any universal or common rules regarding such predicates that might usefully constrain the ways the taxon and character "vocabularies" are defined and used as an XML "grammar" potentially emerges. Also, it is also not yet clear to me how we can use XML to construct more complicated grammars implicit in the higher order grammatical relations of the type "Taxon "A" has character "X" according to Investigator(s) "Y" using observational method "Z" at time "T" (for rapidly evolving retroviruses or seaonal breeding colors, flowering condition. or molecular pathways perhaps). With respect to "messy" names-spaces/vocabularies I suggest we let the computers handle the large lists, multiple spellings, homonymy of terms, etc. They are particularly good at it and we will need a great many to characterize the millions of taxa, billions of molecules, and innumerable different bases for character definitions. I believe we need to concern ourselves more with how a consistent grammar for a "metadata language about characters and taxa" can be constructed/(implemented on the WWW) to facilitate consistent human input and precise meaning. Which specific character strings we ultimately use to formulate the tags or namespaces might take a bit longer to resolve, particularly should foreign colleagues feel strongly that they should be given unicode rather than soley ascii representations. Stuart

1 0

(RQT) inheritance of properties
by Gregor Hagedorn 01 Dec '99

01 Dec '99

> But there is an inheritance hierarchy that I don't know how to > handle using W3C's specifications alone. A genus has properties > values (e. g.flower/petals/color = 'white' ) that MUST be inherited > by the species in this genus. The properties values of a genus CAN > be overriden by some species in this genus. I have no insight on the way it would be handled in XML, but I believe we must consider that (a) inheritance goes up and down, and (b) information may be original or virtual. The genus description is nothing more than the consensus description of all species placed in that genus. This is true regardless, whether it is done by computer or a scientist. The type species (the species to which the genus name is bound) will carry more wheight, some people may wheigh it very strongly. Thus currently placed in GenusX Species1=101 \ Species2=10U > synthesized ("summary function") description of GenusX Species3=10U / results in a description GenusX=[101) (1 = a state applies, 0 = does not apply, U = unknown, not yet observed) Some information, e.g. anatomical or electron microscopy is usually only obtained from one or few species and then assumed to be true for all species in the Genus. Thus, information that was passed up, is now passed down again to species where this has not been studied. Care must be taken that the information is recognized as being passed down, it can NOT be simply copied as if it would indeed have been studied there. Species1=10 1 Species2=10[1] Species3=10[1] Square brackets shall here indicate that the information has been passed down. It will often happen that Species3 is transferred to a different Genus, because the assumption that it belongs to GenusX turned out to be false. The GenusX description must now be revised, and information that was passed up from Species3 to GenusX and down to Species1/2 must be removed. This is a continous dynamic process in science, and must be modeled on the computer. However, the Genus description is not necessarily entirely dynamic. In many cases it will be necessary to copy information from publication, thus the Genus description may be physical and original data, in contrast to the virtual descriptions obtained by automatic synthesis processes. Gregor ---------------------------------------------------------- Gregor Hagedorn G.Hagedorn(a)bba.de Institute for Plant Virology, Microbiology, and Biosafety Federal Research Center for Agriculture and Forestry (BBA) Koenigin-Luise-Str. 19 Tel: +49-30-8304-2220 14195 Berlin, Germany Fax: +49-30-8304-2203 Often wrong but never in doubt!

1 0

(RQT) Character and item hierarchy
by Gregor Hagedorn 01 Dec '99

01 Dec '99

Jean-Marc Vanel wrote under the Subject: RDF Schema design pb: inheritance of properties (Subject renamed, because this post refers to issues secondary to the original post) > There is a containment hierarchy (e. g. a flower contains petals) which > raises no problems in XML. I would like to challenge this assumption, which seems to be generally accepted in the discussion here. It is true: Morphological structures may have containment hierarchies, but I believe that these depent strongly on the viewpoint of the author or user. EXAMPLE 1: The peduncle (stalk of a flower) is usually assumed to be part of the flower, but anatomically it is clearly a part of the stem (in contrast to the petiole, stalk of a leaf!). EXAMPLE 2: Stuff can be in-between: The inflorescence contains part of stem, part of leaves, and all flowers. Which leaves are part of inflorescens and thus called bracts, and which aren't is often a matter of taste, school, country... Thus: there are multiple concurrent or competing hierarchies, which may overlap. EXAMPLE 3: Further, other, non-morphological hierarchies/classifications exist. These may be much more relevant for many purposes, and may be the primary outline used in a natural language description. For example, structures can be classified by function (sexual/asexual propagation), observation method (naked eye/light microsc./electron microscopy, physiological/molecular etc. EXAMPLE 3 (leaving morphology altogether): Gene sequences can be classified as transscribed/non-transscribed, Protein-coding/rRNA coding/non-coding, intron/exon, regulatory/structural, conserved/variable/hypervariable etc. Most of these classifications overlap. *** Call for more examples: Can anybody come up with more good examples (perhaps from animals) as to why a fixed hierarchy in the form of a feature path may be unpractical? > Info. for computer men: biology has a hierarchic classification of > species whose 4 lowest levels are: > order > family > genus > species Sorry this is a gross misrepresentation. Info from a biologist: (incomplete list, reflecting botanical code, and devised with fungi in mind). Number of ranks is a matter of taste, some people like to supra and sub [and hyper+hypo? :-) ] everything: CODE Description fsp. forma specialis nov. -- nomenclatorially facultative subfm. subforma nov. -- subform nomenclatorially obligatory fm. forma nov. -- form nomenclatorially obligatory subvar. subvarietas nov. -- sub-variety nomenclat. obligatory var. varietas nov. variety nomenclatorially obligatory ssp. subspecies nov. subspecies priority mandatory sp. species nov. -- species priority mandatory subser. subseries nov. -- subseries priority mandatory ser. series nov. -- section priority mandatory subsect. subsectio nov. -- subsection priority mandatory sect. sectio nov. -- section priority mandatory subgen. subgenus nov. -- subgenus priority mandatory gen. genus nov. -- genus priority mandatory subtrib. subtribus nov. -- subtribus priority mandatory trib. tribus nov. -- tribus priority mandatory subfam. subfamilia nov. -- subfamily priority mandatory fam. familia nov. -- family priority mandatory superfam. superfamilia nov. -- super-family priority mandatory subord. subordo nov. -- suborder priority facultative ord. ordo nov. -- order priority facultative superord. superordo nov. -- superorder priority facultative subclass. subclassis nov. -- subclass priority facultative class. classis nov. -- class priority facultative subdiv. subdivisio nov. -- subdivision priority facultative div. divisio nov. -- division priority facultative subreg. subregnum novum. -- sub-regnum priority facultative reg. regnum novum. -- regnum priority facultative Gregor ---------------------------------------------------------- Gregor Hagedorn G.Hagedorn(a)bba.de Institute for Plant Virology, Microbiology, and Biosafety Federal Research Center for Agriculture and Forestry (BBA) Koenigin-Luise-Str. 19 Tel: +49-30-8304-2220 14195 Berlin, Germany Fax: +49-30-8304-2203 Often wrong but never in doubt!

1 0

Re: RDF Schema design pb: inheritance of properties (XML)
by Leigh Dodds 01 Dec '99

01 Dec '99

> But there is an inheritance hierarchy that I don't know how to handle > using W3C's specifications alone. A genus has properties values > (e. g. flower/petals/color = 'white' ) that MUST be inherited by the > species in this genus. The properties values of a genus CAN be > overriden by some species in this genus. I believe RDF handles this, see below. > It seems that the "refines" property in XML Schema would bring the > desired semantics, together with the "Schema Information Set > Contribution: augmentations to instance information sets which follow as > a consequence of schema-validation". But the uncomplete state of the W3C > document makes me unshure. Hmm, I'm not 100% sure either, but my feeling is that RDF is better suited to this task than Schema. As I understand things - we're looking to produce a generic schema for capturing character data. In this case, "refines" could be used to extend the schema to include attributes/elements from other disciplines. RDF allows you to describe 'resources' independent of the structure of the schema they conform to, and allows things like sub-classing. > I don't see anything clearly specified in RDF and RDF Schema for > inheritance of properties. >>>From the RDF Schema spec : http://www.w3.org/TR/PR-rdf-schema/ "Classes themselves are often organised in a hierarchical fashion, for example the class 'dog' might be considered a sub-class of 'animal' which is a sub-class of 'organism' etc., meaning that any resource which is of rdf:type 'dog' is ipso facto also of rdf:type 'animal' and so on. This specification describes a property, rdfs:subClassOf, to denote such relationships between classes." ...and (describing subClassOf)... "If class A is a sub-class of some broader class B, and B is a sub-class of C, then A is also implicitly a sub-class of C. Consequently, resources that are instances of class A will also be instances of C, since A is a sub-set of both B and C" ...and (describing subPropertyOf)... "If the property biologicalFather is a sub-property of the broader property biologicalParent, and if Fred is the biologicalFather of John, then it is implied that Fred is also the biologicalParent of John." I think these portions of the spec apply to the sort of 'inheritance' you require - or am I missing a subtle difference here? If you wish to provide default values for properties then I think this needs to be addressed at an application level. > Info. for computer men: biology has a hierarchic classification of > species whose 4 lowest levels are: order family genus species Heres an RDF example of this structure (I've kept it to the lower 4 levels of the classification, ignoring Kingdom, Phylum, Class, etc.). <rdf:Description ID="Order"> <rdf:type resource="http://www.w3.org/TR/1999/PR-rdf-schema-19990303#Class"/> <rdfs:subClassOf rdf:resource="http://www.foo.com/taxonomy-schema#Class"/> </rdf:Description> <rdf:Description ID="Family"> <rdf:type resource="http://www.w3.org/TR/1999/PR-rdf-schema-19990303#Class"/> <rdfs:subClassOf rdf:resource="#Order"/> </rdf:Description> <rdf:Description ID="Genus"> <rdf:type resource="http://www.w3.org/TR/1999/PR-rdf-schema-19990303#Class"/> <rdfs:subClassOf rdf:resource="#Family"/> </rdf:Description> <rdf:Description ID="Species"> <rdf:type resource="http://www.w3.org/TR/1999/PR-rdf-schema-19990303#Class"/> <rdfs:subClassOf rdf:resource="#Genus"/> </rdf:Description> Here we can declare Chrysanthemum to be a subclass of species, forgive me for not knowing the latin name... <rdf:Description ID="Chrysanthemum"> <rdf:type resource="http://www.w3.org/TR/1999/PR-rdf-schema-19990303#Class"/> <rdfs:subClassOf rdf:resource="#Species" /> </rdf:Description> However, and this is where I'm still wrestling with RDF to an extent, there are other ways to achieve this. For example, we may not need Genus and Species classes - instead modelling the Chrysanthemum genus as a class, and species as subclasses. Note that RDF does allow a resource to be a subclass of more than one class. I can see a lot of work ahead in this area, and there seems to be a wealth of different ways to model the same ideas. I'm at present digging through the literature to come up with some decent examples. (I hope the above was vaguely useful). Cheers, L.

1 0

RDF Schema design pb: inheritance of properties
by Jean-Marc Vanel 30 Nov '99

30 Nov '99

Hello We are designing XML vocabularies for the description of biological species. See : http://jmvanel.free.fr and Taxonomic Databases Working Group - Structure of Descriptive Data <TDWG-SDD(a)USOBI.ORG> (mailing list) archive available at http://usobi.org/archives/tdwg-sdd.html This is a large project: 250 000 species of plants, 1 000 000 of insects, relations beetwen plants and insects, etc... There is a containment hierarchy (e. g. a flower contains petals) which raises no problems in XML. But there is an inheritance hierarchy that I don't know how to handle using W3C's specifications alone. A genus has properties values (e. g. flower/petals/color = 'white' ) that MUST be inherited by the species in this genus. The properties values of a genus CAN be overriden by some species in this genus. It seems that the "refines" property in XML Schema would bring the desired semantics, together with the "Schema Information Set Contribution: augmentations to instance information sets which follow as a consequence of schema-validation". But the uncomplete state of the W3C document makes me unshure. Although RDF and RDF Schema's data model are very promising, allowing to : - make links between plants and insects, - state properties about properties (some are morphological, others are chemical, etc ...) , I don't see anything clearly specified in RDF and RDF Schema for inheritance of properties. Would this be a solution: - each instance of genus is a rdfs:Class, e.g. all daisies belong to (are rdfs:subClassOf) the genus Chrysantemum; - a genus has Class properties, - a species, instance of species, and subClassOf a genus, redefines some of the genus' properties, overrides the other properties, and adds properties unspecified by the genus. Sample RDFS: <rdfs:Class ID="Chrysantemum"/> <rdfs:subClassOf rdf:resource="http:botany.org/taxonomy#genus"/> <flower><petals> <color>white</color> <number>10--40</number> </petals></flower> </rdfs:Class> <species> <rdfs:subClassOf rdf:resource="Chrysantemum" /> <name>maximum</name> <flower><petals><color>yellow</color>  </petals></flower> </species> Info. for computer men: biology has a hierarchic classification of species whose 4 lowest levels are: order family genus species

1 0

Re: Item description data in XML (XML)
by Leigh Dodds 30 Nov '99

30 Nov '99

> > General question: Is there any reason why attributes exist at all? Robin Cover covers some of the issues here: http://www.oasis-open.org/cover/elementsAndAttrs.html This debate is also currently raging on the XML-DEV mailing list. There are proponents for both sides of the debate, which can get quite heated! If you care to dip in you can look at the thread here: http://www.lists.ic.ac.uk/hypermail/xml-dev/xml-dev-Nov-1999/0998.html There are strong arguments on either side - even the 'attributes are required for linking' viewpoint isn't firm, as <leaf shape="obovate" /> <leaf> <shape>obovate</shape> </leaf> are semantically identical. It largely depends on which you prefer, and how your schema is defined. The id/idref linking mechanism only works within documents. Its unlikely that this will apply in this context as links will be likely *between* documents - and we'll therefore be using XPointers (which are based on XPath). Its up to the application to decide whether it looks for a shape attribute or element. Cheers, L.

1 0

Re: Markup of text descriptions Vs. structured data (GEN)
by Gregor Hagedorn 30 Nov '99

30 Nov '99

Jim Croft wrote: > Can you explain in more detail why they should be fundamentally different? > Aren't the differences just a matter a matter of degree, different points > on a continuum of data as it were? And aren't the basic principles > applicable across the whole? >>The issues overlap, and it would be beneficial to use as much common >>syntax as possible, but fundamentally I believe them to be quite >>different. I completely agree with you that the issues overlap and should be treated together. However, I still see them a distinct, so that they can be treated by separate groups of commands, that use a common system. But maybe that is only to try to protect my brain from blowing up in confusion :-) Could you perhaps give examples where you see the continuum? My differences (see also the example in the other post) are: o Language in free text that is marked up may have different language, while the underlying data should be language independent o Similarly, on a more subtle level, the concepts used by one authors may only approximately map unto the concepts developed in the character definition o Data may be more detailed, which unambigously may map unto a course natural language representation o Much additional knowledge management attributes, mutliple authoring/editing/revision layers, annotation and assessment, quality control, status regarding data source and passing information up and down taxonomic hierarchy, assumptions about possible misinterpretations, etc. Much information may be used for markup as well as for data, so there will be overlap. However, for example, for a given character definition multiple independent hierarchical classifications (by morphological structure, perhaps according to different schools, by methods, by function) may be defined; the natural language definition will however usually only use a single classification. This classification is superfluous in the item description part of the data, but it is desirable to add to the markup. Or isn't it? But then we could not use the idea to just use a coarse structure-only markup as a first step to limit the scope of XML queries, an idea forwarded by Don Kirkup and Bryan Heidorn at the TDWG meeting. Gregor ---------------------------------------------------------- Gregor Hagedorn G.Hagedorn(a)bba.de Institute for Plant Virology, Microbiology, and Biosafety Federal Research Center for Agriculture and Forestry (BBA) Koenigin-Luise-Str. 19 Tel: +49-30-8304-2220 14195 Berlin, Germany Fax: +49-30-8304-2203 Often wrong but never in doubt!

1 0

Re: Item description data in XML (XML)
by Gregor Hagedorn 30 Nov '99

30 Nov '99

> My XML based application may not render the data at all - it might > just use the data to perform some other functionality - lets not > get too fixed up on the visual representation. OK, you are probably correct. > XSL is based on XSLT which is a means of taking an XML document > and producing another document from it (not necessarily XML). > XSLT currently supports three output methods - XML, plain-text > and HTML. Thanks for the insight into XSL and XSLT. I think I understand the context much better now; I was really thinking in terms of stylesheets alone. The problem is really that I believe using XML is a good idea because it allows reuse of existing tools, but then I really don't have any experience with an existing tools other than IE5, which does not do much. That in principal it is always possible (and necessary) to write custom applications for custom functionality is quite understood. However, it would be nice to have generally accepted mechanisms, that allow to use certain feature right out of the box. When I was asking about how to manage free text with markup and data _efficiently_, I was thinking about how to use standard tools (which don't exist that much yet) to edit the different views of the data, much more than thinking about reporting/screen display with a stylesheet. --- Also, I think I did not do a good job in explaining where I see the problems (which may be due to my lack of understanding XML). Lets go back to examples. Markup of the text: leaves usually obovate (except in young leaves), 3 cm long could look like <leaf>leaves <shape> <frequency modifier>usually</frequency modifier> <categorical value>obovate</categorical value> <reportable note>(except in young leaves)</reportable note> </shape> , <length>3 cm<length> long </leaf> Thats already quite complicated. But how do I add the data? The data contain, e.g. an identifier for obovate (i.e. an alphanumeric code or some GUID) to identify the term obovate in relation to some character schema. Note that the character schema may use the term obovate as identifier, so it could be identical. However, German, Chinese, or Russion descriptions will have quite other text within the markup... Further the data would contain more detailed information what frequency "usually" really is, where the original information has been stored, it may contain in addition to the 3 cm information whether "3 cm" is a single measurement, a guess, or the average of 30 measurements, for each of these informations. Yet another level is the knowlegde management, with authoring, editing/revision, quality control, multiple annotations etc. Welding all this into the above structure is quite complicated, so I am looking for general guidelines, how this can be achieved efficiently. My feeling is, that attributes within elements are not the solution, but I may be wrong. In Don's example <leaf shape="obovate" length_mm="31"/> It is difficult to add information about obovate or length. Noel already noted that we may be running into this "properties of properties" problem. General question: Is there any reason why attributes exist at all? What is the advantage of <leaf shape="obovate" length_mm="31"/> over <leaf> <shape>obovate</shape> <length>31</length> <MeasurementUnit>mm</MeasurementUnit>=" </leaf> ? Gregor ---------------------------------------------------------- Inst. for Plant Virology, Microbiology, and Biosafety Federal Research Center for Agriculture and Forestry (BBA) Gregor Hagedorn Net: G.Hagedorn(a)bba.de Koenigin-Luise-Str. 19 Tel: +49-30-8304-2220 14195 Berlin, Germany Fax: +49-30-8304-2203 Often wrong but never in doubt!

1 0