untitled

(ff) #1

12.3 Ontology Development Tools 289


selves as “ontology development tools.” They are more commonly called
“ontology editors.” This is unfortunate because it fails to distinguish be-
tween the development of the ontology itself and the creation of the know-
ledge base. In bioinformatics, the data being stored in the knowledge base
are often obtained by automated means, either directly or by means of a
transformation (as shown in chapters 10 and 11). Semantic Web ontolo-
gies, however, often presume that the knowledge base will be manually con-
structed. Indeed, SW data are often referred to as “annotations.” As a result,
Semantic Web ontology editors have focused on the annotation task rather
than the ontology development task. Nevertheless, many editing tools (es-
pecially XML editors) do include the ability to develop the ontology as well
as the documents.
The following is a rough classification of the approaches and tools that can
be used for ontology development:



  1. No explicit ontology.One way to develop an ontology is to write the data
    in XML using a text editor or a specialized XML editor, without any DTD.
    XML can be parsed and processed without a DTD. This technique works
    only when the data have a relatively simple structure and the data will be
    used by a small number of people and processes.

  2. DTD generator.A DTD can be automatically constructed from XML files
    using a DTD generator program. Some XML editors (such as XML Spy)
    include DTD generation as one of their features. There are also stand-
    alone DTD generators. This kind of tool allows one to transition from
    using no DTD to using a generated DTD. It is not necessary to obtain and
    install a DTD generator as there are online services that will perform the
    generation for you, such as the Hit Software website (Software 2004). This
    site also has services for generating XML schemas and converting DTDs
    to XML schemas. If the BioML document in figure 1.3 is processed by this
    website, one obtains a DTD that looks like this:


<!ELEMENT bioml ( organism ) >

<!ELEMENT chromosome ( locus ) >
<!ATTLIST chromosome name CDATA #REQUIRED >
<!ATTLIST chromosome number NMTOKEN #REQUIRED >

<!ELEMENT db_entry EMPTY >
<!ATTLIST db_entry entry NMTOKEN #REQUIRED >
Free download pdf