DanC: "EVS encourages you to begin using NCI Thesaurus in the Web Ontology Language (OWL). "
DanC: Thesaurus.owl is 198M
DanC: excerpt: <owl:ObjectProperty rdf:about="#Allele_Plays_Role_In_Metabolism_Of_Chemical_Or_Drug">
DanC: and they're clear about full URIs; not just xml:base but also comments: <!-- http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl#Allele_Plays_Role_In_Metabolism_Of_Chemical_Or_Drug -->
DanC: Thesaurus.owl is 198M
DanC: excerpt: <owl:ObjectProperty rdf:about="#Allele_Plays_Role_In_Metabolism_Of_Chemical_Or_Drug">
DanC: and they're clear about full URIs; not just xml:base but also comments: <!-- http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl#Allele_Plays_Role_In_Metabolism_Of_Chemical_Or_Drug -->
karlcow: reminds me of http://en.wikipedia.org/wiki/AOL_search_data_scandal
History Sniffing: How YouPorn Checks What Other Porn Sites You’ve Visited and Ad Networks Test The Quality of Their Data
mischat: link to the actual paper http://cseweb.ucsd.edu/~d1jang/papers/ccs10.pdf
danbri: HTTP error codes and privacy
danbri: Try the 'graph' tab
webr3: following manu's genome open sourcing, he's now getting forked / issues etc
drewp: a story about the fallout when developers didn't make programs that can follow links inside resources
DanC: Appl Ontol. 2008 January 1; 3(3): 173190.
DanC: "Golbeck and colleagues (Golbeck et al., 2003) developed the initial translation of the NCI Thesaurus into OWL DL. "
DanC: "... representation issues that we address ... include representation of imprecise information (e.g., a possible outcome of a disease), exceptions (e.g., properties of genes that are changed in their alleles), ...
DanC: role chains (e.g., representing a link from gene to gene product and then to the disease, but being able to query the link between the gene and the diseases, bypassing the gene product); reciprocals for existential restrictions (e.g., representing that gene X plays role in a process Y by using an existential restriction on the class for gene X, but being able to query which genes play role in process Y, from the process point of view); and others.
DanC: "The class hierarchy can be subdivided into 22 main subontologies, " but the figure shows only 20. hmm.
DanC: nice: "The solution that we are proposing is to have two properties, such as Disease_has_left angle bracketXright angle bracket and Disease_May_Have_left angle bracketXright angle bracket (e.g., Disease_Has_Finding and Disease_May_Have_Finding)."
DanC: "The OWL language itself does not have role chaining" fixed in OWL 2
DanC: oh... they noted that: "OWL 1.16 also has introduced role chaining in a decidable way (Horridge et al., 2006)."