Clarify COR product vision

[lewismc]

From the ESIP summer 2020 meeting notes

Our stakeholders provided the following input to guide us forward

a. Get pick list of terms from controlled vocabulary for a particular context (e.g. a data entry form. What are the terms for sandstone grain shape?) SMR (Steve Richard)

b. What vocabularies are there that are about soil texture? SMR

c. What is the definition of the term with URI X (displayed in human-readable html) SMR

d. What vocabularies are there on a certain topic and what are their advantages

e. Is there a term with label X in vocabulary Y (the validation use case) SMR

f. Vision needs greater specificity and clarity! << isn’t that a bit recursive? :-) well, you can be clear but non-specific (general), and unclear but specific

g. What vocabulary mappings have mapping from concepts with label ‘X’; what are the target vocabularies in the mapping? SMR Get different representations for a concept referenced by a URI (e.g. various OWL flavors, maybe different rdf vocabularies, html, free text, maybe an image?) SMR

h. A place for wiring up things, e.g., to connect resources, via the API

i. COR should be a hub for semantic harmonization however right now the mapping manager does not support user-defined predicates… so it’s usability and usefulness is limited.

j. Doing lookups in the current model has issues, e.g., lookup ‘flux’ and getting a lot of blank nodes. What good is that as a subject. Underlying structure of graphs may be part of issue Curation may be needed. A UI that understands the underlying resources may fix a lot of issues. UI for improving information retrieval may address this; e.g., sorting search results for nodes by label based on for example pageRank

k. Without curation we will hear increasing frustration from users (first about Syrian movies, then about not finding reasonable ontologies, then about not finding good ontologies)