DC TAP meeting Thursday, March 10, 2022

# DC TAP meeting Thursday, March 10, 2022 **Zoom link:** https://us02web.zoom.us/j/85164307523?pwd=QTMybkFlSTJoUHA3cHp0NkZkU1ZZdz09 **HackMD link:** https://hackmd.io/uZe-siwkTOmKGmiaJy2X3w **Time:** 15:00 UTC ([check time](https://www.timeanddate.com/worldclock/fixedtime.html?msg=DC+TAP&iso=20220310T15&p1=%3A&ah=1)) ## Participants * Tom Baker * Phil Barker * Karen Coyle * John Huck * Nishad ## Announcements * Karen made suggested editorial changes to [Primer](/DErWH403RaWiBippMFosaw) * key/value -> key-value * added "TAP example" above examples (Phil had suggested below but I couldn't find a neat way to do that with markdown) ###### tags: `dctap meetings` ## Agenda Let's see if we can finish up the framework, and look at a suggested modification to the definition of AP for the primer. The parts of the frame work to look at are Karen's rewrite of the "vocabularies and models" section, and then the final section on application profiles. * Framework: https://hackmd.io/7RT1MstvS7yumgVwbtFm2A * Github issue: https://github.com/dcmi/dctap/issues/35 * [The Singapore Framework](https://www.dublincore.org/specifications/dublin-core/singapore-framework/) * Definition issue: https://github.com/dcmi/dctap/issues/66 ## Minutes Karen's "simplified" vocabs and models section. Tom: need to remove "standard", and "schema" Karen: domain model isn't a vocabulary. Like FRBR + BIBFRAME + BIBFRAME profiles Tom: a profile needs to profile something. It profiles the metadata. May draw on a model. Providing rules is the profile. Vocabs and domain models are ok, but not necessarily standards. The raw materials. Phil: An application profile will be based on existing vocabularies and models. Phil & Tom: a profile may supply the model - doesn't necessarily come first kc: e.g. DCAT, have a model and use other vocabularies. But there is a model they are using the vocabularies within Tom: The profile can define the model, e.g. if your profile has shapes. Model doesn't have to be written down before the profile. kc: Agree. The question is how can you create a profile if you don't have (at least in your head) and idea of the metadata model you are creating. kc: we don't want all metadata using existing vocabs to be a profile. What wouldn't be an AP? Phil: DC terms tom: we can say DCAT is build on DC terms in the same way that we can say that a house is built out of bricks. It's not build just out of bricks, but that's still true. John: a defacto profile, could exist, e.g. in XML where you can mix and match in the metadata without a schema. Why is Karen concerned? kc: Too big to bite off. We need a well-defined scope for TAP or our problem is too big. Tom: boils down to the term "uses". Domain model and vocabs DEFINE, but an AP USES. Not always a clear distinction. DCAT already uses, so it's ambiguous (Heery and Patel defintion here) phil: "combined together" is where you get the model. Their "schema" means model kc: I think their "schema" is our TAP tom: add schema to How to confuse everybody. problem is unqualified use of "schema" "... profile draws on vocabularies" tom: vocabs and data models are the raw materials of metadata (added to text) tom: difference between property and class vocab and a concept schema. concept schemes are value vocabularies. phil: concepts aren't terms (as we've defined it) descriptive terms that are properties and concept schemes (that are values?). Not good to go straight into concept schemes - kc: call it list instead of scheme? or move it? do lists belong in AP? tom: part of raw materials. phil: one type = properties, classes; another = lists, enumerations. (Phil made changes) " ... which can be used to describe the characteristics of an entity" - connects with first section. Metadata standards can have usage rules. Thinking of this as a style guide, do we need this? Do we need to state that there is documentation? There is no confusion over how documentation is spoken of. Move to first para - usage rules part of schema. kc: a profile expresses a domain model kc: let's see what we need when we define AP - in terms of documentation. kc: what do we want to say about RDF? First part of each section is general, then RDF comes later. (Tom gives suggestions, Phil types) (Then Phil comes up with language for predicates/properties) John: I've done a paper on the uses of the term "ontology" - We all want to read it! John: Daylight savings. ACTION: kc to send email with options (DONE)