Create 2022-04-17.dcap_zoom_mg.md

Author: kcoyle

meetings/2022/2022-04-17.dcap_zoom_mg.md

@@ -0,0 +1,75 @@
+# DC TAP meeting Thursday, April 14, 2022
+
+**Zoom link:** https://us02web.zoom.us/j/85164307523?pwd=QTMybkFlSTJoUHA3cHp0NkZkU1ZZdz09
+
+**HackMD link:** https://hackmd.io/oCYqvJjlSWijHKaCKCXujw
+
+**Time:** 15:00 UTC ([check time](https://www.timeanddate.com/worldclock/fixedtime.html?msg=DC+TAP&iso=20220414T15&p1=%3A&ah=1))
+
+## Participants
+* Karen
+* John
+* Nishad
+* Tom
+
+## Announcements
+[New issue](https://github.com/dcmi/dctap/issues/68) from Bert relating to tooling, and vocabulary vs. AP. (Also relates to DCAT-AP)
+
+## Agenda
+
+The main agenda item is to decide whether "shape" is a thing in the metadata, in the profile, or in both. See issue #67, but also TimBL's explanation of [shapes, forms, and footprints](https://www.w3.org/DesignIssues/Footprints.html) which are audience-centric, and Ruben's [blog post](https://ruben.verborgh.org/blog/2019/06/17/shaping-linked-data-apps/). Some questions:
+* Are these "shapes" the same as "views"? 
+* Are they in the metadata? or only in the "viewer"?
+* Do we need to address the audience for our framework sections?
+
+
+**Links**
+* What's a shape? https://github.com/dcmi/dctap/issues/67
+* 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
+
+kc: new issue by Bert - difference between AP and vocabulary. We have said that the AP is about usage and vocabulary is about defining terms. This works for RDF; may be less obvious in XML or JSON.
+
+kc: main work today is to talk about shapes. We have said that shapes are in the profile, but not clear what we want to say about what is in the data or the model. I have suggested that we say that metadata has structures; these are different in different models. The metadata has units that gather statements.
+
+john: in xml the structure is always there; RDF has no records, just an open graph. AP comes out of vocabulary side; shapes from the data side.
+
+kc: no one disagrees with talking about structure in model
+
+john: do we need to use the term structure?
+
+kc: can't think of another term that covers XML, RDF, and database storage
+
+tom: separate semantics from structure?
+
+kc: problem is the flow we have here which goes from real world -> metadata model  -> vocabulary. Shouldn't vocabulary precede model? We would need that to be able to talk about semantics.
+
+nishad: we have semantics before we encode metadata in our concepts. and profile contains semantics. (seems to be semantics throughout)
+
+kc: we haven't included semantics at all - maybe best not to. But maybe switch vocabulary and metadata sections. 
+
+john: vocabulary is specific to RDF. Metadata model is more general.
+
+kc: called "data elements" in other formats. "Attributes" in O-O. Also, best not to expand this more.
+
+Tom: "vocab data models and raw materials of metadata" - 
+
+john: that may be too RDF-specific because of vocabs. We havent discussed vocabs and what we mean by them. Brings up difference between vocab and AP.
+
+kc: at end, after vocabulary say "also called data elements and terms".
+
+kc: now: AP section. statement constraints vs constraints in the columns. Edit: shapes can be a view over the metadata, not necessarily the same structures are in metadata. e.g. sparql or db query
+
+john: main concept of profile is that it is the profile of a standard. 
+
+kc: tim b-l document is done in terms of audience. tap can be for multiple audiences. 1) can define metadata to be created 2) can define md to be extracted 3) can be used to validate md. Include "purposes" in the primer. Primer says: creation and reuse. Maybe include more examples in Primer. 
+
+John: Be clear that "node" is very RDF. 
+
+kc: A shape is a node, but may not be a node in the metadata.
+
+john: "for a node in the metadata"