Title: Taming the Graph: Profiles over Linked Data
Day/Time: Friday, 8:30-5:00
Room: Prince William (3rd Floor)

Abstract: The idea of application profile was proposed at a Dublin Core workshop in 2000 as a way to customize metadata for specific application domains, mixing-and-matching multiple metadata vocabularies as needed, and to share these customizations within communities of practice. This day-long event will start with a look back at how the discourse around profiles has evolved since 2000, then it will examine new technologies that can help us tame the boundless sea of Linked Data with controlled metadata. This event is sponsored by the DCMI, which participates in W3C efforts to develop standards relevant to profiles and seeks direction for its future work, and by LD4, a coordinated program to advance the use and utility of linked data in research and cultural heritage organizations, especially libraries.

We ask the following questions:

1. Can Linked Data conceptualization and design be enhanced with metadata profiles?
2. How can profiles provide different views: data validation vs discovery, strict vs tolerant?
3. How can we express profiles for processing by machines?
4. How can we publish profiles for human consumption?

Presenters:

Morning Session 1: The Role of Profiles (90 min)

"Application Profiles" since DC-2000

Tom Baker, DCMI

Pieces of the profile puzzle

Karen Coyle, Consultant

Profiles and Data Quality

Stefanie Rühle, SUB Goettingen

Morning Session 2: Developing and Using Profiles (90 min)

Requirements for BIBFRAME profiles

Kirk Hess, Library of Congress

Structured methods for developing profiles

Mariana Malta, Polytechnic of Oporto

Agile, data-driven methods for developing profiles

Eric Prud'hommeaux, World Wide Web (W3C)

Discussion

Afternoon Session 1: Expressing and Profiling Data (90 min)

JSON-LD: The data syntax and its uses

Gregg Kellogg, Consultant at Spec-Ops

ShEx: the Shapes Expression Language

Eric Prud'hommeaux, World Wide Web (W3C)

Afternoon Session 2: Sharing profiles (90 min)

Documenting profiles and vocabularies on the Web

Paul Walk, Edinburgh University

Maintaining RDF vocabularies in spreadsheets

Gregg Kellogg, Consultant at Spec-Ops

Answering the questions posed

Conference participants are free to attend any or all sessions or parts of sessions. The event assumes familiarity with the basic concepts of RDF and Linked Data. Participants will get more out of Afternoon Session 1 if they prepare beforehand by reading (or watching videos) about JSON-LD [1] and by perusing the ShEx Primer [2].

[1] https://json-ld.org/learn.html
[2] http://shex.io/shex-primer