Discussions around metadata standards haved tended to emphasize their urgency while undervaluing the practical applications of such methods. Sadly, standards often appear misaligned with the very practices they aim to enrich. As Manu Sporny has sharply put it, open standards too often appear to offer "esoteric solutions to non-problems".1

Internet Archive API

Essentially about understanding the relationships between files, echo of the pipeline of treatments the file(s) have undergone. Recognizable situation (and related problems) of simply managing the provenance of files / versions / etc.


snapshots / files as (change)sets

The shift from naming files index_olderone.html, process_usethisone.py to thinking about sets of files that stay together via the snapshot of a commit. With git, and the concept of a "feature branch", essentially addresses the practical problem that the "version" of a complex system can never be captured in just the name of a single file. Modern web projects and scripting involves sets of files that rely on each other. git enables snapshots of files and folders to be made rather than attempting to make naming conventions or manually copying files to duplicate folders.

Dublin core

At its core is all about relationships between resources.

xml + namespaces

The single most exciting feature of XML that I had always managed to overlook (at least based on my experiences as a programmer) is the way namespaces (that otherwise appear to needlessly complicate things) actually enable the very powerful idea of mixed documents.


Feeds and the relationship with metadata (rdf site syndication).


Annotation to practice, inject in "contextualizing" labels and standard descriptions by way of enabling standard tools chains.


Standards are about shared conventions, distillations of practice. Be they embodied in song or software, "good" standards invoke a sense of recognizability, while still providing space for innovation and reinvention.


Lateral composition (examples: file system, functional style programming (NOT object oriented))

  1. The lead solutions for RDF in JSON were things like the aptly named RDF/JSON and JTriples, both of which would look incredibly foreign to web developers and continue the narrative that the Semantic Web community creates esoteric solutions to non-problems. From JSON-LD and Why I Hate the Semantic Web http://manu.sporny.org/2014/json-ld-origins-2/