-
Recognize that a richly descriptive, client agnostic format embedding or linking core content should be the “canonical” description of on-line work, rather than html files which entangle a particular set of very contingent UI choices. Build on the simplicity and existing role of RSS, extended as it is designed to be extended, to serve as that format.
-
Define mix-or-match extensions to RSS that provide metadata across a variety of dimensions, so that these “canonical expressions” can be deployed, managed, aggregated, and presented in a wide variety of useful ways.
“Dimensions” might include:
- An application dimension: Podcasts, and Apple’s extensions of RSS to support them, are a touchstone example. Widely distributed, but aggregated and indexed, RSS-based content should be suitable to undergird a new commonwealth of letters (like the old blogosphere), microblogging and social media, and review applications (Yelp, Trip Advisor, GoodReads, Amazon product reviews, etc) while leaving ownership and control over content in widely dispersed hands. As with podcasts, feeds can include descriptions or recommendations of UI elements that rich, application-specific clients may choose to render while leaving the core data in basic (or near basic RSS), comprehensible to anything and everything.
- An attestation dimension: It should be possible to attach digital signatures associated with identities to elements of RSS, in a context where the validity of the association between signing key and identity can be checked by clients and repudiated by identity owners in the event of compromise. 1
- A degree of curation dimension: An RSS feed can represent everything from a single atomic “item” of content to a full listing of the source it represents. The degree of curation should be representable in machine-readable metadata. Dynamic services should provide information about the request or query that would have generated this curation and the source to which that request or query might be made, so that clients and users might vary or further refine those queries.
- A degree of completion dimension: An RSS feed can signal
- as little as existence
- a resource exists, here’s a link
- minimal metadata (title, authors, date) and optionally a short description or summary
- full content, by embedding or enclosure element
- everything a rich client would require to fully represent the item, as long as it has access to the internet to collect linked subsidiary resources (images, etc.)
- complete embedding
- the full content and subsidiary resources are fully embedded in the file. the file can serve as a stand-alone archive of the items it contains.
- as little as existence
When the degree of curation dimension is full-listing and the degree of completion element is complete-embedding, RSS can serve as a permanently readable backup and storage archive for collections of creative work.
-
Design and specify a kind of RSS service, which represents a dynamic repository of creative work and metadata representable as RSS feeds. Define standard (and very simple!) means of adding to, updating, querying, and archiving such repositories.
- In such services, base feeds (as opposed to representations that are curations) should have owners and should be exportable as complete archival feeds. They should be transportable, as all links should be through owner-specific DNS names. Reimportation into a new service and redirection of DNS to that service migrates an author without breaking links.
Footnotes
-
“PKI” — public key infrastructure — is in general a domain sorely in need of protocol improvement. There exists an XML signature standard that could be deployed for attestations in RSS, but it is not popular or straightforward among developers, and relying upon it might tether RSS to aspects of XML that in the long run we might wish what is logically representable as RSS to be portable from. The question of whether to use XML signature remains worth discussing and investigating. ↩