The dialog that Ed Summers initiated in response to an earlier entry on microformats has me thinking about DC as a microformat (or more accurately, as candidate content that should be microformatized), and how we struggled with this in the early days of DC.
Some of the persistent questions about syntax:
- Are attribute-value pairs in html-encoded meta-tags sufficient to do the job (and what is the job, if the answer is yes)?
- Can XML and XML schemas provide sufficient reuse-flexibility and manageability to satisfy the requirements of a generalized metadata architecture?
- Will RDF add sufficient value to overcome adoption barriers and become the Lego™-interoperability standard for metadata?
The ongoing fragmentation of the community on these issues has meant that we’re left with the requirement to support all these options, with a consequent dilution of focus (and interoperability). Theories and practices do not coincide.
Jonathan O’Donnell commented:
Isn't it a Darwinian popularity contest? Those microformats that are popular rise to the surface (and get used again and again). Those that don't ,sink without trace.
People who back the wrong microformats lose out, but it is a comparatively little loss.
A point well taken, and it may be the best we can hope for. But in the Web, as in nature, we have many niches rather than monolithic Darwinian selection, and while this is fine in any given niche, we still end up in isolated metadata communities, a fate we’d prefer to avoid (It is possible I have twisted the intent of Jonathan’s comment to serve my argument, but hey… getcher own blog, mate!).
RDF was then, and perhaps remains, the best effort to microformatize metadata and make it more re-usable, but if you’re not disappointed in the uptake of RDF, then you live on a different technological planet than I do. Lo, these many years later, I struggle to understand why it has not (yet?) become more widely used as an architecture for sharing metadata.
- solve a specific problem
- start as simple as possible
- design for humans first, machines second
- reuse building blocks from widely adopted standards
- enable and encourage decentralized development, content, services
Excellent principles, all. Each was explicitly part of our early imaginings about our metadata futures. Except one. I hope there is someone in the DC community motivated to explore this a bit. It seems a useful path.
Image: A church in rural