decentralised open indexes for discovery
image: nadh.in
I recently came across an interesting concept proposed by Kailash Nadh in his blog post, "Decentralised Open Indexes for Discovery (DOID)." In it, he outlines a vision for rethinking how resources on the web could be indexed and made discoverable in a decentralized manner.
As someone who maintains a love-hate relationship with SEO, I can relate to Nadh's frustration with the degradation of search and the overrun of low-quality content driven by perverse incentives. Over the past decade, finding genuinely valuable information through conventional search engines has become increasingly difficult.
Some additional ideas I have been thinking could help extend and flesh out DOID:
- Encourage the use of linked data and semantic annotations in indexes to enable more powerful discovery. Tags are a good start, but formal concepts/relationships could unlock new clients.
- Develop common quality/relevance scoring metadata that indexes can optionally include. This data could then be aggregated to generate meta-indexes and impact rankings/trust over time.
- Make indexes themselves queryable via simple APIs to allow search/aggregation without needing a full local copy. Index owners could opt-in for lightweight serving.
- Experiment with collaborative curation models like Wikipedia to maintain index integrity, avoid hijacking, and integrate community feedback.
Overall, I'm excited by his DOID vision.
There is certainly more to ponder and try building. I look forward to following any ongoing discussions or experiments in this space.