The Coherency Manifesto: Towards Communal Data Platforms

August 21, 2021 § 2 Comments

Version 1.0: Sep 11, 2021 (Interdependence Day)

As a community
who produces, consumes, and manages data
we hold these truths to be self-evident:

Our most precious resource as a community is our ability to make better decisions together (“Coherency“)
1. Better decisions are enabled by higher Quality Data
2. Coherency increases as our Communal Data Platform aligns the Syntax of producers with the Semantics of consumers (“The Datocrat’s Conundrum“)
Rational Trust in decision-makers is enabled by immutable public histories of past decisions and context (“Datasets”) plus transparency of the decision-making process over time (“Analyses”)
1. A decision-maker’s Values are best seen in the datasets and analyses they Collect, Retain, Use, Share, Ignore, and Learn from (“The CRUSIL Conjecture“)
Our coherency is constrained by how fast our communal data platform can adapt to new realities
1. Forking Culture enables us to simultaneously pursue both adaptation and coherency
2. The platform Architecture must be built on open source, standards & interfaces so we control our own evolution
3. Service Providers may be used to provide well-defined components and innovative endpoints, where they won’t constrain the overall architecture
We will never do this perfectly, but the better we understand and document our decisions, learnings and failures, the more rapidly the next generation can improve on them

Definitions

Architecture: the fundamental concepts, contracts, and technologies that determine what is possible, feasible, and imaginable
Analysis: The filters, features, priorities, and agencies used to make a decision in a particular context
Coherency: How effectively an entity can harmonize the diverse perspectives of its constituents
Communal Data Platform: An ecosystem of people, pipelines, and policies designed to provide us with continually better data
Dataset: a versioned snapshot of the data and metadata used to make or validate a decision
Forking Culture: using branches and pull requests to enable experimental customizations that may become maintream innovations
Quality Data: (timely | accurate) + (comprehensive | digestible)
Rational Trust: using all available information to estimate the ability of an entity to keep its promises, relative to our values and expectations about the future
Service Provider: any durable entity that promises to add value in exchange for, eg, money, data, or attention
Syntax: formal specifications of behavior, such as those used by a computer
Semantics: the underlying meaning we care about, as experienced by humans
Values: the tradeoffs we are willing to make when faced with hard decisions

Tagged: adaptive, cohesion, data