TechPrufer
Get started
Slack
Slack logoSlack · Feature trail

Conversations API

2 posts · Jan 2018 – Aug 2021

What this trail covers

This thread tracks the evolution of Slack's API design, focusing on creating a unified interface for all channel types. Initial posts detailed the challenges of evolving APIs and the need for backwards compatibility. This post introduces the Conversations API, a new interface designed to unify channel types (public, private, DMs, shared channels) under a single set of endpoints. Key improvements include smarter OAuth scope resolution, simplified and consistent arguments, better performance for l

All Slack trails
  1. How We Design Our APIs at SlackThis post details Slack's API design principles: "Do one thing and do it well" (illustrated by rtm.start vs. rtm.connect), "Make it fast and easy to get started" (Time to First Hello World metric, interactive docs), "Strive for intuitive consistency" (industry, product, and internal consistency), "Return meaningful errors" (actionable, unambiguous errors with longer-form descriptions), "Design for scale and performance" (pagination, rate limiting, illustrated by channels.list vs. conversations.list/members), and "Avoid breaking changes" (philosophy of "what worked yesterday should work tomorrow"). It also outlines the design process involving API style guides, reviews, and testing.
  2. Evolving the Slack APIIntroduced the Conversations API as a unified interface for all channel types, addressing the complexities introduced by Shared Channels. Implemented smarter OAuth scope resolution to prevent unauthorized access to private channels. Simplified API arguments and made non-default arguments the new default for improved usability and performance. Addressed performance bottlenecks for large payloads by defaulting to excluding members and introducing a new conversations.members endpoint with pagination. Committed to cursor-based pagination for all endpoints to improve scalability. Implemented strict JSON schema validation in RSpec tests to ensure API reliability and generate OpenAPI specs.