模組 3 · Decision Lenses: Licensing, Cost, Integration, and Build-vs-Buy
第 5 節,共 5 節 · 4 分鐘 · 英語 · 唯讀
Capability parity does not mean equivalent fit. This lesson adds four decision lenses to the comparison matrix you have been building: license type, cost model, integration breadth, and build-vs-buy fit. By the end you should be able to read the same capability rows and reach a defensible position on whether to adopt an offering or build on an open-source baseline.
Your matrix so far answers one question: does an offering have a capability? That is necessary but not sufficient. Two offerings can fill the same orchestration rows and still be very different choices, because the rows say nothing about what you give up to get the capability.
The four lenses in this lesson each change the comparison in a different way:
The central relationship is this: capability parity does not imply equivalent fit. Open-source baselines trade license freedom and control for operational burden; commercial SaaS trades subscription cost for managed operations; integration breadth determines how much of your existing stack the offering can reach. A row that reads "yes" in both columns can still hide a very different total commitment.
來源支援
來源支援
Hypothetical scenario. A team has filled its capability matrix and finds two offerings that fill the same three orchestration rows: tracker-driven dispatch, per-issue isolated execution, and agent harness selection. Offering A is an open-source baseline in the Symphony family (Apache 2.0 or MIT licensed, self-hosted, no subscription). Offering B is a commercial SaaS offering with tiered subscription pricing and a broad prebuilt integration catalog.
The team's situation: they have two engineers who can run and maintain a long-running service, they already use Linear as their tracker, and they need Slack and Jira integrations for other teams. Assume the capability rows are genuinely equal and the only differences are the four lenses.
Question: which lenses decide the choice, and what would make the answer flip?
Step 1 — Read the license lens. Offering A is permissively licensed (Apache 2.0 for Symphony, MIT for OpenSymphony), so the team can modify and self-host it. Offering B is proprietary commercial terms. This lens favours A if the team needs to modify the code or avoid vendor dependency, and favours B if they would rather not own the code at all.
Step 2 — Read the cost lens. Offering A has no subscription fee but requires the team to run the service, manage credentials, and keep the runtime healthy. Offering B has tiered subscription pricing — for example, Devin lists Free $0, Pro $20/month, Max $200/month, Teams $80/month plus $40/month per full dev seat, and Enterprise on request; OpenHands Cloud lists a Free open-source tier, an Individual tier with BYOK or provider at-cost, and Enterprise custom pricing. This lens favours A if the team's two engineers have spare capacity, and favours B if their time is more valuable elsewhere.
Step 3 — Read the integration lens. Offering A's integration breadth depends on what the team builds or configures. Symphony's Elixir implementation ships tracker adapters for Linear, GitHub Issues, Jira Cloud, Asana, and GitLab, and exposes provider-native tools such as `linear_graphql` and `github_api`. Offering B lists prebuilt integrations — Devin lists Slack and Teams, Linear and Jira, GitHub, GitLab, and Bitbucket, plus custom git provider support; OpenHands Cloud adds Jira and Slack integrations and Git provider integrations. This lens favours B if the team needs Slack and Jira integrations that A does not already provide.
Step 4 — Combine the lenses into a build-or-buy position. The team has two engineers who can run the service, so the operational burden of A is absorbable. They already use Linear, which A supports. But they need Slack and Jira integrations for other teams, and A's breadth there depends on what they build. So the position is: adopt A if the team is willing to build or configure the Slack and Jira integrations themselves, and adopt B if the integration work would consume more engineering time than the subscription costs.
Step 5 — Identify what would flip the answer. If the team lost one of its two engineers, the operational burden of A would become harder to absorb and B would become more attractive. If the team needed to modify the orchestration logic itself, A's permissive license would become decisive and B would be ruled out. If the team's integration needs were already covered by A's shipped adapters, the integration lens would stop favouring B.
來源支援
情境題
Hypothetical scenario. A different team has filled the same capability matrix and is looking at the same two offerings: an open-source baseline in the Symphony family (permissively licensed, self-hosted, no subscription) and a commercial SaaS offering with tiered subscription pricing and a broad prebuilt integration catalog. The capability rows are equal.
What is different this time: the team has no engineers who can run a long-running service, they use a tracker that the open-source baseline does not ship an adapter for, and they need SSO and centralized admin controls for a 200-person organization.
Using the four lenses, identify which evidence in this scenario pushes toward the commercial SaaS offering and which pushes toward the open-source baseline. Then state what information is still missing before you could reach a defensible build-or-buy position.
想一想
Take one row from your capability matrix that reads "yes" for two offerings. Which of the four lenses — license, cost, integration, or build-vs-buy — would most change how you read that row, and what evidence would you need before you could call the two offerings equivalent?