liveBook · Manning

Antigoal

As explained in Machine Learning System Design

An explicitly identified outcome or condition that a system design should avoid.

An antigoal is an explicitly identified outcome or condition that a system design should avoid. It acts as an inverse statement alongside goals, narrowing both the problem space and the solution space. section 4.2

Use antigoals to make tradeoffs visible before implementation:

  • Strike out qualities that do not materially improve the system’s main objective. section 4.2
  • Prevent an endless list of competing goals from making the design impossible to satisfy. section 4.2
  • Keep the boundaries visible during design-document review, where vague requirements can later become change requests. section 4.6

A boutique-store recommender

For an internal reporting system, processing time can be an antigoal when reports only need to be ready by morning. For a boutique store with fewer than 1,000 products and low traffic, scalability is an antigoal: supporting millions of items would add no useful value. section 4.2

Figure 4.1 A shop with <1,000 goods for sale and low traffic should not aim for scalability when building a recommendation system, as almost any tech solution can handle its load these days.
figure

Use the figure to see how product count and traffic can make scalability an unnecessary design target. section 4.2

How Machine Learning System Design covers it

How deep it goes
Defines inverse statements, explains their tradeoff value, and applies them to reporting and recommendation scenarios.
Examples
Uses report readiness, processing time, boutique-store product counts, traffic, and recommendation-system scalability.
What it assumes
Assumes an identified system problem, goals, stakeholders, and rough candidate technologies or solutions.
What you can do after
Apply antigoals to focus requirements, expose tradeoffs, and remove irrelevant system capabilities before implementation.
How it connects
Connects goals, success criteria, tradeoffs, design documents, stakeholders, and design review.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning System Design. The section links point to the parts of the book it draws on.