Desirable

Podcast 2 — In conversation

Speaker A: Can I run something by you? I am writing a job description and keep second-guessing whether to say required or desirable for certain skills.

Speaker B: That is actually an important distinction. Required means it is a dealbreaker if they do not have it. Desirable means: we would love this, but it will not eliminate you from consideration.

Speaker A: So using desirable opens the door to more applicants without diluting the role.

Speaker B: Exactly. Experience with Kubernetes is desirable tells candidates: this is a bonus, not a barrier to applying.

Speaker A: And in everyday speech — would you actually use the word desirable outside of formal writing?

Speaker B: You would, but it carries a slightly elevated register. You would not usually say to a friend: that jacket is very desirable.

Speaker A: You would say I really want that, or everyone is after one.

Speaker B: Right. Desirable implies a judgement about objective worth — not just personal preference. A desirable outcome suggests it is genuinely a good outcome.

Speaker A: Whereas desired outcome is just the one you aimed for — descriptive rather than evaluative.

Speaker B: Exactly. Desirable evaluates. Desired describes. That distinction matters in precise writing.

Speaker A: And what about its nearest neighbours? Desirable versus enviable?

Speaker B: Enviable adds the social comparison dimension — you want it partly because others do. A desirable location is simply an excellent location. An enviable location means others wish they had it too.

Speaker A: And attractive?

Speaker B: Attractive emphasises the pull — the way something draws you in. Desirable emphasises that the thing is genuinely worth wanting. There is more considered judgement in desirable.

Speaker A: So it is the most evaluative of the group.

Speaker B: I would say so. The classic estate agent example: a desirable property in a sought-after location. Both words doing precise work in five words.

Speaker A: And not just marketing puffery — if the property is genuinely good, both words are accurate.

Speaker B: That is the thing about desirable — when used well, it is a mark of genuine quality, not mere salesmanship.

Podcast 3 — Prompt engineering

Instructor: Today we are focusing on how desirable works in AI development prompts. It is underused — which is a shame, because it is remarkably precise.

Student: I would not have expected an adjective to carry that much functional weight in a prompt.

Instructor: When you write desirable in a prompt, you mark something as a goal rather than a requirement. That is a specific prioritisation instruction to the model.

Student: So it teaches the model to distinguish between essential and optional.

Instructor: Precisely. First — feature prioritisation for a product roadmap:

Instructor: You are a senior product designer. For a SaaS project management tool, define the required features for the MVP and the desirable features for version two. Organise under: task management, team collaboration, reporting, and notifications. Format as a two-column structured Markdown table with a priority rationale column.

Instructor: Desirable forces a genuine prioritisation layer into the output. Without it, you get a flat, undifferentiated feature list with no guidance on what to build first.

Student: That is immediately useful for roadmap planning and stakeholder communication.

Instructor: Next — database schema with priority tiers:

Instructor: Design a relational database schema for a recruitment management platform. Mark each field as required or desirable. Include tables for Candidates, Jobs, Applications, Interviews, and Offers. Clarify which fields are operationally essential and which are desirable for analytics and reporting. Output as a formatted schema with column definitions and recommended indexes.

Instructor: Desirable creates a two-tier schema — operational necessity versus analytical value. Very practical for real database design conversations with clients.

Student: It is like building the lean version and the full version in one prompt.

Instructor: Exactly. Here is a full application build — notice how desirable shapes the implementation strategy:

Instructor: Build a web-based HR employee onboarding application using HTML, CSS, and vanilla JavaScript. Required features: employee registration form, document upload checklist, and department assignment screen. Desirable features: animated multi-step progress bar, welcome email preview panel, and printable onboarding summary. Implement all required features fully. Scaffold desirable features with placeholder comments and basic UI shells.

Instructor: This is very powerful. The model builds the critical path completely and stubs the rest intelligently. You get a production-ready foundation with extension points already in place.

Student: And you get a codebase that is immediately extensible rather than a half-finished jumble.

Instructor: Right. Here is one for UX specification — framing design as ambition:

Instructor: Act as a UX designer. Define the desirable user experience characteristics for the checkout flow of a B2C e-commerce platform. Cover: micro-interactions for each input field, error handling with inline recovery suggestions, loading state designs, and order confirmation animation. Output as a numbered UX specification with a brief design rationale for each point.

Instructor: Asking for desirable user experience characteristics invites aspirational thinking. The output frames design as ambition, not as a minimum viable experience.

Student: That framing shift genuinely changes what the model considers worth including.

Instructor: Here is a REST API characteristics prompt:

Instructor: You are a software architect. Define the required and desirable characteristics of a REST API for a document management system. Required: JWT authentication, URL-level endpoint versioning, and rate limiting. Desirable: a GraphQL compatibility layer, webhook support for document events, and full-text search integration. Provide example endpoint definitions and JSON response schemas for both tiers.

Instructor: And finally, an analytics dashboard with layered requirements:

Instructor: Design a desirable analytics dashboard for an e-commerce operations team. Required panels: daily revenue summary, top-selling products list, and low-stock inventory alerts. Desirable panels: predictive restocking suggestions, geographic order heat map, and A/B test result comparison view. Output complete HTML, CSS, and vanilla JavaScript with inline SVG charts and a responsive two-column card layout.

Student: Every single prompt uses desirable to add precision rather than decoration.

Instructor: That is the lesson. Desirable communicates intentional prioritisation. It tells the model: this is a goal worth pursuing, not a strict requirement. Build the essential first, then extend toward the desirable. Use it deliberately and your prompts become dramatically more actionable.