In a mobile system-design round for a city transit app with mobile tickets, what are the functional and non-functional requirements?
answer
- what the rider does vs how well
- plan, departures, buy, show
- the gate may have no signal
- prepaid data, old phones, low battery
- screen readers and large text count too
basics
~20 sFunctional requirements are what riders do: plan trips, check live departures, buy tickets, show one at the gate. Non-functional ones say how well: the ticket opens fast with no signal, data and battery stay cheap, screens are accessible, old phones are supported.
solid answer
~50 sI split the list in two. **Functional**: plan a trip between two places, see live departures for a stop, buy a ticket, and show a valid ticket at the gate or to an inspector, plus a 'my tickets' list and saved stops if time allows. **Non-functional**, the part an installed app adds: the ticket must open at an underground gate with **no signal**; the time from tap to a visible ticket must be short, measured from a cold start; **data and battery** must stay cheap because many riders are on prepaid plans; every core flow must work with a **screen reader and large text**; and there is an **oldest supported OS version and device class**. I make each one measurable, with a target time or a data budget per week, because 'fast' and 'cheap' cannot be designed against.
go deeper
Recall the two lists: what riders do, and how well it must work. Name offline use at the gate, speed, data, battery, accessibility and the oldest supported phone without being prompted.
Explain what each non-functional requirement forces in the design, and turn every adjective into a measurable target with a stated assumption.
Show which requirement reshapes the architecture, the offline ticket, rank the rest by design impact, and park secondary features aloud to protect time for the deep dives.
Connect requirements to the operator's commitments: a device floor and an accessibility bar promised to the public, and a data budget that decides who can afford to use the service.
## Two lists, two questions A **functional requirement** says what the app lets a person do. A **non-functional requirement** says how well it must do it: how fast, under which conditions, at what cost to the device, for which users and on which phones. In a mobile system-design round the interviewer hands over a one-line prompt, here a city transit app with mobile tickets, and the first minutes go to turning that line into both lists. Functional requirements decide which screens and data exist. Non-functional requirements decide how those screens are built, and for an **installed app** they carry most of the design pressure: the app runs on a phone nobody on the team controls, on whatever network, battery and storage the rider happens to have, and an old build keeps running for years. ## The functional list Keep it short and centred on the rider. For a transit app a reasonable core is: - **Plan a trip** from a place or stop to another, for a departure or arrival time. - **See live departures** for a stop: which line, which direction, how many minutes away. - **Buy a ticket** for a fare product such as a single ride or a day pass. - **Show a ticket** at a gate or to an inspector, as a scannable code with its validity. - **My tickets**: active, upcoming and expired tickets. Secondary items such as saved stops, service alerts or a map of nearby stops are named and then either kept or explicitly parked. Interviewers tend to reward four or five core flows over fifteen, because every item kept is something that must be drawn and defended later. ## The non-functional list an installed app adds | Requirement | Why it bites on a phone | How to make it measurable | |---|---|---| | Ticket works with no signal | Gates are often underground; the purchase happened earlier, elsewhere | A bought ticket opens from a cold start in flight mode | | Time from tap to visible ticket | A queue forms behind a rider fumbling at the gate | A target the team agrees, measured offline from a cold start | | Data cost | Many riders are on prepaid plans and pay for every megabyte | A budget per session and per week of normal use | | Battery | Live departures tempt the app to refresh constantly | No background refresh the rider did not ask for | | Accessibility | A public service is used by people relying on screen readers and large text | Every core flow completed with a screen reader and the largest text size | | Oldest OS version and device class | Riders keep phones for years; old phones have less memory and storage | A named minimum version and a reference low-end device | | Freshness of departures | Live times are useful only while recent | A stated maximum age, with stale times labelled as such | Two more usually earn a line: **reliability of purchase**, because a rider charged twice or charged with no ticket is the worst outcome, and **privacy of location**, because the planner wants location and the rider may refuse it. Both are named here and designed in their own deep dives. ## Turning adjectives into targets The weakest requirements are adjectives: fast, light, offline-friendly. No design can be checked against an adjective, and the habit that separates candidates is converting each one: 1. **Name the condition.** Not 'works offline' but 'a ticket bought earlier opens with no network, after the app was closed and the phone restarted'. 2. **Name the number, or ask for it.** 'Ticket visible within about two seconds of tapping the icon' is an example target; the interviewer may prefer another, and that is fine. What matters is that one exists. 3. **Name who it is for.** The data budget matters most to prepaid riders; the device floor matters to the owners of the oldest phones the operator must serve. 4. **Say what it forces.** No signal at the gate forces the ticket to live on the device; old phones limit memory, storage and which platform features the app may rely on. ## Where candidates go wrong - Listing only features, then copying 'scalability, availability, security' from a server design as the non-functional list. - Talking about server throughput, which belongs to the back end's own design, instead of the phone's constraints. - Forgetting accessibility and the oldest supported device, which a public operator may well be bound to provide and which are expensive to retrofit. - Treating every non-functional requirement as equally important. Ranking comes next, and it follows what each requirement forces: the offline ticket reshapes the design, while a label on stale departure times is a small interface decision. A good opening ends with both lists written where the interviewer can see them and the most design-shaping requirement, here the offline ticket, clearly marked as the one the rest of the round must respect.
- Why not list scalability and high availability as the transit app's main non-functional requirements?Because they describe the back end, not the installed client. The phone serves one rider; its constraints are network quality, battery, storage, old OS versions and accessibility. Back-end availability still matters, but from the app's side it becomes a requirement like 'a bought ticket still opens when the back end is down', which is the offline requirement seen from another angle.
- Which non-functional requirement would you rank first for this app, and why?The offline ticket at the gate. It is the one whose failure physically stops the rider, and it forces a structural decision, the ticket living on the device before the rider reaches the gate, that every later box in the diagram must respect. Data cost and accessibility shape many screens; the offline ticket shapes the architecture.
- What do you do when the interviewer will not put a number on 'fast'?Propose one, label it an assumption and move on: 'I'll assume the ticket should be visible within about two seconds from a cold start; correct me if that's wrong.' A stated assumption can be designed against and revised later; an open adjective cannot be checked at all.
- An app's non-functional requirements are just scalability and uptime
- Offline use is a nice-to-have for a ticket shown at a gate
- Accessibility can be added once the design is finished
- Fast and lightweight are precise enough to design against
- Listing every possible feature makes the scope stronger