Web Design & Development · Testing

WDD14 - Testing: usability and compatibility

Thu 10 Dec 2026 · P2/P3 (double)
Content scoped for a double period · matches available time
Learning intentions
Success criteria
Warm up - forms, media and layout recap

WDD14 tests the HTML forms (WDD5), media (WDD4/5) and responsive layout built across the course so far.

WU1
Which HTML attribute makes an input field compulsory, so a form cannot be submitted without it filled in?
 
WU2
WDD3 introduced a fictional user profile used to think about real user needs. What is this called?
 
WU3
Which CSS technique, used throughout this site's own stylesheets, adapts a page's layout for different screen widths?
  • :hover responds to mouse position, not screen width.
  • Correct - @media (max-width: ...) rules adjust layout for narrower screens.
  • float positions elements but does not itself respond to screen width.
  • onclick is a JavaScript event, unrelated to screen width.

Key vocabulary

Persona
A fictional profile of a target user, built from realistic needs and goals, used to guide testing decisions.
Scenario
A realistic task a persona tries to complete on the website, used to frame usability testing.
Test case
A specific, step-by-step set of actions with a defined expected outcome, used to check whether a scenario succeeds.
Input validation testing
Testing a form field with normal, extreme and erroneous data to confirm validation rules behave correctly.
Orphan page
A page that exists on a website but has no link pointing to it from anywhere else on the site.
Media testing
Checking that images and other media load correctly, have appropriate alt text, and do not break the page layout.
Device compatibility
Whether a website displays and functions correctly across different screen sizes, such as phone, tablet and desktop.
Browser compatibility
Whether a website displays and functions correctly across different web browsers, such as Chrome, Firefox, Safari and Edge.

Usability testing: personas, scenarios and test cases

Three tools, used together

WDD3 previewed three testing terms while planning wireframes for Meadows Wildlife Centre's website. This lesson teaches them properly. A persona is a fictional profile of a target user - not a real named individual, but a realistic composite built from the kind of person the site actually needs to serve. A scenario is a specific, realistic task that persona would try to complete using the site. A test case then makes that scenario testable: a precise sequence of actions, together with a clearly stated expected outcome, that a tester can actually carry out and mark as pass or fail.

Why all three, not just one

A persona on its own says nothing about what to test. A scenario without a test case is not checkable - "find a workshop and enquire" is a good scenario, but it needs a test case's precision ("enter school name, class size 28, and a date within the next term, then press submit; expect a confirmation message") before anyone can actually run it and get a pass or fail result. Used together, the three build outward from a realistic person, to a realistic goal, to something a tester can genuinely check off.

Testing input validation, navigation and links

Validation testing needs three kinds of data

WDD5 taught how to build validation rules - presence, length and range checks on form inputs. Testing those rules properly means trying all three categories of test data on each one: normal data that should be accepted (a class size of 20, within an allowed range of 1-35), extreme data at the very edge of what is allowed (exactly 1, or exactly 35), and erroneous data that should be rejected (0, 36, or text typed into a number field). Testing only with normal data can make a form look correctly validated when its boundary or rejection behaviour is actually broken.

Testing navigation, links, and orphan pages

Link testing checks that every navigation link and in-page link actually leads to its correct, intended destination - not just that it is clickable, since a broken or mistargeted link is still clickable. A separate, easily missed check is for orphan pages: pages that genuinely exist on the site, but that no other page links to. An orphan page cannot be reached by a user browsing normally, and is effectively invisible, even though the file itself works perfectly if visited directly. Finding orphan pages means checking in the opposite direction from ordinary link testing - not "does this page's links work", but "does anything on the site actually link to this page at all".

Testing media and cross-device/browser compatibility

Media testing

Media testing checks that every image and other media element actually loads, displays at a sensible size without breaking the surrounding layout, and has appropriate alt text for accessibility (recapping WDD4's semantic HTML). A missing or incorrect alt attribute is a genuine testing failure, not a cosmetic detail, since it directly affects whether a screen reader user can access the same content.

Device compatibility

Device compatibility testing checks that a page displays and behaves correctly across different screen sizes - phone, tablet and desktop - since a layout built and only ever viewed on one screen size can look completely broken on another. This site's own stylesheets already use @media (max-width: 640px) rules throughout for exactly this reason.

Browser compatibility

Browser compatibility testing checks the same page in multiple different browsers - Chrome, Firefox, Safari, Edge - because different browsers can render CSS and run JavaScript with small but real differences. A page that looks correct in the browser it was built in can still have a genuine, undiscovered fault in another.

Same page, two different testing questions does this page work?
Link testing (from the page) Do this page's own links work?
Orphan page check (to the page) Does anything link TO this page?

Worked examples

Example 1 - Persona, scenario and test case for Meadows Wildlife Centre
1
Persona: "Primary teacher booking a workshop for 28 pupils."
2
Scenario: "Find a suitable school workshop and send an enquiry."
Test case: Enter school name "Bridge Primary", class size "28", preferred date "next Tuesday", then press submit. Expected result: a confirmation message appears and the enquiry is recorded. Pass/fail is decided directly against this expected result.
Example 2 - Validation test data for the class size field (allowed range 1-35)
Test dataCategoryExpected result
20NormalAccepted
1ExtremeAccepted
35ExtremeAccepted
0ErroneousRejected
36ErroneousRejected
"twenty"ErroneousRejected
All three categories are tested, not just normal data - the extreme values at 1 and 35 confirm the boundary is inclusive, and the erroneous values confirm out-of-range and non-numeric input are both correctly rejected.
Example 3 - Finding an orphan page
1
A site has four pages: Home, Animals, Workshops and Prices. Home links to Animals and Workshops. Animals links back to Home. Workshops links back to Home and to Prices.
2
Checking each page's own links (ordinary link testing) shows all four pages' links work correctly - nothing is broken.
Checking the reverse direction shows Prices is only ever reached from Workshops, and nothing links to it from Home directly - if a user never visits Workshops, Prices is effectively unreachable. Whether this counts as a genuine orphan page depends on whether at least one path exists at all; a true orphan page has no path in from anywhere, which link testing alone (checking outward from each page) will never reveal on its own.
Now you try

A form field accepts a date within the next 90 days. Suggest one normal, one extreme, and one erroneous piece of test data for it.

Normal: a date 30 days from today.
Extreme: exactly 90 days from today (the boundary).
Erroneous: a date 91 days from today, or a date in the past.

Practical: testing your own earlier work

Open one of your own WDD1-13 practical files (or the Meadows Wildlife Centre example from WDD3) and apply today's checklist directly to it.

  1. Write one persona and one matching scenario for a genuine feature of the page.
  2. Write a specific test case for that scenario, with a clear expected result.
  3. If the page has a form, design normal/extreme/erroneous test data for one field.
  4. Check every link on the page leads to its correct destination, and check whether any page in the site is never linked to from anywhere else.
  5. Open the page in at least two different browsers, and view it at both a narrow (phone-width) and wide (desktop-width) browser size, noting anything that looks or behaves differently.
Common mistakes
Exam tip

Use the precise terms persona, scenario and test case as three distinct things, not interchangeably. For validation testing, always name all three categories - normal, extreme and erroneous - rather than describing testing in general terms. Remember an orphan page is found by checking what links to a page, not by checking that page's own links outward.

Task Set A - Core questions

Task Set A - Core questions
Complete all questions. Written answers reveal a model answer for self-assessment.
A1
What term describes a fictional profile of a target user, used to guide testing decisions?
 
A2
What term describes a realistic task a persona tries to complete on a website?
 
A3
What term describes a specific step-by-step check, with an expected result, used to test whether a scenario succeeds?
 
A4
Which three categories of test data should validation testing always include?
  • These describe speed, not validation test data categories.
  • Average and typical both describe the same "normal" category - they miss extreme and erroneous data.
  • Correct.
  • Testing only normal data can miss broken boundary or rejection behaviour.
A5
What term describes a page that exists on a website but has no link pointing to it from anywhere else on the site?
 
A6
A pupil checks every link on a page and confirms they all lead to the correct destinations. Have they confirmed the site has no orphan pages?
  • This checks outward from each page, not which pages are linked to.
  • Correct.
  • Site size does not change what checking a page's own links reveals.
  • They can be detected, but by checking incoming links, not outgoing ones.
A7
Explain why testing a website on only one browser and one device is risky, even if every test passes on that combination.
Model answer
A8
Which of these is a genuine media testing failure, not just a cosmetic detail?
  • Correct - this directly affects whether a screen reader user can access the same content.
  • This is a cosmetic/design concern, not a testing failure.
  • This is a layout preference, not a testing failure.
  • This is a stylistic choice, not a testing failure.

Task Set B - Extension

Task Set B - Extension
Beyond the core specification - written answers are self-assessed against a model answer.
B1
Design a new persona, scenario and test case for a Meadows Wildlife Centre visitor with an accessibility need not covered by Example 1.
Model answer
B2
Normal, extreme and erroneous data are also used to test programs in the SDD unit. Explain the connection between that idea and input validation testing here.
Model answer
B3
Explain why an orphan page is a problem both for real users browsing a site, and for how easily the page can be found through a search engine.
Model answer

File this in OneNote under:
Higher Computing Science > Web Design & Development > WDD14

Teacher notes - Shift+T to toggle

Timing: content is scoped for a double period and lands on one (Thu 10 Dec 2026, per WDD.md §2a/§4 table) - an "OK" match. This lesson deliberately covers six sub-topics that Chris compresses into half of one lesson (L14): usability testing (persona/scenario/test case), input validation testing, navigation/link testing including orphan pages, media testing, device compatibility, and browser compatibility. Suggested pace: 8 min warm up, 15 min usability testing notes + Example 1, 20 min validation/link/orphan-page notes + context box + Examples 2-3, 15 min media/device/browser notes, remaining time on Task Set A, then Task Set B and the practical.

Scope control: no TYPE 4 code block in Task Set A, since this is a testing/evaluation lesson, not a coding lesson, per CLAUDE.md's instruction to remove it when not applicable. The practical deliberately reuses pupils' own earlier WDD files rather than introducing a new build, keeping the double period's generous time focused on testing skills rather than new construction.

Persona/scenario/test case terminology was already previewed in WDD3 (Meadows Wildlife Centre wireframing) - this lesson is pupils' second, formal exposure to the same terms and reuses the same fictional project for continuity, matching the two-stage pacing WDD.md documents for this pair of lessons.

The orphan page concept is the highest-value gotcha in this lesson - pupils very naturally check a page's own outgoing links and assume that is sufficient, missing that orphan detection requires the reverse, incoming-link check. A5/A6/context box/Example 3/B3 all test this from different angles.