How to test your archive’s accessibility on older browsers
Older browsers remain part of everyday access in Australia, particularly for people using shared computers at libraries, community centres and small organisations. Testing an archive’s accessibility on older browsers therefore means more than loading a page in an emulator. It asks whether a visitor can find, understand and retrieve an item when scripts fail, fonts are missing or a device has limited memory. The aim is a usable service, not perfect rendering on obsolete technology.
A practical programme combines a representative browser matrix, repeatable user journeys, integrity checks and clear reporting. Start with current accessibility standards, then add legacy conditions that reflect real use. Preserve the test build, evidence and original files so another steward can repeat the work years later. That discipline suits the papajohnphillips resource hub, where historical material, domain research and long-term care are connected responsibilities.
Choose a representative browser set
| Test environment | What it reveals | Common failure to inspect |
|---|---|---|
| Internet Explorer 11 on Windows 7 SP1 | Behaviour on legacy managed computers and embedded document tools | Modern TLS, JavaScript and CSS support |
| Mozilla Firefox 115 ESR on Windows 8.1 | Older Gecko rendering, forms and extension compatibility | Media codecs and platform certificate roots |
| Google Chrome 109 on Windows 7 | The final Chromium generation supporting that operating system | Unsupported APIs and modern JavaScript syntax |
| Safari 12 on macOS Mojave | Older WebKit behaviour and media playback | Autoplay rules and retired media formats |
| Current Chrome, Firefox, Safari and Edge | The present-day baseline for archive users | Regressions hidden by legacy-only testing |
The browser mix should separate a present-day baseline from genuinely legacy environments. Use browser images only in isolated virtual machines or disposable hardware, never in a workstation that handles production credentials. Record the operating system, browser version, user agent, installed fonts and relevant extensions, because two installations of the same browser can behave differently. Include zoom and text-size settings in the label so another tester can reproduce the result.
For an Australian collection, include the paths used by community and public-library audiences. A catalogue record may be discovered through Trove and the National Library of Australia before a visitor reaches a state library service or local archive, so redirects and domain changes matter. Australian not-for-profits may still support members on donated or refurbished Windows computers and rely on volunteer technical help rather than a managed enterprise endpoint. A legacy test can reveal where a modest device or constrained support budget changes the experience.
Build an isolated test environment
Treat each legacy system as evidence rather than a browsing tool. Work from a versioned copy of the archive and a fresh disk image, and exclude mail, banking and other personal accounts. Background reading on local storage cleanup can help explain why temporary caches, downloads and application remnants consume scarce space. Clear only disposable test data; never purge the preservation copy or alter its timestamps to create extra room.
Create snapshots before each pass so a failed plug-in cannot contaminate later runs. Capture both wired and wireless connections, because Australian users in regional and remote areas may depend on an NBN connection with uneven performance, a fixed wireless service or a mobile hotspot. Record latency, packet loss and download speed, since a timeout may look like a broken page even when the interface itself is sound.
Set the viewport, zoom, colour scheme, cookie preference, JavaScript state and font configuration before testing. Run paired passes with scripts enabled and disabled, then repeat at 200 and 400 per cent zoom. Save each screenshot with a visible timestamp, the network profile and the URL. This creates a defensible trail when someone returns the issue six months later.
Test complete user journeys
Build the test around journeys rather than isolated pages. A repeatable path might start with an archived news or community article, move through search, apply a date or place filter, open a catalogue record, inspect metadata, download an item and report a problem. Use the same wording and expected outcome for every browser. Record where the visitor stops, what error text appears and whether a keyboard, mouse or touch action caused the failure.
Check accessibility features as carefully as rendering. Move through every interactive control with Tab, Shift+Tab, Enter, Space and the arrow keys. Confirm visible focus, useful labels, logical headings, skip navigation and a clear way past video, audio or maps. Where licensing permits, pair the browser with the screen reader combinations that Australian users may still encounter. Test captions, transcripts, alternative text and colour contrast instead of assuming a page is accessible because it resembles the original.
Archived objects often need special handling. A PDF may require an old plug-in, an audiovisual item may depend on a retired codec, and OCR text may be embedded in a scan that cannot expose it to an assistive tool. Check that each item has a meaningful fallback page with title, date, provenance and access instructions. Confirm that searching still returns results when rich previews fail, and that download filenames retain useful context for someone sorting files offline.
Confirm integrity while checking access
A page can look correct while its underlying file has changed. The checksum verification guide provides a practical method for comparing published hashes with stored files. Record the expected value, tested value, filename and timestamp together; when no trusted checksum exists, create a baseline only after confirming that the source is the authorised copy. Never alter the archived original to make a test pass.
Integrity is only one part of access. Test the full redirect chain, certificate path, mixed-content warnings and behaviour for moved or deleted objects. A legacy browser may reject a certificate root introduced after the original capture, while a modern browser may redirect through a host that no longer matches the archived service. Keep 404 and maintenance responses informative, and make sure search indexes do not present a dead URL as the only route to surviving material.
Run security-sensitive checks only on systems you are authorised to test. Keep legacy browsers offline from production wherever possible, disable macros and active content, and do not open unknown files outside a sandbox. Note mixed-content and certificate errors as evidence rather than dismissing them as cosmetic. Where a defect could expose users, keep the exact reproduction steps and affected URL in a restricted issue log while publishing a non-sensitive summary.
Record results for Australian users
Translate technical results into obligations that an Australian archive can act on. Use WCAG 2.2 AA as a detailed benchmark and check the current Australian Government Digital Service Standard when the service supports public users. An old-browser failure can still be an accessibility barrier when it blocks magnification, keyboard operation, authentication or a text alternative. Assign severity according to the blocked task, not the visual prominence of the error.
Recruit testers who reflect different access conditions, including regional community groups such as those in Bendigo or Alice Springs, as well as people using older equipment and assistive technology. Participants may call a usable result a “fair go” or ask whether the “old computer” can manage the item; preserve that plain Australian English in quotations, without turning one informal phrase into a sweeping claim. Offer a remote session when travel or unreliable connectivity makes an in-person test impractical.
A useful report states what was tested, when, by which method and with what expected result. Spell out month names and identify AEST or AEDT when coordinating testers in different states. The domain context belongs in the record too: for domain context, the DNS history overview explains how host and routing changes can affect links that appear unrelated to page design. This helps readers separate a content defect from a domain migration and avoid “fixing” an archive that is merely resolving elsewhere.
Publish findings and preserve evidence
Turn findings into a prioritised release note and a maintained test record. A small archive can begin with the most common browser and the single most important retrieval journey, then expand by collection type and user need. Publicise known limitations without making them a substitute for remediation. The following measures provide a practical baseline for each review cycle:
- Maintain a browser matrix containing current standards and one representative legacy environment for each collection type.
- Reuse fixed search, filter, metadata and download journeys with explicit success criteria.
- Record exact software, fonts, settings, network conditions, screenshots and failure notes.
- Separate visual faults from barriers that prevent a task, download or alternative format.
- Publish plain-English results, severity, owners, target dates and the next review date.
- Verify that navigation from archived news and community stories still reaches its intended historical section.
- Store test images, scripts, checksums and consent-protected participant notes with versioned evidence.
Make each completed pass part of stewardship rather than a one-off technical exercise. Attach the report to the archive’s release history, update the preserved materials and keep a direct route for corrections. Visitors can browse the historical material in the papajohnphillips resource hub and use its enquiry channel for domain or stewardship questions. Run one dated legacy-browser pass, publish its evidence beside the archive, and route domain or stewardship enquiries through the same hub.