Never risk customer data
Features involving deletion, storage, export, recording or licensing must fail safely and preserve recoverability.
Retrace search
Retrace Enterprises ยท Quality Assurance Plan
This plan defines how Retrace Enterprises will test products before release, classify defects, protect customer data, verify performance and compatibility, and respond to problems after launch. Quality assurance is treated as a release gate, not a final cosmetic check.
The standards that apply to every Retrace product and release.
Features involving deletion, storage, export, recording or licensing must fail safely and preserve recoverability.
The main customer task must be tested end to end on supported systems before release.
Known restrictions, unsupported systems and unresolved non-critical issues must be documented honestly.
When a release causes harm or failure, Retrace must investigate, communicate and correct the problem promptly.
A release cannot proceed until every mandatory gate has passed.
Features, expected behaviour, supported systems and acceptance criteria are documented.
Primary workflows succeed without crashes, data loss, corruption or misleading results.
Existing features still work after the new change.
Supported operating systems, screen sizes, permissions and hardware conditions have been tested.
Startup, memory, CPU, disk and long-running operations remain within acceptable limits.
Permissions, local data, network requests, licensing and sensitive operations have been inspected.
Installer, update process, versioning, licence activation, rollback and documentation are correct.
Every release should include a balanced mix of automated and manual testing.
Verify each feature against its intended outcome, including valid, invalid and edge-case inputs.
Re-run critical workflows after every meaningful code or dependency change.
Confirm that labels, flows, errors, permissions and recovery steps are understandable.
Test operating systems, hardware classes, display scaling, permissions and installation conditions.
Measure startup time, memory, CPU, disk I/O, battery use and long-running operations.
Review command execution, local storage, update delivery, licensing and exposed interfaces.
Test interrupted operations, damaged files, failed updates, cancelled tasks and restart behaviour.
Verify clean install, upgrade, uninstall, reinstall and licence restoration.
Defects are prioritised by customer harm and business risk.
| Severity | Definition | Examples | Release Rule | Target Response |
|---|---|---|---|---|
Critical |
Data loss, unsafe deletion, security exposure, app unusable for many users | Corrupted projects, deleted protected files, licence system bypass or lockout | Release blocked | Immediate |
High |
Major workflow broken with no acceptable workaround | Scanning fails, recording cannot save, import/export unusable | Release blocked | Same day |
Medium |
Feature impaired but workaround exists | Incorrect sorting, minor activation issue, intermittent display defect | Must be reviewed and documented | Within 2 business days |
Low |
Cosmetic, wording or low-impact inconsistency | Spacing issue, minor typo, non-blocking visual defect | May ship if logged | Scheduled |
Each product carries different risks and therefore needs different emphasis.
Testing must reflect the actual environments customers use.
Test current supported Windows versions for desktop products and supported Android versions for Stone.
Include constrained systems, average consumer devices and modern hardware.
Verify normal, high-DPI, small-screen and large-screen layouts.
Test restricted folders, external drives, low disk space, read-only locations and interrupted access.
Performance targets should be measurable and product-appropriate.
Core interface should become usable promptly without unnecessary blocking work.
Memory use should not grow continuously during repeated scans, edits, exports or recordings.
Heavy operations should be progressive, cancellable and isolated from the interface.
Stone should minimise wake locks, background processing and unnecessary screen activity.
Every release must minimise unnecessary access and protect local customer data.
Request only the permissions required for the documented feature.
Protect customer files from corruption, accidental exposure and insecure temporary storage.
Verify update source, package integrity, version handling and recovery from failed updates.
Test valid, invalid, expired, offline and restored activation states without exposing licence weaknesses.
The final sign-off required before publishing a build.
Quality assurance continues after the product is published.
Monitor activation, installation, crashes, data issues and support spikes.
Analyse support tickets, refunds, repeated complaints and performance reports.
Compare defect rate, crash rate, refund rate and customer satisfaction with the previous release.
Use a hotfix, temporary mitigation or rollback when customer harm exceeds the risk of another release.
Weekly and release-level measures for tracking product health.
Measure defects discovered by customers after release.
Track sessions completed without unhandled failure.
Measure whether previously fixed defects return.
Identify releases that create abnormal support demand.
Track how often a published release must be withdrawn or replaced urgently.
Compare startup, memory, CPU and disk behaviour across versions.
Separate quality-related refunds from preference-based refunds.
Measure how quickly serious production defects are fixed.
The order in which the quality system should be established.