GIS, KOTLIN, MULTIPLATFORM
A 6-year-old WPF/C++ GNSS monitoring app was fully rewritten into a modern Kotlin Multiplatform with Compose UI solution, achieving faster updates, higher stability, and cross-platform growth opportunities.
Reduced device setup time
60%
faster
Reduced memory use
38%
lower
Onboarding time
3x
faster
GIS Platform
USA
Geospatial hardware & software
Architecture assessment and roadmap creationConcept validation and migration planning (KMP + Compose)Full rewrite with legacy feature parityUI/UX redesign (Material 3)CI/CD setup (Gradle KTS, GitHub Actions)GNSS SDK integrationLegacy installer adaptationSupport and new features Complete rewrite of a WPF/C++ app into a Kotlin Multiplatform application
Business Challenge
the inability to provide instant tooling updates required for your product?
The client was having trouble operating the tool built 6 years ago. It resulted in numerous restraints and shortages, becoming a significant burden for the company.
The code was outdated, brittle, hard to maintain, and slow to update.
Senior WPF/C++ developers are expensive and hard to find.
Low multiplatform opportunities restrained the potential for extending the customer base.
An outdated, x-86-only solution didn’t allow for providing full functionality on ARM tablets.
The client was tied to one installer because of a required signed USB driver, complicating builds
Conducted the code audit: Checked 42k lines of legacy code and identified 19k for a full rewrite.
Selected hexagonal architecture to separate GNSS core logic from UI and platform code.
Recognized packaging as a critical risk and integrated KMP output into the old installer.
Our engineers can update your legacy app the same way to tackle the issues of aging code, platform limits, installer dependency, and unmet product demands.
At the beginning, we created the foundation for future speed, stability, and cross-platform capacity. Arranged KMP modules, dependency injection, coroutines, and a CI pipeline.
We reconstructed the core GNSS port, re-implementing the binary parser, NTRIP client, and RINEX exporter with unit-test coverage. This guarantees accuracy, reliability, and a future-proof setup.
We upgraded the user journey with an advanced Compose UI, implementing a convenient dashboard, satellite sky-plot, and accuracy charts, for conducive interactions and insights.
We expanded integration by adding SDK support for the latest receiver model. This enabled real-time telemetry and battery monitoring, ensuring smooth compatibility and extra value for users.
To guarantee seamless delivery, we wrapped the new KMP EXE into the legacy MSI installer, retaining the signed USB driver. This connected old and new systems, ensuring stability and compliance at launch.
We launched a beta version first to test it with survey partners. The research provided data on performance and user experience, helping confirm stability and polish usability before the full release.
In less than a day,
We confirmed mobile readiness, with the opportunity to craft a working Android version from the desktop codebase in 1 day. Thus, we exploited the power of a single source for flexible growth across platforms.
WPF on .NET 4.x (C#)
C++ libraries for telemetry
WiX-based MSI installer with embedded driver
Manual build/deployment
Kotlin Multiplatform shared core
Compose Multiplatform UI
Gradle KTS, Coroutines, Koin, kotlinx-serialization
Automated CI/CD with nativeDistributions
We faced the inability to change the MSI installer due to a required signed USB driver. This complicated the pipeline builds.
GNSS receivers sent data at 10 updates/sec, leading to intermittent performance. We resolved this with lazy diffing and Skia canvas pooling.
Windows had a USB driver, while macOS didn’t. The team used libusb with Kotlin/Native bindings to access the devices on macOS.
Rewriting all legacy math would be a lengthy job. Instead, we used Kotlin/Native CInterop to wrap the existing C++ code, saving time and ensuring functionality.
Results
Migration from legacy WPF to KMP/Compose delivered faster feature delivery, improved stability, reduced memory use, and true cross-platform capability.
With legacy WPF, adding a new GNSS device took 4 months. With KMP/Compose, it dropped to 6 weeks
The cold-start RAM has been reduced from 420 MB with legacy WPF to 260 MB with KMP/Compose
The system’s stability significantly increased compared to the legacy code, reaching virtually crash-free operation
With legacy WPF, onboarding new developers took 6 weeks, while with KMP/Compose, the timeline was cut to 2 weeks
We transformed a Windows-only setup into a multiplatform solution, ready for ARM tablets, macOS, Linux, or mobile devices
2x
faster feature delivery, including updates for new devices
38%
lower memory use
Near 100%
performance stability with minimized system crashes
Remove the limits of the legacy MSI by replacing it with a Kotlin‑native installer using Gradle and driver-injection scripts. This will reduce build times by ~15% and eliminate dependency on WiX.
Develop macOS and Linux versions and release Android and iOS companion apps. This extends user access throughout desktop and mobile platforms.
Prepare the GNSS diagnostic library as a Kotlin Multiplatform SDK to drive effortless integration of core functionality into partners’ apps and solutions.
Legacy Software Modernization