Our History

OUR HISTORY

Our use of the high integrity language, Ada, has long played an important role in formulating and implementing leading-edge software development processes in mission critical systems. 

For example, our internationally recognised electronic voting and counting system (eVACS®), for parliamentary elections. Ada is a proven language that strongly supports the construction of safe, secure, quality systems.
 
The discipline associated with the use of the Ada language we apply to improve the efficiency and effectiveness of all our clients’ software development efforts.  A natural outcome of applying these principles is to model both requirements and architecture as the essential input to automated implementation.

In 2016, we partnered with Stratum Communications (now Stratum ICT) to form Digital Elections Pty Ltd to specifically provide election solutions and services based on eVACS® and netVote® an online system operating within a secure cloud environment.

More recently we were engaged via TAQNIA (Saudi Development and Technology Company) (https://taqnia.com/en/) to guide and contribute to the development of a prototype product by the King Abdulaziz City for Science and Technology (KACST), utilising our expertise in developing critical systems using model driven development and the open source software BridgePoint for xtUML Modelling and Model Compiler Development, available from OneFact (https://onefact.net).  Software Improvements has been using BridgePoint internally and providing support to other users since 1998.

ONGOING HISTORY

Our history is still being written. The arrival of high-reasoning AI systems is among the most significant shifts we have seen in systems and software engineering, and we are adopting it the way we have always adopted powerful new technology: inside a discipline.

We now use AI across the whole development life cycle: sharpening requirements, drafting models and architectures (AADL and xtUML), and drafting SPARK Ada code with its proof contracts. What has not changed is where we place our trust. Every AI-produced artefact is checked by qualified, deterministic tools, such as GNATprove proofs and architectural timing analysis, and signed off by an experienced engineer. In short: a fast generator, a strong verifier, and a human who signs.

We presented this approach at the Australian Systems Safety Conference (ASSC2026), and write about it in our News & Insights.