Measure
I separate the execution into stages and record volume, time, repetitions, errors, and the verified result.
Engineering for real systems
Ioan Marian
Senior .NET Engineer · Enterprise platforms and performance
I find the breaking point before the user reaches it. I measure what actually happens, remove unnecessary work, and build systems that stay fast, explainable, and transferable as data and requirements grow.
MeasureVolumes, stages, time, repetition
UnderstandThe real cause, not only the symptom
RemoveUnnecessary data, work, and steps
ContinueClear architecture and transferable decisions
28 yearsof hands-on professional work
20+ yearsof C# and .NET
3 generationsof SharePoint responsibility
~100×documented historical paging case
The working method
Simple questions, asked until the system becomes clear
Performance is rarely one spectacular instruction. It is the accumulated cost of ordinary decisions repeated at scale.
Simplify
I reduce the amount of data moved and the work repeated before adding hardware or complexity.
Transfer
I make the model, log, tests, and decisions clear enough for another specialist to continue.
The book
The Critical Path
How I Find the Limit Before the User Reaches It
A practical book about measurement, logs, breaking points, pagination, search, processing, reusable models, and the decisions that keep systems useful as they grow.
Reproducible evidence
The demonstrations can be inspected, downloaded, and run
The public Visual Studio solution uses synthetic data and keeps the code, five measured runs, verified outcomes, detailed logs, and a SHA-256 manifest. It contains no client code, schema, or production data.
18independent demonstrations
90primary measured logs
5runs for each demo
0client datasets
Selected impact
Many applications, complex systems, and responsibility over time
I have tested the same method in two complementary ways: by reusing a clear model across many applications, and by remaining responsible for operational platforms long enough to see what endures.
I do not optimize for an attractive number. I remove unnecessary work so the application can remain fast, understandable, and easy to change as data and requirements grow.