When people talk about feedback on a thesis, they often mean wording, structure, or a few corrections around the edges.
That was not really the important part in my case.
The feedback that mattered most changed the shape of the work itself. Early on, it narrowed the framework. Later, it changed how that framework had to be written and presented. Those were different kinds of feedback, and they helped in different ways.
What early feedback changed
At the beginning, Vigil was much broader and more ambitious.
Coming out of the literature, I had six check types, six variation families, and three provenance domains in the model. On paper that looked rich and thorough. In practice it was becoming unmanageable. It was too much to implement cleanly, too much to explain clearly, and too much to defend without the whole thing turning into a taxonomy exercise.
The provenance side was the clearest example. My earlier model relied much more heavily on capturing information from the systems themselves. That sounds appealing in theory, but it assumes a level of visibility and feasibility that many of these systems simply do not offer. It also assumed more change and more hidden movement than was really necessary to make the framework useful.
That feedback pushed the thesis in a better direction. Instead of trying to reconstruct everything that might have happened inside a system, the framework moved toward declared variation: explicit changes to inputs, configuration, or environment that are stated up front and can be rerun later.
That shift made Vigil smaller, but it also made it more realistic. The framework stopped trying to capture the whole universe of possible behavioral change and focused on variations you can actually declare, execute, and inspect.
The real improvement was reduction
This was probably the most useful lesson of the whole process.
The thesis did not improve by becoming more comprehensive. It improved by becoming more selective.
Good feedback did not add more layers. It removed categories, reduced ambition, and forced clearer boundaries. A framework with fewer moving parts but stronger internal logic is better than one that gestures at everything and supports little of it well.
That is easy to say in hindsight and much harder to accept while you are building. Broad models feel intellectually satisfying. Smaller ones feel like compromise. But in technical work, reduction is often what turns an interesting idea into a usable one.
What later feedback changed
By the later stage, the framework itself was mostly stable. The feedback was no longer really about what Vigil is. It was about how the thesis reads.
The main tension there was style.
My instinct is usually to write in a way that feels coherent and story-like. That works well for me as a reader, and I think it often makes technical material easier to follow. But academic writing also has genre expectations, and parts of the thesis read as too narrative or too soft in that context.
What changed after that feedback was not the substance of the work. It was the presentation. Clearer headings. Clearer structure. A more visibly scientific frame around the same core argument.
That middle path was useful. I did not want the thesis to read like a dead template, but it also needed to signal more clearly what kind of document it was.
What the colloquium confirmed
The colloquium mostly confirmed that the important changes had already happened earlier.
By that point, the framework itself was stable. The questions were not really about whether the core idea made sense. They were mostly clarification questions and follow ups on the case studies. That was useful in its own way. It suggested that the central argument held up well enough that discussion moved to application rather than basic justification.
It also showed that spoken presentation exposes different weaknesses than writing. In a thesis, a reader can slow down and reconstruct a vague section. In a talk, weak transitions or vague examples become obvious immediately. So the colloquium was less a source of new conceptual feedback and more a pressure test for clarity.
That is a modest insight, but still a real one. Not all good feedback changes the work. Sometimes it confirms that the structure is now strong enough that the remaining task is simply to explain it cleanly.
Good feedback did not just decorate this thesis. It made it clearer, smaller, and stronger. And for this kind of work, that was exactly what it needed.
