I'm a Technical Solutions Architect. I write about what I'm working through — AI in real applications, architecture trade-offs, and the gap between how systems are designed and how they actually behave.
The systems I build — PromptSpark, WebSpark, GitSpark, DevSpark — aren't just portfolio pieces. They're how I test whether my ideas hold up under production constraints.
Where I think out loud about architecture, AI, and what building software actually teaches you.
The Mountains of Misunderstanding map the gap between what we think we know and what we actually know — a gap that AI widens by packaging fluency as expertise. A year after writing this article, spec-driven development became my structural answer to staying off the mesa.
Getting one .NET site online behind Cloudflare is manageable. Hosting several low-traffic demonstration sites on one Windows VM to keep cost and maintenance low forced me to think less about hosting checklists and more about blast radius, boundaries, and what I would do if one of them ever outgrew this setup.
Writing the Cloudflare and IIS article made me realize I needed the same kind of governed workflow for content that I expect from code. This is how I layered write-article, critique, editorial, and SEO prompts on top of DevSpark so a rough idea could become a stronger, more publishable article through deliberate iteration.
I've been writing software since the early 1990s — through client-server, the web, cloud, and now AI-augmented development. Each wave brought new tools and new versions of the same fundamental questions about how systems should be structured, how decisions should be made, and how to know when something is working.
I started writing publicly because the act of explaining something is how I find out what I actually understand. The blog posts here are where I work through those questions — not polished conclusions, but honest explorations. Some of them turn out to be wrong. That's fine.
More about my backgroundEach system holds some constraints constant and changes others. The comparisons expose trade-offs that are hard to see from specs alone.
Spec, build, validate, evolve
DevSpark is an implementation of Adaptive System Life Cycle Development (ASLCD), a full lifecycle model for building software with AI collaborators. It carries intent from specification through implementation, validation, and evolution.
Operational prompt workflows for real applications
PromptSpark.Chat explores how prompt systems behave when they become part of an application workflow. It focuses on structured conversations, reusable prompt assets, and practical integration patterns.
Documentation that becomes implementation context
DocSpecSpark frames documentation as a working system asset instead of a static afterthought. It supports the bridge from architectural intent to implementation-ready specifications.
A controlled frontend architecture comparison
TailwindSpark is a React and Tailwind implementation that consumes shared service patterns while emphasizing composition and utility-first design. Its value is clearest when compared with BootstrapSpark against the same architectural constraints.
TailwindSpark, BootstrapSpark, PromptSpark, and supporting services make architecture visible by holding some constraints constant and changing others. The point is not to crown a framework; it is to see how complexity moves.
The shared service layer connects WebSpark, PromptSpark, GitHub Repository Spark, and markhazleton.com content. That connection turns portfolio work into a living architecture study.
Explore the Architecture