--- title: "adam bien's blog" url: https://daily.dev/posts/adam-bien-s-blog-6qwj0q4p8 source_url: https://adambien.blog/roller/abien/entry/149th_airhacks_tv_htmldb_frameworks_obsolete_zero_depende type: article source: "Adam Bien" published: 2026-08-13T09:23:29.069Z updated: 2026-08-13T09:46:21.446Z tags: ["java", "ai-agents", "quarkus", "web-components", "jakarta-ee"] reading_time: 3 upvotes: 0 comments: 0 language: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # adam bien's blog **[Adam Bien](https://daily.dev/sources/adam-bien)** · 3 min read · 0 upvotes · 0 comments ## Summary A wide-ranging episode recap covering API gateways versus nginx/HAProxy, the Boundary Control Entity (BCE) architectural pattern, and the release of HTMLDB - a zero-dependency, XML-valid HTML database written in ~550 lines of Java 25, designed as agent memory storage. Also discussed: floci.io, a Quarkus-based local cloud emulator for Lambda/DynamoDB/Neptune that hit ~20k GitHub stars in 5 months; a thesis that web frameworks like React/Angular/Svelte may become obsolete since LLMs generate framework-free web component code directly; grounding LLM agents against Jakarta EE Core/MicroProfile so generated code runs on Quarkus without the agent knowing it; a 40-line zero-dependency Java 25 agent with 60ms startup; and practical notes on JWT testing, thin WARs, EclipseLink versus Hibernate, and query caching migration from EHCache to Infinispan. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is HTMLDB and how is it built in Java 25? HTMLDB is a zero-dependency database built directly on HTML that is also valid XML, requiring no external libraries. It is implemented in roughly 550 lines of Java 25 using the built-in Java XML APIs, supports defining tables and columns, adding rows via CLI, and viewing data rendered as semantic HTML in a browser. It was designed as agent memory storage since LLM agents understand XML well. _Developers experimenting with lightweight agent memory stores can follow releases like HTMLDB on daily.dev._ ### How can I ground an LLM coding agent to generate Jakarta EE code that runs on Quarkus without the agent explicitly targeting Quarkus? Ground the agent exclusively against Jakarta EE Core and MicroProfile APIs, which cover roughly 80% of typical needs using CDI, bean validation, and JAX-RS. Instructing the agent to use only these standard, minimal boring platform APIs lets the generated code run unmodified on Quarkus even though the agent has no awareness that Quarkus is the target runtime. _Teams deciding how to steer AI-generated backend code can track grounding techniques like this on daily.dev._ ### How fast can a zero-dependency Java agent start up and how small is it? A zero-dependency agent written in plain Java 25, using only a main method with an HTTP endpoint deployed on AWS agent core runtime, can be implemented in about 40 lines of code, start up in roughly 60 milliseconds, and occupy about 5 kilobytes. No external frameworks or libraries are used. _Developers weighing framework-free versus framework-based agent runtimes can follow benchmarks like this on daily.dev._ ## Similar posts on daily.dev - [adam bien's blog](https://daily.dev/posts/adam-bien-s-blog-rdt3jp0wb) · Adam Bien · 0 upvotes · 0 comments - [adam bien's blog](https://daily.dev/posts/adam-bien-s-blog-cc7smrsm2) · Jakarta EE · 0 upvotes · 0 comments - [adam bien's blog](https://daily.dev/posts/adam-bien-s-blog-zaztp0zyx) · Adam Bien · 0 upvotes · 0 comments --- Tags: [#java](https://daily.dev/tags/java), [#ai-agents](https://daily.dev/tags/ai-agents), [#quarkus](https://daily.dev/tags/quarkus), [#web-components](https://daily.dev/tags/web-components), [#jakarta-ee](https://daily.dev/tags/jakarta-ee) [View this post on daily.dev](https://daily.dev/posts/adam-bien-s-blog-6qwj0q4p8)