<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>Dipperstein</title>
  <link>https://dipperstein.com/</link>
  <description>Essays, experiments, and code from a working engineer</description>
  <language>en</language>
  <lastBuildDate>Mon, 09 Nov 2026 10:14:35 GMT</lastBuildDate>
  <atom:link href="https://dipperstein.com/feed.xml" rel="self" type="application/rss+xml"/>
  <item>
    <title>The 10:14 A.M. Consensus: A Month of Robot Calls, Sorted by Hour</title>
    <link>https://dipperstein.com/the-1014-am-consensus-month-of-robot-calls/</link>
    <guid isPermaLink="true">https://dipperstein.com/the-1014-am-consensus-month-of-robot-calls/</guid>
    <description>For 31 days I pointed my home number at a cheap SIP trunk and treated every ring as a row in a dataset. The area codes were noisy and the STIR/SHAKEN headers were only somewhat useful, but the clock turned out to be the best robot detector I had.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Mon, 09 Nov 2026 08:19:07 GMT</pubDate>
  </item>
  <item>
    <title>A Cheap Rain Gauge, a Month of Bucket Tips, and the Limits of a Trusted Number</title>
    <link>https://dipperstein.com/cheap-rain-gauge-month-bucket-tips-limits-trusted-number/</link>
    <guid isPermaLink="true">https://dipperstein.com/cheap-rain-gauge-month-bucket-tips-limits-trusted-number/</guid>
    <description>A twenty-dollar tipping-bucket gauge, a Raspberry Pi, and a month of Prometheus logs made for a tidy little telemetry project, until the numbers ran low against the nearest National Weather Service station. The fix was part statistics, part plumbing, and mostly a lesson in how much trust we hand to a sensor.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Mon, 09 Nov 2026 04:41:11 GMT</pubDate>
  </item>
  <item>
    <title>When the Sun Is Your Sysadmin: Thirty Days of Off-Grid CI</title>
    <link>https://dipperstein.com/when-the-sun-is-your-sysadmin-thirty-days-of-off-grid-ci/</link>
    <guid isPermaLink="true">https://dipperstein.com/when-the-sun-is-your-sysadmin-thirty-days-of-off-grid-ci/</guid>
    <description>I moved my self-hosted CI runner onto a solar panel and a Raspberry Pi for a month. The builds were slower, the failures were stranger, and the lessons about queueing, backoff, and the carbon cost of free compute were worth every brownout.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Sun, 08 Nov 2026 08:13:51 GMT</pubDate>
  </item>
  <item>
    <title>One Star, Nine Names: A Resolver for Betelgeuse and for Your API</title>
    <link>https://dipperstein.com/one-star-nine-names-betelgeuse-catalog-identity/</link>
    <guid isPermaLink="true">https://dipperstein.com/one-star-nine-names-betelgeuse-catalog-identity/</guid>
    <description>Betelgeuse answers to at least nine identifiers across the catalogs we&#039;ve built over four centuries, and not one of them is the &quot;real&quot; one. I wrote a small resolver to stitch them together, then spent a week staring at what it taught me about primary keys, aliases, and deprecation in the software I ship.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Sun, 08 Nov 2026 00:15:21 GMT</pubDate>
  </item>
  <item>
    <title>My Code Reviews Were Mostly Vibes: A 2,300-Comment Autopsy</title>
    <link>https://dipperstein.com/my-code-reviews-were-mostly-vibes-comment-autopsy/</link>
    <guid isPermaLink="true">https://dipperstein.com/my-code-reviews-were-mostly-vibes-comment-autopsy/</guid>
    <description>I pulled every code review comment I had left over several years and sorted them by what I was actually trying to accomplish. The results were less about bugs and more about my personality with a keyboard. Here is the dataset, the methodology mistakes, and the rubric I now use before I hit comment.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Sat, 07 Nov 2026 20:15:46 GMT</pubDate>
  </item>
  <item>
    <title>The Leap Second Is a Legacy System, So I Broke My Homelab on Purpose</title>
    <link>https://dipperstein.com/leap-second-legacy-system-homelab-fire-drill/</link>
    <guid isPermaLink="true">https://dipperstein.com/leap-second-legacy-system-homelab-fire-drill/</guid>
    <description>I simulated a leap second across my home lab to see which assumptions snapped first. Cron jobs duplicated, logs arrived out of order, and one alert fired twice. The lesson: UTC is not a trustworthy API, it is a distributed system with sharp edges.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Sat, 07 Nov 2026 16:14:19 GMT</pubDate>
  </item>
  <item>
    <title>The Night Sky Has No Search Bar: A 2 A.M. Wayfinding Experiment</title>
    <link>https://dipperstein.com/night-sky-no-search-bar-app-wayfinding-experiment/</link>
    <guid isPermaLink="true">https://dipperstein.com/night-sky-no-search-bar-app-wayfinding-experiment/</guid>
    <description>The night sky is the oldest navigation system we have, and it works because it offers a few reliable landmarks and a lot of empty space. I spent two weeks opening apps at 2 a.m. to see how they compare. Most of them failed.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Sat, 07 Nov 2026 12:09:08 GMT</pubDate>
  </item>
  <item>
    <title>The Kettle, the Ladle, and the Compiler: A Soaking-Hot Tour of Type Strictness</title>
    <link>https://dipperstein.com/kettle-ladle-compiler-type-strictness/</link>
    <guid isPermaLink="true">https://dipperstein.com/kettle-ladle-compiler-type-strictness/</guid>
    <description>I spent a week dipping things into hot water and thinking about type systems, and I&#039;m only slightly sorry. Here&#039;s what a candy thermometer, a soup ladle, and a very cautious toe can teach you about JavaScript, TypeScript, and Rust.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Sat, 07 Nov 2026 08:10:22 GMT</pubDate>
  </item>
  <item>
    <title>My Well Pump Is the Worst Database I&#039;ve Ever Administered</title>
    <link>https://dipperstein.com/well-pump-worst-database-capacity-planning/</link>
    <guid isPermaLink="true">https://dipperstein.com/well-pump-worst-database-capacity-planning/</guid>
    <description>I strapped cheap sensors to a residential well in the Midwest and spent a year watching an aquifer do capacity planning with zero elasticity and no rollback. Turns out groundwater is just a database with a very unforgiving SLA. Here&#039;s the monitoring rig, the data, and what it taught me about provisioning everything else.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Fri, 06 Nov 2026 20:08:43 GMT</pubDate>
  </item>
  <item>
    <title>Before You Refactor That Monolith, Thank a 1947 Warehouse Manager</title>
    <link>https://dipperstein.com/before-you-refactor-that-monolith-thank-a-1947-warehouse-manager/</link>
    <guid isPermaLink="true">https://dipperstein.com/before-you-refactor-that-monolith-thank-a-1947-warehouse-manager/</guid>
    <description>The metaphor behind &#039;technical debt&#039; didn&#039;t come from software — it came from postwar American logistics, and it&#039;s quietly steering your sprint planning. I traced the lineage, then ran a six-week experiment on my own repo to find out which debt payments actually compound.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Fri, 06 Nov 2026 16:08:28 GMT</pubDate>
  </item>
  <item>
    <title>The Big Dipper Doesn&#039;t Move, But My Tripod Sure Did: A Star Tracker Built Out of Junk Drawer Physics</title>
    <link>https://dipperstein.com/star-tracker-from-spare-parts/</link>
    <guid isPermaLink="true">https://dipperstein.com/star-tracker-from-spare-parts/</guid>
    <description>Armed with a $12 webcam, a 3D printer, and an unreasonable amount of confidence, I set out to build a star tracker that could find true north without a GPS chip. Here&#039;s what went right, what melted, and why the night sky is the most patient debugging partner I&#039;ve ever had.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Fri, 06 Nov 2026 08:14:22 GMT</pubDate>
  </item>
  <item>
    <title>Burying a Library With Dignity: Confessions of an Open-Source Undertaker</title>
    <link>https://dipperstein.com/burying-a-library-with-dignity-open-source-undertaker/</link>
    <guid isPermaLink="true">https://dipperstein.com/burying-a-library-with-dignity-open-source-undertaker/</guid>
    <description>I adopted a dependency with three users, two of whom were me, and spent six months figuring out how to give it a good death. Here&#039;s what I learned about code archaeology, writing migration docs for an audience of zero, and why a farewell release is a kindness, not a formality.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Fri, 06 Nov 2026 04:02:05 GMT</pubDate>
  </item>
  <item>
    <title>Open Mic Night at the Compiler: Why Programming Languages Act Like Local Music Scenes</title>
    <link>https://dipperstein.com/open-mic-compiler-languages-as-music-scenes/</link>
    <guid isPermaLink="true">https://dipperstein.com/open-mic-compiler-languages-as-music-scenes/</guid>
    <description>Every language community has its house style, its gatekeepers, and its one-hit wonders — just like the local music scene. I spent months digging through forum archives and talking to developers to see whether &#039;scene dynamics&#039; can actually predict which languages break big and which never leave the coffeehouse.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Thu, 05 Nov 2026 04:02:41 GMT</pubDate>
  </item>
  <item>
    <title>I Ghosted My Own Codebase for 90 Days: Notes From a One-Man Bus Factor Experiment</title>
    <link>https://dipperstein.com/ghosted-my-own-codebase-bus-factor-experiment/</link>
    <guid isPermaLink="true">https://dipperstein.com/ghosted-my-own-codebase-bus-factor-experiment/</guid>
    <description>I spent a weekend writing handover docs for a side project nobody else uses, then deliberately stopped touching it for three months. Here&#039;s what rotted, what survived, and the checklist I now run on everything I ship.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Tue, 03 Nov 2026 08:02:32 GMT</pubDate>
  </item>
  <item>
    <title>Your Living Room Is Talking Behind Your Back: A Packet-Capture Tour of the American Smart Home</title>
    <link>https://dipperstein.com/smart-home-network-teardown-packet-capture/</link>
    <guid isPermaLink="true">https://dipperstein.com/smart-home-network-teardown-packet-capture/</guid>
    <description>I put a mirror port on my home network and logged every DNS query and packet my thermostat, TV, and light bulbs emitted for two weeks. The results, ranked by data volume, surprised me — and led to a firewall ruleset you can steal.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Tue, 03 Nov 2026 08:02:32 GMT</pubDate>
  </item>
  <item>
    <title>A Year of Release Notes From the Software Nobody Brags About</title>
    <link>https://dipperstein.com/year-of-unglamorous-release-notes/</link>
    <guid isPermaLink="true">https://dipperstein.com/year-of-unglamorous-release-notes/</guid>
    <description>For twelve months I read every changelog entry from a password manager, a PDF reader, a terminal emulator, a notes app, and a small database. What I found wasn&#039;t feature news — it was a working portrait of what software maintenance actually is, warts and all.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Tue, 03 Nov 2026 04:04:47 GMT</pubDate>
  </item>
  <item>
    <title>Every Millisecond Has a Price Tag: Running the Numbers on Slow Software</title>
    <link>https://dipperstein.com/every-millisecond-has-a-price-tag/</link>
    <guid isPermaLink="true">https://dipperstein.com/every-millisecond-has-a-price-tag/</guid>
    <description>I spent a month injecting artificial delays into my side project to find out exactly what slow responses cost. The numbers between 200ms and 2 seconds are starker than most engineering roadmaps admit — and the framework I built afterward tells you which endpoints are worth fixing.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Tue, 03 Nov 2026 04:04:47 GMT</pubDate>
  </item>
  <item>
    <title>We Log Every Request Except Our Own Lives: Notes on Personal Telemetry</title>
    <link>https://dipperstein.com/personal-telemetry-for-engineers/</link>
    <guid isPermaLink="true">https://dipperstein.com/personal-telemetry-for-engineers/</guid>
    <description>We instrument production to the millisecond but fly blind on our own habits and career choices. After six months of running a handful of personal counters and a Friday review, here&#039;s what the data actually changed — and what it didn&#039;t.</description>
    <author>Priya Raghavan</author>
    <category>Deep Dives</category>
    <pubDate>Tue, 03 Nov 2026 00:03:41 GMT</pubDate>
  </item>
  <item>
    <title>Ten Thousand Packages, Five Surprises: A Random Walk Through the npm Registry</title>
    <link>https://dipperstein.com/npm-registry-random-sample-analysis/</link>
    <guid isPermaLink="true">https://dipperstein.com/npm-registry-random-sample-analysis/</guid>
    <description>I pulled a uniform random sample of 10,000 packages from the npm registry and ran five passes over the data, each one surfacing something I didn&#039;t expect. Full Python notebook included so you can rerun every dip against live registry data.</description>
    <author>Michael Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Tue, 03 Nov 2026 00:03:41 GMT</pubDate>
  </item>
  <item>
    <title>I Turned Myself Into a Dashboard for 30 Days and the Data Was Humbling</title>
    <link>https://dipperstein.com/turned-myself-into-a-dashboard-for-30-days/</link>
    <guid isPermaLink="true">https://dipperstein.com/turned-myself-into-a-dashboard-for-30-days/</guid>
    <description>I instrumented my own workday like production telemetry — commits, context switches, meeting load, coffee timing — and pointed Grafana at my life. The bottlenecks I found weren&#039;t the ones I would have guessed.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Mon, 02 Nov 2026 16:04:14 GMT</pubDate>
  </item>
  <item>
    <title>Design for the Engineer Who Got Woken Up: A Field Guide to 2 A.M.-Proof Systems</title>
    <link>https://dipperstein.com/design-for-the-engineer-who-got-woken-up/</link>
    <guid isPermaLink="true">https://dipperstein.com/design-for-the-engineer-who-got-woken-up/</guid>
    <description>Most of our best practices are written by well-rested people for well-rested people. This piece argues that the real quality gate is whether a sleep-deprived responder can make sense of your failure at 2 A.M. — with a checklist to test your own systems against.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Mon, 02 Nov 2026 16:04:14 GMT</pubDate>
  </item>
  <item>
    <title>Slow Code, Long Shadows: Rebuilding a 1960 Algorithm From a Scanned Paper</title>
    <link>https://dipperstein.com/slow-code-long-shadows-rebuilding-1960-algorithm/</link>
    <guid isPermaLink="true">https://dipperstein.com/slow-code-long-shadows-rebuilding-1960-algorithm/</guid>
    <description>I found my grandfather&#039;s algorithm papers from Saarland and spent a month re-implementing one of them from scratch. What started as a sentimental project turned into the best argument I&#039;ve made to myself against shipping fast and skimming shallow.</description>
    <author>Dipperstein</author>
    <category>Deep Dives</category>
    <pubDate>Mon, 02 Nov 2026 12:04:46 GMT</pubDate>
  </item>
  <item>
    <title>My Lab Notebook of Side Projects: 7 Wins, 3 Craters, and What Each One Taught Me</title>
    <link>https://dipperstein.com/lab-notebook-side-projects-wins-and-craters/</link>
    <guid isPermaLink="true">https://dipperstein.com/lab-notebook-side-projects-wins-and-craters/</guid>
    <description>Over three years I ran ten experiments on nights and weekends, from load-testing a homelab to building a toy database that never worked. Here&#039;s the honest scorecard, with the surprising results and the lessons worth stealing.</description>
    <author>Dipperstein</author>
    <category>Experiments</category>
    <pubDate>Mon, 02 Nov 2026 12:04:46 GMT</pubDate>
  </item>
</channel>
</rss>