Archive note, January 30, 2025: This post is based on the recovery work used to retrieve articles from the defunct InlandEmpire.com archive.
There is a strange kind of loss that happens when a site disappears.
The work is still real. The bylines were real. The shows happened. The photos, deadlines, venue notes, and late-night drafts all existed. But if the site goes dark and the database is gone, the public record starts to dissolve.
That is what made the Wayback recovery work worth doing.
The goal was simple: recover as much of my old Inland Empire writing as possible and preserve it in a form that could be reviewed, archived, and eventually republished under my own control.
The Recovery Problem
The original site was not available as a normal export. There was no friendly WordPress XML file waiting to be downloaded. The remaining path was forensic:
- Search archived snapshots.
- Identify article URLs.
- Filter for pages with my byline.
- Extract the article body.
- Preserve dates, titles, source URLs, and archive URLs.
- Save the result in a format that could survive the next platform change.
The Wayback Machine became the source of truth, but it was not a clean source. Some snapshots had missing content. Some pages captured navigation but not the article. Some URLs existed in search results but were not captured by any usable archive.
That meant the scraper had to be conservative. A page with a title but no article body was not a recovered article. A page with a short fragment was not enough. The recovery script looked for meaningful content, author signals, original URLs, snapshot dates, and enough text to justify keeping the result.
Why JSON Mattered
The useful output was not just copied text.
Each recovered article was saved as structured JSON with the original URL, Wayback URL, title, author, publication date, HTML content, plain text content, and image references. A plain text version was also generated for review.
That structure mattered because the immediate goal was recovery, not instant republication. Once the article was in JSON, it could be inspected, transformed, imported, or ignored. It was no longer trapped inside one archive snapshot.
The importer was designed to create WordPress drafts rather than publish immediately. That was the right default. Recovered content needs review. Images may be broken. Formatting may be odd. Dates need validation. Categories and tags need human judgment.
Automation did the shovel work. Editorial control stayed human.
What Was Recovered
The cleanest recovery set included theater, music, and event writing from 2019 through 2023:
- Musical Theatre West reviews.
- Broadway at the Fox coverage.
- A Flogging Molly concert review.
- A Paranormal Circus article.
- A Riverside Lunar New Year preview.
Not everything survived. Some articles were visible through search traces but missing from the archive. Others likely existed but were not captured by crawlers. That is frustrating, but it is also clarifying. If the work matters, relying on someone else’s database is not an archival strategy.
The Lesson
This was not a glamorous engineering project. It was digital archaeology with Python, BeautifulSoup, the Wayback CDX API, and patience.
But it changed how I think about publishing.
Owning a domain is not the same as owning the record. Owning the CMS is not the same as having a durable archive. A real archive needs portable content, structured metadata, source references, and a way to rebuild when the public surface changes.
The recovered articles have historic value more than commercial value. That is enough. They are part of the path that led to SoCalNomad, the photography work, and this habit of treating project history as something worth preserving.