The old blog is back: 26 articles, with their old addresses and the images recovered alongside them.
Among them is a post about building a Lego case for a Raspberry Pi. The proper case arrived in the post by the time I finished writing about it. There is also a 2015 post about making shapes with my daughter, and another about how I thought work should be organised.
They are small pieces of an older website. Getting them back meant finding the pages, separating the articles from the surrounding WordPress machinery, and giving them somewhere to live again.
Find out what survives
The recovery record from 18 July 2026 contains 2,819 successful Wayback captures covering 404 unique resources. That does not mean there were 404 articles. The list includes images, category pages, tags and other addresses WordPress generated around the writing.
There were 27 original article URLs. After reconciling a duplicate, those became 26 canonical articles: one preferred address for each piece.
That distinction matters when recovering an old site. A long list of captured URLs can make the job look more complete than it is. The first useful document is an inventory: which addresses contain writing, which contain its images, and which are alternate routes to the same thing?
If you are doing this for your own site, begin with a few articles you remember. Check their individual captures and follow their image addresses. Then expand to the rest of the inventory. A working homepage is only one recovered page.
Recover the article and its attachments
The recovery script extracts the article body, removes the old scripts, forms and sharing controls, and passes the remaining HTML through an allowlist. It rewrites recovered local links and image addresses so reading the restored site does not depend on moving through the Wayback viewer.
The July recovery recorded 40 downloaded media files. The original upload paths are retained under /wp-content/uploads/.
The 2015 Lego post is a useful example. Its triangle photograph belongs with the description of how it was made. Recovering only the words would leave a fairly important part of that post missing.
There is a limit to what this process can retrieve. The Internet Archive explains that an archived page may have missing images and that coverage is not guaranteed. A capture is something to inspect, not proof that every linked file survived. Its Wayback help page explains how to look up individual URLs and their available captures.
This recovery rebuilt the public articles as static pages. It did not recreate the WordPress database, administration area or plugins. There is no need for those things to serve an old piece of writing.
Keep the addresses
One article about MySQL’s FIND_IN_SET appeared under two dated paths in the archive:
/2015/04/08/find_in_set-handy-mysql-function/
redirects to
/2013/04/08/find_in_set-handy-mysql-function/The restored site keeps the 2013 address as the canonical page and redirects the duplicate. The rest of the recovered articles keep their historical paths too.
A reader arriving through an old link should reach the article that link referred to. Redirecting every missing address to the homepage would leave them to work out what had happened.
Some of the old WordPress listing pages have relevant redirects. Preserving the articles does not require recreating every way the old theme listed them.
Keep the dates honest
The old posts carry their original publication dates. An archive note explains that links, products and technical advice may now be out of date.
The recovery makes the writing readable again; it does not make old technical advice current. A post about a database function in 2013 should still look like a post from 2013.
New writing belongs beside that archive with a date of its own. The events in a recollection may have happened years before the piece was written.
A manageable recovery job
For another small personal site, I would break the work into six passes:
- List the old article addresses and check available captures.
- Record which images and other files each article needs.
- Extract the available content and remove obsolete page machinery.
- Resolve duplicate addresses and keep a redirect map.
- Check the restored pages, image files and internal links.
- Label the archive and preserve the dates of the original writing.
The resulting website can be much simpler than the one that disappeared. This one serves static pages. The Raspberry Pi case is still there, including the part where the purchased case made it unnecessary.
The old Lego article also prompted a new six-brick programming activity. Recovering a post gives you something to respond to.