Colophon
About this archive
This site is a reconstruction. The original lived at travellingchris.blogspot.com between 2006 and 2007, with the photographs spread across Blogger and a Flickr account. Both were exported and put back together here.
What is here
182 posts, 25,467 words and 203 comments from the people following along at the time. 3,071 distinct photographs across 14 countries and 388 days.
The writing is exactly as it was posted, typos and Spanish-keyboard apostrophes included. Nothing has been rewritten, and no captions have been invented for photographs that never had any.
Where the photographs came from
- 2,724 from Flickr — the South America half of the trip, April 2007 onwards, with capture dates and tags intact.
- 334 from the Blogger album — mostly Central America, November 2006 to March 2007. This is the only surviving copy of that half of the trip.
- 13 recovered — images from the first weeks that appear in neither export, pulled back from the original 2006 image host while it is still answering.
The two archives overlap: photographs uploaded to both places appear twice, at different sizes and with no bytes in common. Those were matched perceptually and merged, keeping the better copy and everything either copy knew about it.
How the locations were worked out
Nothing in either archive records where a photograph was taken. The cameras predate GPS, and all 2,724 Flickr records have an empty geo field. The only source of location is the writing.
So each post was matched against a gazetteer of the places it mentions, and photographs embedded in a post inherit that post's location. Because those photographs carry their own capture dates, they become dated anchors across the whole year — and every remaining photograph is placed by its date against that timeline.
How well that worked:
- 761 certain — published in a post that says where it was.
- 2,205 confident — taken on a day with a firm anchor.
- 100 inferred — placed by proximity to the nearest anchored day.
- 5 unplaced — no usable date, so no location.
Where the automatic match got it wrong it was corrected by hand against what the post actually says — the pre-departure posts are London, the volcano weekend is Pacaya, and the flight home for a friend's stag weekend in April 2007 is reconstructed from the post that lists the itinerary.
What did not survive
Every image reference in all 182 posts was accounted for. 13 could not be restored:
- Eleven photographs hotlinked from other people's Flickr accounts — mostly the Death Road ride, where the tour company took the pictures. Those accounts are gone.
- One Blogger image whose URL no longer resolves.
A video embedded from Google Video is also gone; that service shut down in 2011. The two YouTube videos still play.
How it is built
Static HTML, generated from the exports by a handful of Python scripts and Astro. The map is inline SVG projected at build time from public-domain Natural Earth data — there is no tile server and no mapping library, so it cannot break when someone else's CDN goes away. Photographs are served as WebP in three sizes, with the untouched originals downloadable from every photograph in the viewer.