Finishing the 2020 site restoration

Wednesday 23 June, 2021 at 1.00pm

  • tagged
  • Back in 2020 I wrote an article about starting a restoration of my website. At the time I was fairly confident that this would not take more than a few weeks. And yet, I am writing this summary having finished importing all the content I wanted before I considered it "done" a whole year later.

    I can sense something of a pattern in my meta-posts about my site. The very first “about this site” page I wrote in 2005 said:

    This site is still under construction, and this page will continue to be updated as this site evolves. The styles are still a little rough right now as I tie down the code [...]

    For the last year I've had a banner on the site that said "Please hold… rebuild in progress" that linked to an explanation as my intention was to have it done in a few weeks. Two things derailed that entirely. Firstly, I spent two weeks suffering from vertigo and the summer heat, and then I pivoted almost entirely from writing words and code to making YouTube videos for six months.

    The last imported content

    Apart from one or two things that I'm not yet able to import to the site, the very last thing to import was the original content. I put that off until the end as I knew it would be mostly manual editing, rather than being able to write and tweak a script that did it all, and be done.

    To start with, I decanted the posts stored in a database backup from the last version of this site. v2, circa October 2011 v1, circa September 2009

    Some of them were written in Markdown, which made the process of editing them to work for this site easier. But many were written in a format called textframe, an old side project of mine to create a lightweight markup language that would support many more HTML elements. For those I converted the output HTML into Markdown using markdownify, then compared that to the original source, correcting as necessary.

    I then did a pass comparing these restored versions to those recorded in the Wayback Machine for the site, checking links and rewriting them in my current posts to the Wayback Machine from that period in time if they had subsequently broken or no longer represented the content I was referring to.

    By going back far enough, I then realised that when I moved my content from being on to I had either missed or deliberately left out some of my earliest content. So I recovered those from a much older backup I had kept. v2, circa August 2007 v1, circa November 2005

    A few observations: