Miscellany, Vol. 4

It’s been an exciting couple of months since the last instalment. I added a dark colour scheme, improved the article headers, cleverly lost some historical data, perfected image captions for all time, moved to self-hosted builds, sped up broken-links-inspector by not using it, and demoted the AVIF image format. At the same time, I’ve kept up the smaller, less obvious improvements to A Place For My Head:

• I changed the presentation of inline ins and del elements to clarify their meanings. While I don’t think they look very good, this was a necessary change. I experimented with putting the insertion date inline too, but it only further cluttered an already crowded block of text.

• I further improved the metadata in the HTML across the site and various descriptions of pages as well as the site overall.

• It turned out I’d been linking to non-HTTPS IMDb pages. This shocking lapse in security has been rectified. I apologize to all the IMDb entries I may have inadvertently offended.

• In a mild course correction, I slightly increased the space around h2s and images.

• I added support for crediting multiple authors with different labels, as in my thoughts on I Kill Giants.

• Speaking of thoughts, I replaced the line saying ‘Previously read’ (or ‘watched’) with a parenthetical under ‘Finished reading’ (or ‘watching’).

• code, samp, and kbd had some odd wrapping behaviour. I fine-tuned this for differently sized screens. To be safe, I allow breaking the line at any point in narrower viewports, preventing horizontal overflow.

• In the same vein as preferring AVIF, I’ve put all the use of custom fonts inside a prefers-reduced-data media query. This likewise won’t have any effect until the query is supported by browsers.

• In lists like those on the Thoughts & Spoilers page, the text indicating when I finished something had a title, and therefore tooltip, with the full date and time in ISO 8601 format. It struck me that the time was redundant: I only record the date I finished the work, so the time was always inferred as 00:00:00.000+00:00.

  I spent 10 minutes playing with the format and eventually tried ‘Jan 28, 2022’. It bothered me that this didn’t match the other dates around it, so I changed it to ‘28th January, 2022’, and proceeded to admire my handiwork: now anyone could point the mouse at the line containing the date and, if they were patient, they would be rewarded with… a tooltip showing almost exactly the same text, sans superscript. After a few moments of giggling and contemplating the advisability of my ever writing another line of code, I removed the title.

• I limited the Atom feeds to only the last 50 entries, newest first, preferring my updated field to date when available for the purposes of both sorting entries and producing the XML. I’ll further reduce the number to 20 at some point.

• Articles in a series have sprouted a link to the next one at the bottom. As with my article headers, although I resisted adding anything of the sort for a long time, starting my new series on ‘The Death and Rebirth of a Cluster’ led me to examine how I expected it to be read, prompting this change.

• In a less controversial development, requiring little internal debate, I added Atom feeds for each individual series.

• What’s more, pages can now point to all relevant feeds in the HTML, not just the full feed.

• I tried a few variations on my blockquotes: removing the border, changing the colour to match the text, using quotation marks, and combining the border and quotation marks like you see now. This style is excessive too, but my primary motivation is demarcating quoted text in complicated scenarios.

On a general note, I’m happy to report that, in keeping with my promise, I’ve been publishing more. Much more. I’ve already added 106 entries in the past two and a half months, reducing my Thoughts & Spoilers backlog from 14 months to a mere ten. I’m generally writing longer entries there, too, and starting to learn how to weave my disorganized, reactionary notes into comprehensible prose that—I optimistically assume—is significantly easier to read than what I would have produced nearly two years ago.

At the same time, I’ve improved my technical articles—again, I assume—by making them smaller and more focused. The newest series includes several especially short entries. I’m not entirely comfortable with that, but there’s a vast amount of material to cover, and the only way it’ll be of any use to either myself or anyone else is if I ruthlessly prune and segment it.

I hope I’m able to keep up the pace. There’s still so much more to tell you about, like how I set up auto-archival for this site but it doesn’t work properly yet.

Next in series: Splitting Lines of Code with rehype(#37 in Colophon: Finding A Place For My Head)