Weeknotes #31

Last week’s entry was so late because Monday was Mountain Day and I lost track of time. Mountain Day…? That sounds like a made up holiday that’s just an excuse to avoid going into work in the middle of summer. Yes.

I continued to spend huge amounts of time on my iOS app. I’ve now implemented almost all functionality: it’s possible to reorder processors, enable processors, add custom processors, remove custom processors and set options for processors that support them. There’s now also basic persistence of user preferences.

I’m now at the stage where I’m trying to design an icon. A while back, I decided to end my subscription to Photoshop and make do with indie image editors like Acorn and Pixelmator. Until now, that’s worked fine but designing an app icon has been an exercise in acute frustration. Photoshop has a feature that lets you create a gradient based on the selected foreground and background colours. Acorn and Pixelmator both have gradient tools and both have the concept of foreground and background colours but neither has this—at least in my eyes—rudimentary feature.

It was published back at the beginning of the month, but the best thing I read this week was Eugene Wei’s terrific post on TikTok and its algorithm. I’ve never used the app but reading Wei’s description makes it sound like the Chinese (or at least the app’s parent, ByteDance) have finally realised the promise of algorithmic timelines. Algorithmic timelines have developed a reputation in the West as an anti-pattern; a trick social networks use to prey on the human weakness for random displays of random novelty. Their use for that purpose is detestable but I’ve long been frustrated by the apparent impossibility of training them to better perform their ostensible goal of picking content that’s of interest to you (perhaps because consistently displaying interesting results wouldn’t have the same addictive quality?). TikTok’s success indicates that it’s definitely possible. Perhaps it’s easier because it’s video, although by all reports ByteDance first cracked this nut with their news app, Toutiao, and, when it comes to crappy examples of algorithmic timelines, one of the first that comes to mind is YouTube.

That blog post also contains this amazing backhanded compliment buried in a footnote:

In the West, Facebook is the master of the fast follow. They struggle to launch new social graphs of their own invention, but if they spot any competing social network achieve any level of traction, they will lock down and ship a clone with blinding speed. Good artists borrow, great artists steal, the best artists steal the most quickly? Facebook as a competitor reminds me of that class of zombies in movies that stagger around drunk most of the time, but the moment they spot a target, they sprint at it like a pack of cheetahs. The type you see in 28 Days Later and I Am Legend. Terrifying.

(Spoilers for the end of the first season of Perry Mason. Skip to the next paragraph if you want to avoid those.) I finished the end of Perry Mason and was left a little disappointed. The show eventually engaged in that most detested of premium cable tricks: the subverted audience expectation. In this case, it was Mason asking his estranged former partner, Pete Strickland, for help just before the end of the trial that’s been the focus of season one. You’re of course expecting that it’s going to be Strickland digging up the evidence Mason needs to pin the murder on Joe Ennis, the crooked cop. This is 2020, though, so of course that’s not it. Instead, the reveal at the end of the episode is that Mason paid Strickland to bribe one of the jurors to hang the jury. Ha! Mason’s not so heroic after all. Expectation subverted. Expect even that’s not enough and this subversion is itself subverted almost immediately when it’s further revealed that Mason had in fact successfully hung the jury, convincing two other jurors with the force of his arguments to resolutely insist on acquittal. I’m sure this seemed very clever in the writers’ room but it doesn’t make for a very satisfying conclusion for those of us who were watching Perry Mason in the hopes that Perry Mason would, you know, Perry Mason some shit. Maybe in season two.

I usually don’t listen to songs on repeat; the exception is when I’m programming. My background track of choice this week as I’ve been immersed in Xcode was the Âme Remix of Howling’s 2012 self-titled track, ‘Howling’ (Apple Music).