Index  Comments

As with last year, and that before it, I write about the preceding year and reflect on it. Compared to 2020, I found 2021 oddly unfulfilling in the usual ways, albeit more fulfilling in some new ways.

Much of my future work is reliant on a backlog of other work, preventing me from releasing any of it lately. The last was that first year in which I wrote less often than that preceding year, but this was inevitable sooner or later. I realized that failing my goals for writing and programming didn't bother me, and shouldn't, because absolutely none of it has imparted lasting joy unto me anyway, and so there's no reason to feel grief about that which wouldn't've made me happy regardless. I so well recall the joy of finishing my latest CHIP-8 MMC in 2020, and how it had died within days. Oddly, I still feel joy relating to something I wrote that year, not published on my website or with my name, and so seemingly insignificant in comparison; it was among other people, which is what makes it odd.

I successfully avoided allowing the last two months of the year to dominate my writing done therein.

I resumed my Latin studies with the last year, after roughly seven years, and it has given unto me a joy which I believe is truly lasting. This joy is greater than any I've ever received from using my machines; I understand increasingly more of this world around me, seeing through what was previously opaque, and I'm slowly becoming able to think in a language that isn't English. For many years, I'd thought English was a fine lady, but with this year I've seen her for what she's: a whore whose best qualities come from wearing the clothing of better women. I'm amazed by how humans can communicate.

I've in general had a nice time doing many new things with the past year that I'd never done before.

Rather than continue to focus on the MMC, which is at least in a usable state, I spent the last year focussing rather exclusively on my language modelling, Elision. The English Elision is of a similar complexity to the toki pona Elision toy, in design, but it's with a Latin Elision that I realized an extension of the idea that truly makes Elision a high-level language model suitable for programmatic and symbolic manipulation of language; whereas the other design is still better than that character-by-character model of text, and I still intend to pursue it, it's simply not as nice, but that Latin is a dead language does help. I've been unable to find a dictionary organization with an exhaustive list of all declensions and conjugations in English, although I've also not bothered to search well.

The only thing stopping me from making progress on the English Elision is sloth, and the issues with the dictionary meaning initial work will be closer to compression than a language model, but this is necessary for progress. The only thing stopping me from making progress on the Latin Elision is the immaturity of my grasp on the language, which I fear could damn any early version through an unknown requirement, and which has me think starting soon to be foolish. For both cases, I've the necessary algorithms and tables in mind, and constant thought about this has relieved me from needing to write down these ideas, for the most part. There's so much I can do once these prerequisites be in place.

Thinking about Elision so, for which I want to have tables storing all that's interesting, with only simple algorithms to compose them in ways of which a finite function is incapable, has influenced my thoughts about programming in general. I increasingly grow to view computation as disgusting, where it be unnecessary. I struggle to consider ways in which partially or entirely listing those answers of an algorithm in a table to be referenced be inferior to the method of using code for such. I see code increasingly as undesirable. Recently, I was considering many different ways to model the mode of a program, each with good or bad qualities, and at first considered that more context were needed to determine which would be better in a scenario, but decided whichever decreases code most be best.

I've struggled with what is likely tendinitis this last year; I attempted to mitigate it by learning how to use a chording keyboard, and see it as an ideal for data entry and light typing, but realized singlehandedly typing everything with my left hand was placing an unacceptable strain on it. Rather than purchase a second chording keyboard, I bought a nice model of mechanical keyboard, and found it reasonably nice. I've yet to find a voice-or-sound-to-text system I may use for the brunt of typing I do, and I've lately reconsidered purchasing a second chording keyboard. This is certainly part of the reason why I wrote less this year, but isn't the whole of it. More recently, I've a new bed, as its predecessor was decades old, and I believe this will help. Currently, the office chair I use is cheaper than the keyboards I've purchased, individually, so that's something to replace rather soon.

I expect the coming year to be better than the last, and perhaps I'll finish some nice work therein.