Decker prefers integral upscaling. On most browsers, "Decker -> Fullscreen" in the menu (or "app.fullscreen:1" from a script) will enable "stretching", but this feature is not supported by Safari on iPhones last time I tested.
Decker fonts are composed of fixed-size bitmaps. The included example "fonts.deck" has a variety of additional fonts to choose from in various sizes, and an editor:
Asking for good-faith compliance, and failing that, simply waiting for slop enthusiasts to publicly brag about using LLMs to break the rules. All evidence suggests that it is impossible for slop enthusiasts to resist outing themselves for any significant span of time.
LLMs are in essence an attack on the concept of written language, harvesting and dissolving the social contract that underpins it; it is no longer safe to assume that text has intent, let alone content, simply because it is grammatically well-formed.
Perhaps the most darkly amusing consequence of this particular mania is that by poisoning the majority of our information environment with hallucinated slop, we have likely crippled the next several generations of machine-learning techniques before they're even invented! Small, locally-hostable LLMs will rattle along spewing spam long after the broader "genai bubble" pops, and building clean training datasets will permanently be more difficult and expensive.
A Quadra 700 could run A/UX 3.0 or higher, which would make it relatively pleasant for the macs and unix workstations to interoperate (provided you spared no expense).
The articles you publish are, in principle, a celebration of the cleverness in programming and hardware design behind video games; a vehicle for appreciating human craftsmanship and artistry.
LLMs are the product of taking the creative output of countless millions of people and turning it into a statistical model which completely erases the attribution for all that material. They can regurgitate interpolated, plagiarized copies of someone else's code, divorced from its original context, but they are structurally incapable of showing where that code came from, who wrote it, or why.
Using these tools, and using your public profile to contribute to the normalization of these tools, is profoundly disrespectful to the programmers, artists, and writers whose work was strip-mined by megacorporations to build those LLMs. You are prioritizing your personal convenience while enriching bad actors who are bent on obliterating the open web, open-source culture, and the craft of programming. It's dissonant.
It's tremendously disappointing to see Fabien stoop to using slop for his research; completely antithetical to the love for software craftsmanship I had inferred from his previous work.
I'd like to know more about how this is built. A stripped-down mac emulator? Does it leverage any of the existing libraries for decoding the HyperCard stack format, or something novel?
Lisa and early MacOS are tremendously different in their details than the Alto operating system. While there was clearly a transfer of inspiration, Apple engineers like Bill Atkinson made countless small and large innovations to simplify the Xerox GUI model and improve its usability based on extensive in-house R&D and user testing (and in some cases implement features that the Apple team presumed Xerox had but actually didn't exist on the Alto). It is simply ahistoric to build narratives around Apple stealing Xerox ideas wholesale.
I think the key thing is that nothing about Decker or Lil was specifically designed or tweaked for this use case; Everything I needed to build the hershey module and make it reasonably efficient was already at my fingertips. I considered bundling up some of the patterns I describe in this article as a library, but I think it's much better to view them as transparent idioms to adapt to wherever they're needed, letting the abstraction live in your head instead of the source file.
As I wrote about two years ago in "The Benefits of Conforming", Lil's approach to handling mismatched dimensions for conforming arose from a need to plug the semantic gap left by K and Q, since Lil doesn't believe in runtime errors. While evaluating my options I made a series of pleasant idiomatic discoveries that ended up also being useful for "hershey". There is perhaps an argument that if I'd "preferred" the length of the right argument instead of the left some expressions would require fewer parentheses, but I suspect it's a wash either way.
The design of heading/mag/unit were influenced by my work with oK and iKe, particularly my annoyance at the clunkiness of writing things like {(cos x;sin x)}' or wrangling with atan2[;]. The Lil primitives help expressions "flow" together nicely, with few intermediate variables and minimal nesting. I think I spent more time agonizing over reasonably intuitive names for them than I did puzzling over semantics. I'm quite pleased with these primitives as some of Lil's hidden pearls, and I hope I can inspire other arraylangs to incorporate something similar.
The point is that https-everywhere people have successfully pushed the majority of the sites that could previously be served over vanilla http to require https anyway. Now if you're on and old device you need an https proxy or else you're out of luck.
"Accessibility" is an open-ended set of functionality, not a checkbox; it is never "complete", there is always room for improvement. Colorblind support (which ones)? High-Contrast mode? Adjustable text size? Screenreader integration? Localization? IME support? Keyboard navigation? Keyboard remapping? Functional entirely without a keyboard? Touch support? Pen support? Dyslexia-aiding typefaces? The list goes on and on.
Not especially. Depending on where you set your standards for "holding a conversation" you can satisfy the requirement with a classical markov chatterbot, a well-trained parrot, a copy of Eliza, or a telemarketer flowchart drawn on a sheet of paper. Only the markov bot is made out of "weights" in the sense of a statistical model.
Parrots are intelligent animals, albeit with a limited capacity for vocabulary and syntax compared to a human, and Eliza and the flowchart are made out of explicitly encoded rules and conversational tactics.
If you want to use a REPL as a pleasant calculator you really want an APL-family language or a modern concatenative language like Factor. Sexprs are horrendously verbose in comparison.