Re: unboxed structure fields (or class slots)

> On 8 Jan 2026, at 14:47, Tim Bradshaw <[email protected]> wrote:
> 
> I can't imagine writing a construct like this other than in Lisp: sadly I can really imagine how utterly miserable this makes most programmer's lives, because I have been those people.  They're basically mediæval serfs toiling in the mud with no understanding that there could ever be or ever was a world without mud and lice.  Occasionally they dig up shards of a strangely shiny metal, too hard to beat into tools: fragments of a long-forgotten flying city.  Curiously, those flying cities still exist: they just can't see them.

Well, I have a thing or two to say about that too.

The current inhabitants of the once great civilisation are a scattered mix: researchers using it for specific problems, hobbyists fascinated by its design, a few companies maintaining legacy systems or using it for niche applications where its strengths matter. They pick through the ruins, use what works, build makeshift extensions where the ancient builders left gaps (like binary I/O), and occasionally marvel at both the brilliance and the strangeness of what was left behind.
The treasure hunters are particularly apt - people coming from Python or JavaScript, poking around, finding genuinely powerful ideas (macros, conditions, interactive development), taking those ideas back to their own languages. Rust's error handling owes something to CL conditions. Python's object system shows CLOS influence. The ideas get plundered and adapted elsewhere.

Meanwhile the ruins themselves sit there, still functional, still inhabited by those who appreciate them, but clearly not the thriving metropolis they were designed for. No one's building new wings onto the cathedral. They're just keeping the lights on in the parts that still matter to them.

It's a melancholy image, but not entirely negative. Ruins can be beautiful, and sometimes they outlast the more fashionable cities that replaced them.