Hacker News new | more | comments | ask | show | jobs | submit login
Scheme 9 from Empty Space, Reimagined (t3x.org)
107 points by nils-m-holm 4 months ago | hide | past | web | favorite | 26 comments

The site is sincerely refreshing. I feel like I'm walking into a calm forest, compared to the din of most any other website (HN would probably feel the same if it were dark-themed).

I like the minimalism, but not the colour-scheme. I far prefer dark-on-light.

HN itself has it about right - minimalistic and easy to read.

Your preference is backed by science:

> However, most studies have shown that dark characters on a light background are superior to light characters on a dark background (when the refresh rate is fairly high). For example, Bauer and Cavonius (1980) found that participants were 26% more accurate in reading text when they read it with dark characters on a light background. Reference: Bauer, D., & Cavonius, C., R. (1980). Improving the legibility of visual display units through contrast reversal. In E. Grandjean, E. Vigliani (Eds.), Ergonomic Aspects of Visual Display Terminals (pp. 137-142). London: Taylor & Francis


There is a reason why most sight-impaired people, at least those I know, prefer light-on-dark. A bright background blurs out the foreground and makes reading hard. Dark-on-light only makes sense when switching between paper and screen frequently.

To me, dark-on-light becomes increasingly hard to read as I grow older. So much so that I have disabled CSS in my browser entirely.

Like almost half the globe I have astigmatism:

> People with astigmatism (aproximately 50% of the population) find it harder to read white text on black than black text on white. Part of this has to do with light levels: with a bright display (white background) the iris closes a bit more, decreasing the effect of the "deformed" lens; with a dark display (black background) the iris opens to receive more light and the deformation of the lens creates a much fuzzier focus at the eye.

> Jason Harrison – Post Doctoral Fellow, Imager Lab Manager – Sensory Perception and Interaction Research Group, University of British Columbia

> Jason Harrison – Post Doctoral Fellow, Imager Lab Manager – Sensory Perception and Interaction Research Group, University of British Columbia

Borrowed from this answer in stackoverflow: https://graphicdesign.stackexchange.com/questions/15142/whic...

Looks like this study is about people with mild vision problems. As I said before, I know some people with more limited vision (myopia with more than 10 diopters, optical nerve damage, etc), and all of them prefer light letters on dark background. Myself I can hardly read on paper or a paper-white screen. Whatever science may say, this is how I experience it! :)

These are my terminal colors. Goes back to the green on black of my Commodore PET in 1978. It works well for my aging vision.

Thank you! I like the "calm forest" analogy!

> The test hardware was a Dell E6410 with a Core i5 CPU running at 750MHz

Is there a reason to downclock your CPU when running benchmarks? Is is trying to reduce the effect of memory stalls?

The Minix developer Kees J Bot once postulated that "the sum of CPU power and user brain power is a constant", so I'm underclocking all my hardware in the hope of gaining some IQ points.

More seriously: I cannot afford to buy new hardware on a regular basis and heat is what kills hardware. That, and 750MHz is plenty for everything I do.

At one point many years ago I realized that I was happy with approximately 486 CPU level performance. I wished at the time for a 486 built on Pentium fabs. Of course simply scaling down the size of an old CPU probably doesn’t work, but still, the energy performance trade-off is a valid one.

But then we figured out how to make software slow (all good and appropriate abstractions, I’m sure) and I find myself wishing for faster CPUs again.

I’ve often wondered if it is just that I’ve gotten impatient or if performance has actually not perceptionally improved much over the years.

This is why I like to visit the Living Computer Museum here in Seattle once in a while. It turns out that even very old systems feel snappy at some tasks even by today’s standards.

UI and interactive features are generally fast, but there’s no getting around the slowness of say rendering a fractal.

Kudos to you for this work and a fun and interesting website. It reminds me of the recent “This Old Lisp” posting.

Worth noting that an 80s computer would already draw the key you pressed on the screen, while the same keypress would still be bouncing around the transistors of a modern USB keyboard.



"Old computers did it better!" https://www.youtube.com/watch?v=0wDtxYeJdzg

I think it depends on what you are trying to do: any kind of modeling and you wouldn't be happy with 66 mhz.

But for most people, they don't really have a computation problem, they have a communication/information problem, and I think computers could be made more effective in solving it.

Does the choice of shallow rather than deep binding have any semantic impact in practice?

The difference cannot be exposed by a Scheme program, so no.

Why this implementation over Chez now that Chez is free? Or Racket?

There are dozens of Scheme implementations. Are they all obsolete now that Racket exists and Chez Scheme is free? I don't think so! There are lots of sweet spots and lots of implementation to fill the niches. This is something I really like about Scheme!

Scheme 9 is small, super-portable (it only requires a C89 compiler, no libraries, no language extensions), and quite easy to grok if you are into studying the internals.

Edit: typos. This is what you get for replying right after your afternoon nap. :)

Tl;dr: a modern, non-naive implementation, much faster than previous.

I wonder if the book will be updated?

Clicked too fast.

From the website: "The book might eventually get an update, if circumstances permit."

Where's the source code?

It's a bit hidden, but linked: https://t3x.org/s9fes/index.html

Ah, I somehow thought that that was the previous version (not "reimagined"), but it's actually up-to-date.

cool. Will this help klong as well?

Klong already uses bytecode internally, although it is not as efficient as the one in S9 Reimagined. I doubt that adopting the new abstract machine would help that much, because Klong spends most of it time evaluating built-in operators anyway.

I hope its as good as the movie (considered by some to be the worst movie ever made). Read the trivia/goofs on its IMDB page, its hilarious.



Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact