Rolling the dice with CSS random() | Hacker News Reader

Rolling the dice with CSS random()

webkit.org

145 pointszdw31 comments

This seems like a good opportunity to bring up the old, more hacky, but also more performant and predictable CSS random effect using backgrounds of prime number sizes to achieve a "random" distribution. The "cicada principle"

https://www.sitepoint.com/the-cicada-principle-and-why-it-ma...

https://lea.verou.me/blog/2020/07/the-cicada-principle-revis...

In this case you would use multiple transparent tiles of different star patterns (images, or gradient/clip-path tricks), each one a different prime number in size. It should work with anything you can tile and overlay in CSS though.

I'm not sure if I'll ever get a chance to use that but it was very informative nonetheless.

(oh, I should note that the 2nd link uses nth selector to apply any rules pseudo-randomly, not just tiles) ... and, hm, I guess you could "seed" the pseudorandom nth selectors if your pages had unique attribute selectors, by adjusting the primes and offsets. Like with drupal you could do different ones based on digits of the nid in the body tag.

The starfield example is cool but it seems like that might be exactly where random() wouldn’t work as well as people hope: true randomness often looks pretty bad when you want to make graphics out of it, because true randomness has clumps and voids, and a lot of observers think it looks less random than pseudorandom sequences with more evenly-spaced points.

The term for this is “low-discrepancy sequences”, there have been a handful of HN posts on it over the years. I know I’m bikeshedding the API already before it even really exists, but for image presentation I think a lot of applications might actually find that more useful.

Related: Animated starfield in pure CSS

https://codepen.io/ArneSava/pen/BaWxOaR

Really laggy on an M1 MBP; probably `box-shadow`'s fault.

Have you tried different browsers?

Nice. Currently I have to set CSS custom properties with JS to achieve the same effect.

Wonderful to see how CSS gets a usable random function before JS does.

Maybe "usable" is your qualifier but what's wrong with Math.random()?

JS also has Crypto.getRandomValues()

To generate random number in a specific range you need to do something I always forget and need to google.

    Math.floor(Math.random() * (max - min + 1)) + min;

(Google AI summary says this is the thing)

The CSS function would be random(min, max)

Also the CSS function seems to take a number of steps, it is not immediately obvious to me how to do that with Math.random()

Why not add a Math.randint?

I imagine there's some deep ideological war over whether to add more programming functionality to css...

Currently under discussion in the standards committee: https://tc39.es/proposal-random-functions/

    Math.floor(Math.random() * Math.floor((max - min) / step)) * step + min

So now we can add a random data prop to a hidden dom element, then query that from JS. You know, to make your JS random function simpler. ;)

That was my second idea. I've done worse.

Nice but... no dice!

Perhaps you can set the seed to a fixed value on page loads? I kind of like the idea of the same "random" star field even if the user refreshes the page. Or rather, it would somewhat bother me if it changed for a refresh since a refresh is supposed to simply re-present the same web page.

Says who? Why would I refresh to see the same page? Usually I refresh because I want to see some different content.

I yern for the day we can have react-type pages without any javascript. Keep chugging webkit I believe in you.

These examples feel a bit contrived, are there any other cases where random CSS values would be useful? I don't often reach for randomness when doing business apps.

I can only imagine the groundbreaking and innovative MySpace themes that would have been possible with this new random technology.

Where is the spec? I can't find an entry on MDN.

Is there a way to get reproducibility? In the same browser or across browsers? Even if it's not the default mode.

Spec: https://www.w3.org/TR/css-values-5/#randomness

WPT test results: https://wpt.fyi/results/css/css-values?label=master&label=ex...

Only Safari supports it for now, it seems.

And only in a prerelease build; no browser has yet shipped this to users by default.

Having seen too many "this randomness function was never meant to be used for security, but people use it for security anyway" vulnerabilities in the past:

Can we PLEASEPLEASEPLEASE have this secure by default from the beginning?

Security in the Stylesheet? Come on, you need to set boundaries for expectable use.

"Look, I implemented diceware in pure CSS!" is unfortunately not that hard to imagine.

I would bet someone is already working on it as we speak.

I don't disagree on that point.

Introducing cryptography in the STANDARD for stylesheets adds complexity where it doesn't belong. Ultimately a browser vendor isn't responsible when a company sells insecure cryptography.

Adding crypto to CSS will bring us nearer to bitcoin mining in the CSS engine.

If you implement security protocols in a production app using CSS then you deserve to be hacked and then sued for negligence.

Counterargument: it would make for a very funny post-mortem.

Reply on news.ycombinator.com