When the sun is directly overhead it's meant to be 12:00 - IN THEORY! However as Timezones are pretty wide, most of the time you'll be at least 15 minutes out. Sometimes you'll be out by as much as 3 hours - and you've probably never even noticed! Telescopes already have to compensate for this (as well as for summer time). Leap seconds make a shambles of book keeping too. What is "2022-07-17T12:00:00" + (60 x 60 x 24 x 365 x 5) seconds? No one knows! And the answer to that question will change depending on when you calculate it and which updates you installed! So I say ditch the leap second and let it drift. In a few hundred years we could update our timezones if we _really_ want to (timezone changing is actually pretty common, so code should already be handling this edge-case).

Let's plug one of my "joke" domains:

http://amibehindacaptiveportal.com

If you get any response other than "No" then you're behind a captive portal.

People can simply set the volume on the player louder if there's more background sounds. The trick to the loudness wars is to sound louder (than other songs) at the _same_ volume. If a bar or radio is in the background at a fixed volume, the loudest songs stand out.

Not if the volume is uneven. The either the silent parts will be inaudible or the loud parts will be extremely loud.

This site is a bit scarce on details. There's more details and some photos here: https://de.wikipedia.org/wiki/Cave-Link http://www.hoehlenverein-blaubeuren.de/index.php/cavelink-te... (Use Google Translate)

I wonder how the licensing issue referenced in the Wikipedia article is handled in the UK. I imagine Ofcom will have had to approve and license these installations, or will they have made these exempt (I can't find any evidence of this)?

I can't see these devices specifically mentioned in https://www.ofcom.org.uk/__data/assets/pdf_file/0028/84970/i..., though the frequency range they operate in (20-140kHz) is allowed to be used by induction loop type applications, at a suitably low power.

... but the cave link uses relatively high power levels (20-30 Watts if I remember correctly)

Inductive stove tops are also In the KW range

as far as I know similar general licensing exists for such purposes. Lots of inductive applications need it.

Something that is mentioned here and not on the main page: the devices form a mesh network and can relay messages between each other.

Seems strange, feels like a pretty big selling point.

I'm really surprised ELF are not commonly used for this. Even a simple single-frequency CW transmitter for emergencies.

I wonder if audio frequencies below 20KHz can be used to make a cheaper system.

At the extremity of that band, fewer nations have constructed transmitters than have nuclear weaponry.

You can combine acoustic with radio in TARF (yes, really).

I meant acoustic frequencies in ELF, not using sound... but TARF looks amazing.

see e.g. https://www.mit.edu/~fadel/papers/TARF-paper.pdf for those who didn't hear about TARF (translational acoustic-RF communication)

I think "1 million" means 1e6 everywhere. But the confusion might be because "1 billion" means 1e9 in USA but 1e12 in UK/Ireland. "1 trillion" similarly is 1e12 in USA but 1e18 in UK/Ireland. This is called the short/long scales. It's a headache.

I knew billion was different, but the commenter mentioned million specifically as having a different meaning.

You _can_ deny cookies. It's just that this breaks almost every website. This was true before Chrome existed.

If Chrome made denied cookies by default and required an explicit opt-in caused by a user action (basically deter un-prompted cookie prompts like we managed to deter popups) then that would change very quickly. I wonder why they don't?

I have my browser setup to delete cookies each time I close it. And I run ad, tracker, and script blockers on top of that.

So do you just deal with logging in every time on every website when you open your browser again?

A password manager really helps here, yes.

The downside is that I always see the cookie banners, which I mostly try to ignore but some of them block most of the page.

Tangentially related, but does anyone have a trick to Google search for holograms without getting pages of lenticular/pepper's ghost items instead?

I'd like to buy a couple of pre-made holograms just to get a feel for them, but can't find anywhere selling them?

Add some detailed keywords. In this case "reflection hologram" is a good keyword and I see a bunch of reasonably priced items on Etsy.

> I would jump on an optimised version of Illustrator with the feature set from 10 (or more) years ago.

Genuine question, can't you just install an old version of Illustrator?

Not anymore! Adobe has been taking the activation servers offline for its older products for some time now. Newer versions as I understand it are the service model "Creative Cloud" meanwhile

You can, just not an official version :/

So... it was basically a subscription all along.

Other issue is version compatability if you're working on a team

Growing plants/trees and then storing them underground _is_ carbon negative.

Nah, some asshole will come by in a million years or so and frack them to get fuel again ;)

I'd worry a bit about the chance of collisions.

They correctly state you can generate 50 million of them per millisecond with only a one in a billion chance of a collision (per millisecond).

Unfortunately this means the chance of a collision in a year is 99.999999999998% which isn't great!

(I think they've given too many bits to the timestamp.)

Also, the uuid format it creates isn't a valid uuid which is a pity.

If you want to have "only" a one in a billion chance of a collision a year, you can only generate about 300 a millisecond.

Look for a different solution if you need 1 quintillion IDs per year. Noted.

I guess you assume to build one timeflake per ms for one year. That's about 3e13 timeflakes. Ordinary UUIDs have a collisions at the order of 1e18 hashes (cf. https://en.wikipedia.org/wiki/Universally_unique_identifier#...). So we are "only" 5 orders of magnitude away. For me the collision quality feels the same.

If you assume one per ms then timeflake and UUIDv1 both have a 0% chance of collision as they encode the timestamp.

You might go above this rate some of the time though and we might as well strive for perfection.

One of the nice things about UUIDs is that you can just merge data together willy-nilly without worrying about collisions - for example, if you gets bought by Google. You can't really do this with timeflake.

>> If you assume one per ms then timeflake and UUIDv1 both have a 0% chance of collision as they encode the timestamp.

They can be generated anywhere at any time. It two machines are generating even 1 per hour, they will have collisions if their clocks are synchronized. The random bits are important.

You'd run into a bigger problem than collisions a lot sooner (IMHO): time ordering. "ms" precision is very misleading, because keeping clocks really in sync while processing this much data is a real challenge.

Sure, maybe it doesn't matter that one server/emitter and thus a bunch of IDs are out of order, but if this is really just for helping indexes, then a simple 0.1 second precision would make more sense. (Or even simply truncating the timestamp to however many bits to get ~2-30 sec precision, and using a data store with LevelDB-like staged compaction. Which is virtually a requirement at this scale.)

> Unfortunately this means the chance of a collision in a year is 99.999999999998% which isn't great!

How was this computed? My intuition begs to differ.

Probability of no collision in a single millisecond: p_1 = 999999999/1000000000

Milliseconds in a year: t = 1000 * 60 * 60 * 24 * 365

Probability of no collision happening for every millisecond of a year: p_2 = p_1^t

Probability of a collision happening for at least one millisecond of a year: p_3 = 1 - p_2

https://www.wolframalpha.com/input/?i=1+-+%28999999999%2F100...

There are 31,536,000,000 milliseconds in a year. If you do (1-1E-9)^3.1536E10, you get a 2E-14 chance of no collision in a year.

I think the presumption is you will not be generating 50 million of them each millisecond.

Which is a pretty safe assumption, considering just generating those 128bits 50 million times a millisecond results in over 20 million TB of data in a year. I think it's safe to say this is not a common scenario.

There are close to 32 billion millisecond in a year. If the chance of a collision happening within a millisecond is 1 in a billion then my intuition is that the probability of a collision happening at least once a year is quite high, indeed.

It's like a dice: You have 1 in 6 chance of getting a 6 every time your roll the dice. Now, if you roll the dice 100 times the chance that you'll get a 6 at least once is then pretty high.