Looking forward to checking your site later gershy (and placing this comment as a reminder :P ), but I'll admit I'm glad in a horrible sort of way that it means I was able to see this article.
It’s trivial to survive HN traffic with a static site (assuming you know in advance that the site needs to be prepared for HN traffic) - just stick it behind a CDN.
I found HN front-page to still be less than about 1 user action per second. Ie. Pageload.
That means if your site loads fast, you typically won't ever be serving more than a single user at once, so unless you're serving it on a burstable CPU and you run out of burst capacity, you should be fine.
Yeah, my site which is hosted on fly.io's smallest VM with 256MB RAM and only part of a CPU core was on the HN front page the other day and I was surprised to see that it barely even registered the resource usage: consuming just 50MB RAM with a 0.2 load average.
It is a fairly simple webpage, and it's rendered in Rust so it's probably more efficient than most. And perhaps it didn't get as much traffic as other things on the front page, but I was still expecting worse given how much some sites seem to struggle.
> It’s trivial to survive HN traffic with a static site
It’s fairly trivial to survive HN traffic, full stop. I’ve seen the RPS live from several “#1 on homepage” posts with hundreds of comments (let alone upvotes) and as another commenter stated, it’s at most single digit requests per second.
Edit to add:
I’m not judging harshly hobby projects for not surviving it, it was a hobby project after all, but that doesn’t change the fact it’s fairly trivial to handle if you set that as a goal.
> If you're serving a lot of traffic I'd consider switching to a faster server
I don't think anyone with a small personal site should spend too much time planning for that single, improbable, day when traffic might increase by several orders of magnitude.
With most hosts I've had, that type of increase would just trigger DDOS protection and/or take down the site with its itty bitty quota limits.
I assume it works just like Diplomacy. Meaning they both bounce. Unless one of them has support to move by another piece, and that piece itself is not under attack.
So how does castling, En Passant, and moving one's king into a checked (but wouldn't have been checked, but for the move of the opponent) position work?
I was very confused when I was in check mate, but the other player didn't capture my king for some reason.... so in the end I won. Didn't expect that. They probably just didn't see the check mate even if it was obvious to me!
A clever way to escape checkmate is to move your king towards or away from the piece that placed the king in check (as long as your new square doesn't also place you in check by a different piece you should be able to avoid the initial capture and still be safe since they won't be able to move this same piece right away).
Yes...I started playing and then realised none of the moves work...then the other player left the game but there was no way of me knowing they had left.
"Neither player moves first" The way that's phrased makes it sound like no one ever gets to take a turn, but they actually meant that players move at the same time.
Doesn't seem to be working at the moment. Console shows this error:
An Error lacking a "stack" property:
Form: TypeError
Desc: TypeError: NetworkError when attempting to fetch resource.
Keys: [ fileName, lineNumber, columnNumber, message ] html.room:49:9
Uncaught (in promise) TypeError: NetworkError when attempting to fetch resource.
Firefox kan geen verbinding maken met de server op wss://chess2.fun/?trn=sync&hutId=000000LFCzgqdaBb.
I made a whatsapp group with a friend then kicked them so it's only me, then I message myself all the links I need to check out after work. Works relatively well
Your opponent may have disconnected, giving you the win by default! I've gotta make that wording a bit clearer... right now it always says "checkmate".
As true as this is, it's also immaterial to the game itself. Since both white and black move simultaneously, the only difference in gameplay is literally which side your king and queen start on.
So yeah, it should be fixed, but it's also a purely cosmetic bug. And I imagine it's probably safer to fix it by swapping the assets and flipping the board rather than actually remapping white and black pieces.
Just played my first game; the dynamics here didn't take long to get used to and it's brilliantly fun to play. Bugs aside, at least.
A few people pointed out that the king and queen are in the wrong positions, but if you think it through, you'll also note that this is entirely cosmetic when both players move at once. If anything, it might be easier to fix this by swapping the assets and mirroring the board rather than swapping white and black, at least depending on how the game's written.
Seems like it got the hug of death from many users - few interesting quirks I saw while it was still running:
* You need to actually capture the king to win. Check/Checkmate does not apply here
* You can't move the same piece twice in a row (I think)
* You can actually skip a turn - This is somehow more strategic than it sounds, but I'm not sure how to use it - maybe just wait with your "venus flytrap" and then capture the pieces?
The last bullet point reminds me of the description from a game in China Miéville's Kraken.
> The “universal leaper” was usually thought the most powerful piece, he read, as it could go from where it was to any other square on the board. But it was not. Kraken was... It could move to any square including the one it was already on. Anywhere including nowhere.
Good. The entire check system is a pointless bolt-on esotericism. Not being able to place your king in check has never improved a single game of chess.
What about the rule that the king is not allowed to castle if one of the squares moved over by the king is under attack? If you allow the king to be placed in check then you surely would also allow the king to step over squares attacked by opponents peaces? But this would change the game in the sense that it will not be equivalent to chess any more.
Even tho I can't see the site, I love the idea of rule changes. We live in a golden age my friends; computers, automation and a massive online player base opens up so many opportunities to tweak the rules of existing games and see what falls out.
Maybe my rule change idea has already been implemented, but, other than building it myself, is there a way to play / automate chess / checkerboards of arbitrary sizes? Bigger / smaller boards, odd numbers of tiles, rectangles, etc. How about adding and removing pieces, an extra rank of pawns for example. Letting it run / learn for a million games or so and see what comes out.
Please let me know if yous guys have seen something similar.
Can't try it because it is being hn-hugged to death.
So, I will just comment that I've really enjoyed chess.com's fog of war variation. Basically, you can only see the squares where you are legally allowed to move, adding a whole different level of strategy to the game.
IIRC the Lichess dev isn't too fond of random people adding big new features. They tend to be a bit sloppily written (by people who don't have as much Scala experience as other languages) and then people ditch the project and don't maintain what they built.
And he said it's really difficult to remove features. The smallest most obscure thing will have at least one user. Removing a game mode or big feature makes loads of people angry. But he hasn't got time to maintain the millions of things people submit and then never touch again.
118 comments
[ 3.9 ms ] story [ 179 ms ] threadIn the meanwhile, others might be interested in visualizing a Reddit (and HN) hug of death:
https://www.tylermw.com/visualizing-a-reddit-hug-of-death-an...
That means if your site loads fast, you typically won't ever be serving more than a single user at once, so unless you're serving it on a burstable CPU and you run out of burst capacity, you should be fine.
It is a fairly simple webpage, and it's rendered in Rust so it's probably more efficient than most. And perhaps it didn't get as much traffic as other things on the front page, but I was still expecting worse given how much some sites seem to struggle.
https://danuker.go.ro/how-to-protect-your-personal-data.html
It’s fairly trivial to survive HN traffic, full stop. I’ve seen the RPS live from several “#1 on homepage” posts with hundreds of comments (let alone upvotes) and as another commenter stated, it’s at most single digit requests per second.
Edit to add:
I’m not judging harshly hobby projects for not surviving it, it was a hobby project after all, but that doesn’t change the fact it’s fairly trivial to handle if you set that as a goal.
I don't think anyone with a small personal site should spend too much time planning for that single, improbable, day when traffic might increase by several orders of magnitude.
With most hosts I've had, that type of increase would just trigger DDOS protection and/or take down the site with its itty bitty quota limits.
2. Get featured on HN
3. Suddenly Jeff can hire all the folks back he's laying off, from the bandwidth proceeds.
I think you mean "likely" rather than "possible". It's definitely possible. I had it happen.
https://en.wikipedia.org/wiki/Chess_boxing
It’s a curious game. The only way to win is not to play.
It’s a curious game. The only way is not to play
1. The accuracy or embellishment of the headline
2. The quality of the content behind the headline
This does not seem to strongly violate either condition.
So yeah, it should be fixed, but it's also a purely cosmetic bug. And I imagine it's probably safer to fix it by swapping the assets and flipping the board rather than actually remapping white and black pieces.
A few people pointed out that the king and queen are in the wrong positions, but if you think it through, you'll also note that this is entirely cosmetic when both players move at once. If anything, it might be easier to fix this by swapping the assets and mirroring the board rather than swapping white and black, at least depending on how the game's written.
Played a game, turns out I didn't castle--just moved my queen for no reason.
* You need to actually capture the king to win. Check/Checkmate does not apply here
* You can't move the same piece twice in a row (I think)
* You can actually skip a turn - This is somehow more strategic than it sounds, but I'm not sure how to use it - maybe just wait with your "venus flytrap" and then capture the pieces?
> The “universal leaper” was usually thought the most powerful piece, he read, as it could go from where it was to any other square on the board. But it was not. Kraken was... It could move to any square including the one it was already on. Anywhere including nowhere.
Any position like this would essentially be drawn if both sides could skip their move
Good. The entire check system is a pointless bolt-on esotericism. Not being able to place your king in check has never improved a single game of chess.
Maybe my rule change idea has already been implemented, but, other than building it myself, is there a way to play / automate chess / checkerboards of arbitrary sizes? Bigger / smaller boards, odd numbers of tiles, rectangles, etc. How about adding and removing pieces, an extra rank of pawns for example. Letting it run / learn for a million games or so and see what comes out.
Please let me know if yous guys have seen something similar.
So, I will just comment that I've really enjoyed chess.com's fog of war variation. Basically, you can only see the squares where you are legally allowed to move, adding a whole different level of strategy to the game.
I wish lichess could add it!
And he said it's really difficult to remove features. The smallest most obscure thing will have at least one user. Removing a game mode or big feature makes loads of people angry. But he hasn't got time to maintain the millions of things people submit and then never touch again.