I visit HN for experiments like this. Love it.

The aliasing or whatever on the gridlines is causing that moire effect optical illusion for me and making it hard to enjoy looking for more than a few minutes.

Good point. I'll fix that. Do you mind telling me your screen resolution? I think this problem may be more pronounced on relatively low-res screens.

When I get home I'll give you a dump of info. It's a 4k screen with something like 150% UI scaling. Windows. Chrome. I'll test it in Ubuntu and on my other pc.

Something to be aware of is to use integer math where possible when dealing with canvas coordinates and drawing. I found I ran into similar issues when I allowed fractions to get into my system.

It seems it was hugged!

Nice game, was able to play it for a bit, got 8 right tops.

It seems like the density of bombs is much larger than a regular minesweeper? Or I've just been playing on easy mode in other times...

Not working for me either right now. There was a game like this before, something like minefield.net, but it eventually disappeared. Played it every now and then with a couple colleagues during lunch break.

Needs a different size grid for mobile so you don't fat finger mines, otherwise pretty cool.

Do not trust other people's flags though as they can be wrong - possibly intentionally.

Yeah, adjustable viewport size is on my TODO list.

> Do not trust other people's flags though as they can be wrong - possibly intentionally.

Yes, this is intentional. I want to see how people start collaborating across very limited channels. Or how they start playing against each other :)

Great work, and the adversarial aspect is an interesting twist.

One quick suggestion on the scoring.

Currently scoring favours quantity over quality so the best strategy is to just find the big empty areas rather than thoroughly clearing around bombs. If the score also factored in the adjacent bomb count for each revealed position it would make that a more viable strategy.

Nice but:

1. Do not debug in production (console.log ...)

2. Secure the session (do not display the session ID in high-scores)

3. This is not infinite (in64 * in64) square + weird border :)

4. Balance the score for big discovered area

5. For the new players start the position on an playable area

6. Limit the total of high-scores returned by the ws

7. Limit the ws data as much as possible

...

Thanks and sorry :)

1, 2, 6, 7: All good points, I've fixed those.

3: 64 bits are close enough to infinity :)

4: Good idea. Maybe something like a logarithmic curve on the number of discovered fields would be good.

5: Also a good point. I'll try to build that

502 bad gateway

Probably someone found a mine already.

In a way, yes. It's fixed now.

504 again. It sounds pretty cool, though!