Splat ruler: Newbies think full height is for walker, incl/excl ambiguity

Started by Proxima, February 17, 2024, 02:26:51 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Simon

All right, cool, thanks for clarifying.



I found this new problem with the white triangles in my proposed ruler. Look carefully at the highest row of the ground: I've stuck the ruler 1 pixel deep into the ground, which is a mistake. The ground is light, and that's common: Bricky tilesets often have such bricks with lighting from the top-left or top-right.

Curiously, the 1-pixel red line would have prevented this accident.

I think this new problem comes from the purely white end. If we put multiple colors in the ruler's end, this accident shouldn't happen. I'll experiment with a stick-shaped ruler, sized 3 x 63, with three columns of 1 x 63 in white-black-white.

-- Simon

WillLem

Quote from: Simon on February 22, 2024, 12:42:17 AM
I found this new problem with the white triangles in my proposed ruler. Look carefully at the highest row of the ground: I've stuck the ruler 1 pixel deep into the ground, which is a mistake. The ground is light, and that's common: Bricky tilesets often have such bricks with lighting from the top-left or top-right.

Curiously, the 1-pixel red line would have prevented this accident.

I'm not sure I understand the issue here. Is the ruler the incorrect length? Or, have you incorrectly placed the ruler because of the colour of the block?

If the latter, couldn't that potentially happen with any colour of ruler? If we make the ruler red or blue, for example, it's similarly difficult to use with a red or blue block.

Simon

Quote from: WillLem on February 22, 2024, 12:53:36 AM
incorrectly placed the ruler because of the colour of the block?

If the latter, couldn't that potentially happen with any colour of ruler? If we make the ruler red or blue, for example, it's similarly difficult to use with a red or blue block.

Correct, I've put the correctly-sized ruler at the incorrect spot.

Yes, this will happen with any single-colored tip. To prevent that, my idea is now to make the tip multi-colored. E.g., white-black-white (from left to right). If this 3-pixel width proves too thin, I'll try white-black-white-black-white. I'll see if that's too distracting.

2023 NL's ruler with the red line would have prevented that even in a red block. Reason: We knew that the red line has to be somewhere, and we'd have supposed it under the ruler's white. Likewise, the red line would have prevented it in a white block. Reason: We knew that the ruler's white section had to be above the red line.

-- Simon

WillLem

Quote from: Simon on February 22, 2024, 01:06:13 AM
Yes, this will happen with any single-colored tip. To prevent that, my idea is now to make the tip multi-colored
...
2023 NL's ruler with the red line would have prevented that even in a red block

I see, good idea. I suppose that this is why ruler snap is a thing in Lix.

If NL's different-colour-tipped ruler works, then perhaps the tip could still be a single colour, as long as that colour is different from the rest of the ruler?

If you stripe only the wider tip of the ruler, the main part can still be a single pixel width to reduce distraction. In fact, if we make the end a straight "T" rather than a triangle, this gives another visual clue when accidentally overlaying a similar-colour block, and perhaps even removes the need for stripes.

Dullstar

Ruler Snap like in Lix is probably the best solution.

Rather than making the end multi-colored, I think a better solution (although it would require changes to the code) would be to make the colors change. It wouldn't be entirely new functionality at least, since that visual effect is used in clear physics mode, just not with the splat ruler.

WillLem

Quote from: Dullstar on February 22, 2024, 03:05:13 AM
Rather than making the end multi-colored, I think a better solution ... would be to make the colors change

Nice idea :thumbsup:

I managed to get this working! This will definitely be used in SLX, happy to provide the code for NL if namida approves.

Here's a video demo.

WillLem

Looking again at the video, should the ruler be centred on the cursor?

Simon

First hunch from the video: This constant color rotation makes it harder to see where exactly the ruler is.

Reason 1 is distraction from the constant animation. It usually takes over a second anyway to verify that the ruler is exactly where you want it. And now, you aren't examining a still picture.

Reason 2 is monitor subpixels. Now, the ruler moves on the screen even after you take your hand off the mouse entirely. Yes, a hatch in clear physics mode will color-rotate, but there, it doesn't matter for looking at a big, fat hatch. Here, it may matter, we need the ruler to analyze pixel-perfectly. I can't tell how problematic the subpixels will be in reality (reason: video compression; another reason: you'll prefer different zoom levels than I will). It's up for experimentation.

-- Simon

WillLem

Quote from: Simon on February 24, 2024, 10:45:14 PM
First hunch from the video: This constant color rotation makes it harder to see where exactly the ruler is.

I know what you mean. I'll see if I can keep the colour rotation within a certain range of values, so that it only shows lighter shades. This might help to keep it more visible.

Quote from: Simon on February 24, 2024, 10:45:14 PM
Now, the ruler moves on the screen even after you take your hand off the mouse entirely

Does it? ???

If you're referring to the video, anytime the ruler moves it's because I moved it - it absolutely doesn't move on its own. Or, did you mean something else?

Simon

Quote from: WillLem on February 25, 2024, 02:05:34 AM
Does it? ??? If you're referring to the video
you mean something else?

Not necessarily in the video. It's a physical property of LCD monitors. When the ruler is green, it must be elsewhere (physically on the monitor) than when it was red.

http://www.lagom.nl/lcd-test/subpixel.php

-- Simon

WillLem

Quote from: Simon on February 25, 2024, 02:48:25 AM
It's a physical property of LCD monitors. When the ruler is green, it must be elsewhere (physically on the monitor) than when it was red.

Yikes! I suppose that does need to be taken into account, then.

My instinct is to make this optional. Thankfully, the ruler can look however we want, and the colour change effect will still overlay each pixel - so, we can have the best of both worlds. If people prefer a solid colour ruler, we can give it stripes if needed. If people prefer the colour-change effect, we provide this as well.

Both rulers solve the visibility problem, IMHO.

mobius

Just to be clear my post may have been misunderstood slightly. I wasnt saying "i couldnt be bothered to figure out how the splat ruler works." I meant I honestly never needed it for any of these more precise things people are complaining about so it was never really an issue. I just used the one line and ignored everything else.

More pointedly; why is there such a difference between dangerous fall distance bewtween walker/climbers etc? And if im playing a level that relys on such precision, im very likely to stop playing said level or pack. Thats bad level design imo
everything by me: https://www.lemmingsforums.net/index.php?topic=5982.msg96035#msg96035

"Not knowing how near the truth is, we seek it far away."
-Hakuin Ekaku

"I have seen a heap of trouble in my life, and most of it has never come to pass" - Mark Twain


WillLem

Quote from: mobius on February 26, 2024, 02:09:24 PM
why is there such a difference between dangerous fall distance bewtween walker/climbers etc? And if im playing a level that relys on such precision, im very likely to stop playing said level or pack. Thats bad level design imo

100% agreed :thumbsup:

Proxima

Quote from: mobius on February 26, 2024, 02:09:24 PMMore pointedly; why is there such a difference between dangerous fall distance bewtween walker/climbers etc?

Because climbers stop when they hit their head on a ceiling, but splat height is measured by how far the feet fall. In other words, the fall distance is the same but the smallest fatal distance from floor to ceiling is slightly more than the splat height.

QuoteAnd if im playing a level that relys on such precision, im very likely to stop playing said level or pack. Thats bad level design imo

Sometimes it's avoidable: for instance on "The runaround" (where we first noticed the issue) it would be possible to make the wall smaller so it's more immediately obvious that climbers will survive. That's because the solution doesn't (and can't) involve building under the fall from the ceiling.

But any time you have to build under a ceiling to make a fall safe for a falling climber, it's helpful to be able to measure how high your bridge needs to be.

WillLem

Quote from: Proxima on February 26, 2024, 06:22:21 PM
Because climbers stop when they hit their head on a ceiling, but splat height is measured by how far the feet fall. In other words, the fall distance is the same but the smallest fatal distance from floor to ceiling is slightly more than the splat height.

So, the Climber splat ruler is (fall distance + distance between the top of the fall and the ceiling)?

That makes sense, since we don't know where the Climber is actually falling from; I now understand better the need for the extra measurement. Also, I can see how this situation might come about unintentionally, rather than as a result of bad design.

I do still agree with mobius though; as useful as the ruler is for this purpose, a level designed in such a way that necessitates its use is poor design. Discontinuing the ruler at this point would be too drastic a solution to this, though. Better to provide the ruler (albeit separately from the main one, perhaps), but discourage levels which can't be solved without using it excessively.

I plan to add the Climber ruler as a separate entity, callable via shift modifier of the regular ruler hotkey and/or its own hotkey. This seems to be the best way to avoid any possible confusion.