Hacker News Reader

Comment by sandworm101 | Hacker News Reader

I would argue that for something to be trademarked it must somehow be recorded. A vague description is not enough. For a color such as Ferrari Red, we have specific values. For a sound such the windows noises, we have exact recording. If I were a judge, I would need more than "it smells like vanilla" before enforcing a trademark against someone with an allegedly similarly-smelling product. At least run the scent through a mass spectrometer or something.

Are we going to have vineyards suing each other because X wine smells too similar to Y?

View on HN

Aside from the separate conversation about the validity of a scent trademark --- trademarks have very different purposes for existing than a copyright or a patent. The point of a trademark is to protect consumers from being confused in the marketplace. You can violate trademark without copying a mark exactly. You can also copy a mark exactly and be completely justified in your usage. Extreme precision in defining the mark isn't totally necessary.

You say Ferrari Red is described in terms of ‘specific values’, but ultimately it’s still just a description. Who defines what the values mean?

I actually don’t know. Does each set of values (in whatever colour system you’re talking about) correspond to a list of quantities of specific physical dyes?

I went to dig into this a bit (since the obvious way to find out what is trademarked is to read the filing) but the euro trademark viewer fails to load for me in the US; a bit more digging suggests that not only is it not trademarked, but for a bunch of time they were driving Marlboro Red cars instead anyway, as part of a branding deal.

So, anyone have leads on any actual evidence here?

>> correspond to a list of quantities of specific physical dyes?

No. A trademark is not a recipe. What matters is the actual look. As for values, "Ferrari Red" is has defined RGB values according to colorhexa (RGB: (239, 26, 45), CMYK: (0, 89, 81, 6).0) and I am sure is outlined in any contract involving Ferrari sponsorship. No doubt Coke also has very specific values for its version of "red".

But if it’s specified in terms of CMYK, that is a list of quantities of specific dyes, isn’t it?

CMYK is a representation ratios of unspecific dyes, on an unknown medium.

Ah — true. I thought the dyes used for each colour would be fairly standard, though.

Dyes are not like light. Mix any three, regardless of ratio, and you basically always get brown. That's why paint stores carry so many colors.

They're close, if we're just talking about printer ink. But if you're coloring some other item, you might be using entirely different materials.

> Does each set of values (in whatever colour system you’re talking about) correspond to a list of quantities of specific physical dyes?

Yes. Pantone colors (an industry standard) are defined as a ratio of Pantone colorants. Those colorants then have specific formulations depending on the type of paint they go in. No matter what though, they're formulated such that the colors reproduce exactly every time (insofar as chemistry and physics allows). This is the same general process that most (all) standards use AFAIK.

But at the crux of it, we can measure color very precisely. A specific color correlates to a specific distribution of light wavelengths. Spectrometers are not particularly expensive these days.

Believe me, there has been a ton of effort put into this question. Color accuracy is a whole industry that most people have never heard of.

Pantone is a way to standardize on colors. But that level of precision isn't really necessary for a trademark case. The courts care about whether or not a consumer is confused by the color, and that level of specificity is way more coarse than the details Pantone is concerned about. If you rip off Coke and change the color to the next closest Pantone color, it will make zero difference in your court case.

It's difficult, that's why Pantone charges thousands of dollars. You know those chameleon car paints that change color based on sunlight or angle? Turns out everything is like that somewhat. You can have two parts match perfectly under a fluorescent light and then look like totally different colors under sunlight. I had to come up with a color scheme for a factory where all the colors matched, but all the different machines used powder coat, wet paint, polyurethane dip, all sorts of things. In the end, how we had to do it was buy a very expensive Pantone physical set of chips and light boxes, which costs thousands of dollars each and physically give these to all of the suppliers to match. We had to include this light box contraption with various bulbs and angles and a camera. You could define the spec by Pantone's written list, but you can't ever meet the spec without Pantone shipping you this science fair project.

I guess that’s why even Apple seem to struggle to make ‘Space Grey’ a consistent colour.

Can't smells be recorded by chemical composition?

20+ years ago, CSI (U.S. TV Show) had an episode where they used a device to identify fragrances; seemed plausible to me.

CSI is notorious for inventing lots of forensic tools that never have and never will exist. In fact, their portrayal of forensic science has warped public perception enough that it has led to the corruption of actual forensic science, because people find it extremely easy to believe that all of these magical investigative tools actually exist.

A show about actual investigative science would not be exciting to watch though, so they throw shit in like "freeze and enhance" because people love fantasy science. They just don't know enough to know the difference between what's real and what's not.

In the real world, we do have something called an "Electronic Nose", which is essentially what you are describing here. But, if you're actually trying to document the composition of a fragrance, you wouldn't work backwards by analyzing the smell, you would just document the formula you used to create the fragrance.

https://en.wikipedia.org/wiki/Electronic_nose

Ah yes, the "CSI Effect", I forgot about that. Thank you.

Probably, but couldn't there exist a situation where another set of chemicals produced a similar sensation?

There's a starup working on producng a wide variety of odors from a more limited number of ingredients.

Using AI, "naturally".

Let me guess. The start the machine, say the magic incantation "Zoom. Enhance." and the answer pops up in an animated sequence on a 72-inch high-def television screen after showing a brief animation of some kind of chemical process complete with flashing outlines and sound effects.

No.

IP works differently than code, and it is often subjective. See: copyright infringement.

> Are we going to have vineyards suing each other because X wine smells too similar to Y?

No, because the smell of (most) wines is not trademarkable, being a function of the product. Trademarking scents is difficult[1].

Attempts to quantify scent trademarks / music copyrights are doomed to fail. Partly because every person experiences these sensory inputs differently. Two smells could be very different chemically yet indistinguishable to real people.

1. https://www.orlaw.com/intellectual-property/yes-can-trademar...

Reply on news.ycombinator.com