Wikipedia talk:WikiProject Color/Principles

Explosion in the paint factory

Color itself is not merely one abstract concept; the word refers to several distinct ideas, each of which is a world in itself. To experts in each distinct field in which color plays a part our community adds honest amateurs and the usual crowd of listcrufters who think they know far more than they do.

O how they cling and wrangle, some who claim
For preacher and monk the honored name!
For, quarreling, each to his view they cling.
Such folk see only one side of a thing.

As someone who puts bread on the table working professionally with color, let me set out a few facts and some conclusions of my own.

Facts

There are an infinite number of colors. This should be obvious but not all people have the experience to visualize the extent of this infinity. Let me illustrate.

Let's say we create a page describing a single hue of paint that can actually be mixed from a tube of Winsor & Newton Artists' Oil Colour Cobalt Blue (180 AA S5) and a tube of Cadmium Yellow (108 A S4) -- just sitting there on the palette in a blob. Clearly, one such page could be made for every mixture, from the entire contents of the blue tube plus an infinitesimal dab of yellow, through half-and-half, to all yellow and a dab of blue. Hair-splitters will claim otherwise but for all practical purposes we would require an infinite number of pages.

There are several dozen paints in this line, from this manufacturer alone. We might mix any or all of them in any combination. This manufacturer sells other kinds of paints, too -- water colors, acrylics. Then there are other manufacturers, not to mention house paints.

What happens when we apply any one of these infinite hues to a medium? Depending on the method of application, we actually get a range of tints; the white canvas blends with the paint in various proportions. This range is -- yep, infinite. For each one of an infinite number of hues, on a single medium, an infinite number of tints. There are, of course, many possible media besides canvas bleached and unbleached.

But wait! We have not so much as mentioned light. The same artist standing the same canvas in the same window will see different colors throughout the day as the sun rises and sets. There is no moment when the impinging light is really white; that concept is empty of meaning. Then there is firelight, candlelight, tungsten, halogen, and (ick) fluorescent. Each illumination produces a different color as seen. Infinite hues, infinite tints, infinite lights.

We've still only discussed oil on canvas and related paint-on-media colors. What about neon signage? The colors you see on your computer screen? On mine? The packaging at your supermarket? The overproduced covers of artbooks? Each application (and thousands of others) produces its own sensation of color.

Mathematicians may disagree but I envision an infinite number of universes, each with an infinite number of planets, each with an infinite number of 'pedias, each with an infinite number of pages.

WP not paper but this is absurd.

Common names for colors are not strictly defined. There is no single dominant or authoritative Red; nor is there one Fuchsia. Color names are pointers in the general direction of vast realms of color. This includes such specific-sounding names as Burnt Umber. Many authorities define colors by name, number, and swatch; none agree.

Here's a little social experiment. Go to your local art supply store and hang out near the oil paints. You're more likely to strike up a conversation with a "real" artist if you have a used sketchbook in one hand and dress in rumpled but unstained clothing.

(The general public wanders without picking anything up or buys and leaves; art school students are wearing Daddy-funded trendables; desperate housewife/retiree amateurs wear their paint-spattered work clothes to the store to show off; "real" artists dress up to go out. Any shirt without paint on it is "dress up".)

Don't steer the conversation but wait patiently. You are guaranteed to hear a complaint that Store X does not stock Brand Y paint -- and the speaker's grievance is that, for all the brands in stock, Color Z is not "right". This is not subjective illusion; tubes of, say, Rose Madder from different manufacturers may well not match.

Put that same oil painter in a room with a web offset process ink printer and watch the fur fly. Both will agree on one thing only: They do not even see colors the same way, let alone use the same names for them. These are color professionals talking about well-defined colors. The general public has other ideas.

It's patently absurd to point to any one thing as a definition of a color.

Color models are incompatible. There is no formula for conversion between most pairs of color models. Most formulas that do exist are strictly ad hoc, with very little theoretical basis. The human eye is the most constant thing across color models: If you think two things are the same color, they are (maybe).

A physicist will tell you that there is only one true color model. He will say that any light falling upon his apparatus consists of photons, regardless of source or method of coloration. The incoming photons may individually vary in energy; the greater the energy, the bluer the light. The only other variable is the rate of infall, the number of photons arriving per unit time. More photons mean brighter light. This does not mean that any light can be specified along only two axes, though. Any light may contain a mixture of photons of different energies; there is an infinite number of possible combinations or spectra. To top it off, the visible spectrum is a tiny fraction of the electromagnetic spectrum and the physicist sees little reason to treat it as a special case.

The classical color model is epitomized by the Runge-Itten color sphere; an equatorial cross-section resembles the common color wheel. The north pole is white, the south pole black, the exact center of the sphere gray. This model is highly useful to the artist but has little in common with any other technical conception of color.

The paint salesman has a completely different definition of color. To him, a particular color of paint is whatever you get when you open a can or tube. Every manufacturer has its very own swatch book and every product has its own proprietary code number. Guess what? You are not guaranteed that the paint you buy will match the swatch in the book; such books often carry lengthy, explicit disclaimers. You are guaranteed only that if you order Paint Code 123, you will get a container labeled "123".

Commercial printers (operators of printing presses) are another class of professionals who have strict definitions of color. They will start by explaining the CMYK model and almost immediately mention that there is not one such model but a dozen, depending on the set of process inks chosen, the medium, and the press. Enormous quantities of money have been spent over the years to enable printers using one set of these variables to imitate the colors seen by printers using another -- and you still will not get a printer to commit to a guarantee of color fidelity. Moreover, for one single color, each combination of process variables will use cyan, magenta, yellow, and black process inks in different proportions in order to simulate the same color. Anybody who thinks this is an exact science needs to hang out at a print shop and wait until a fussy pressman bangs the press on the exact spot needed to make the proofs come out the far end the way he thinks is "right".

These same printers, after lecturing you for an hour or three on CMYK process, will then say the magic word gamut and walk right into a discussion of spot inks. Spot inks are sold the same way as paints, by color code number, and subject to the same waffle. Walk into a print shop wearing a turquoise dress. Your printer will never guarantee to print business cards matching your dress; he'll only guarantee to print them in Pantone 321 CVU.

Web designers have a completely different view of color. Either directly in code or through image file formats such as PNG, they specify colors numerically as percentages of red, green, and blue on your computer screen -- assuming you have an RGB monitor. Since they have no idea of your monitor's black level, white point, or gun response curves, they really have no idea what you see.

The invasion of home and workplace by computers loaded with Photoshop have made many a rank amateur think he knows something about color. As is usual with calculating machines, false precision leads many astray. Touch a button and the machine displays a numerical value for what is surely some color on screen. Since the machine said so, that must be the right number. Never mind that your monitor's color calibration might be far distant from mine.

Toss in this mix the great unwashed hordes who go shopping for a "red" car ready to buy anything that isn't white, black, gray, green, yellow, or blue.

You cannot convert reliably between these different color models. They are not just different ways of specifying the same colors; they are different ways of perceiving the world around us.

Conclusions

Like most WP articles on potentially technical subjects, color-related articles suffer from an overdose of technical information and a dearth of social context. Some articles are moving in the right direction; I like Green. Some are only pointless grasping after an impossible technical precision.

Almost all of the hundreds of individual color articles are pointless. List of colors gives all the technical information present in the articles, which lack social context. Meanwhile, the list itself is sufficient forum for the technical foolishness that pretends to define, say, Ultramarine by swatch and a variety of numerical models. Simply associating a name with a vaguely representative swatch and the hex code used to produce it is all that's really warranted.

Wikipedia ought not attempt to duplicate or subtitute for color tools and methods. In order to work with color in any useful or accurate way, you need to have one or more tools, such as Photoshop and a Pantone swatch book, or a box of oil paints and a well-lit model. These tools incorporate precise and comprehensive (though incompatible) methods of specifying color. We have not the capability to duplicate or subtitute for these tools; the attempt deceives the public and arouses the contempt of experts.

If it lacks social context, an individual color article should be redirected to an article with social context. (For example, Coral (color) → Orange (colour).) WP is not a technical manual; articles are of interest to the general public only insofar as they relate to the general human experience. Resist the temptation to legitimize individual articles by dropping in a single line of social context. That line can be used to make the main article stronger. Merge and redirect such substubs.

What do I mean by "social context"? Here are some quotes from actual articles:

Green: "In ancient China, green was the symbol of East and Wood..." Social.

Peach "is color that combines pink and orange colors." Technical.

Fuchsia "The color gained currency in computer usage as an alias for #FF00FF" Technical. "Fuchsia was a popular color for interior decoration along with orange in the 1970s in the United States." Social.

Banish ugly, bulky, misleading {{Infobox Color}}. Readers who want bullshit numerical values for colors can get them from List of colors, which should be linked from every color article. Templates such as {{Shades of orange}} are acceptable. I do think they need work, though. For one thing, most of the shades shown shouldn't even have articles. They should all be derivatives of one standard template and that should be made more compact and attractive.

The key image for each color article should illustrate the color as seen in the world around us, in multiple tints, shades, and hues. This illustration of Green is exemplary. It should be top right. An inability to illustrate an article with real things in that range is a red flag.

My conclusions are not novel; other voices have spoken to the same issues, drawing many similar conclusions. I hope to support these. John Reid 02:15, 31 August 2006 (UTC)[reply]