# On the use of \color in equations

I recently see more and more people utilize the \color function to make equations easier to understand, e.g.:

$$\frac{3\color{blue}{x}}{4}\times\frac{42}{\color{blue}{x}} = \frac{3\times 42}{4}=\frac{63}{2}$$

However, I would like to remind you that there are about $5\%$ of the population which suffer from one or another form of color-blindness. To me the colors are often confusing and sometimes even cause a notable eye-strain when trying to understand what is written.

It is impossible to meet the needs of all the different people with their different deficiencies, so instead I will make the following appeal:

Please, if you think about using colors - think about using \mathbf, \mathit or \underline first. Try not to overdo the colors, too.

• I definitely agree with the sentiment, but doesn't much of this come down to the difference between using color for content and using color for emphasis? I would think most people would be fine with your example even in the absence of color, for instance, since the color here serves to emphasize but not to define anything; where color is an essential part of the answer then it ought to be treated much more carefully... – Steven Stadnicki May 10 '12 at 18:17
• Did it occur to you that someone might downvote who has his own troubles parsing content and feels greatly helped by color? I certainly did not downvote this and I will take your problem into account in the future, but as you say yourself, it is not possible to meet the needs of everyone, so ... you expect that you should be accommodated. Assigning anti-handicapped people motives to downvoters is unwarranted. – Phira May 10 '12 at 19:07
• I have some trouble understanding your argument, but it looks like you're saying that the removing of the color annotations in your example, without doing anything else, would be an improvement to the post. That sounds backwards -- in my view color that is not essential for the message is an appropriate use of color. It's a bonus that doesn't detract from the experience of people who can't see it. In contrast, it would be a problem if the use of color was essential to understand what the author meant. – hmakholm left over Monica May 10 '12 at 19:32
• Presumably there are already ways for color-challenged folks to force monochrome rendering. If your argument held weight, the entire web would be monochrome. I use colors when they greatly aid discerning structure. I encourage everyone to do so in such contexts. – Bill Dubuque May 10 '12 at 21:18
• @Asaf This problem appears to be easily solvable. Simply configure your MathJax environment so that \color does nothing. I'm sure there are MathJax wizards who would be happy to assist you. – Bill Dubuque May 10 '12 at 21:52
• @Asaf Of course one can do smarter things too, such as mapping the $\LaTeX$ colorspace into one that works for you. There are many solutions that do not involve throwing out the baby with the bath water. – Bill Dubuque May 10 '12 at 22:00
• @Asaf Even if all of our users, current and future, were aware of these issues, and did their best to help, it probably would help little, since most folks have little knowledge of the complex design decisions needed to address such issues. But since we cannot even get folks to read the FAQ, there's little chance of teaching them about such. If you really desire to solve it once and for all then your best bet is to do as I suggest, and force the changes you desire via software mods on your end. – Bill Dubuque May 10 '12 at 22:16
• @Bill: Unfortunately this is impossible, my color blindness is very atypical and none of the standard modifications are close to my deficiency. The result is that it looks bad no matter what I try, so I stick with the original. On the other hand, since most of the users that I saw are also meta readers I would think that a meta thread which merely tries to raise some awareness was in order. In fact, it is not the software design which I am worried about. (...) – Asaf Karagila May 10 '12 at 22:20
• @Asaf So how would you propose to modify the color emphasis in the multiplicative telescopic cancellation in my post here? – Bill Dubuque May 10 '12 at 22:24
• @Asaf But tweaking the colorspace need be done only once, and it can be done optimally for your needs. I do analogous things all the time, e.g. being an Emacs developer, I heavily tweaked the Gnus newsreader to suit my preferences (not only for colors, but for scoring, and many other things). If you're not already a software developer, it's not hard to get up to speed for simple mods, esp. with expert coding knowledge at your fingertips via the stackexchange sites, one can go very far quite quickly. – Bill Dubuque May 10 '12 at 22:54
• @Asaf Interesting. Should you decide to attempt software mods, then I'll be happy to offer my help. I'm sure others would to. I suspect we have a fair amount of software expertise in our community. – Bill Dubuque May 10 '12 at 23:17
• This needs more development before it's a solution: chrome.google.com/webstore/detail/…. I find the color to be quite helpful, and I often wonder why examples are limited to black and white in a situation where the cost of color ink is not really a factor. Admittedly, the extension needs a lot of work, and in particular I think it is currently restricted to images and not text, but there have been other extensions (like addons.mozilla.org/en-US/firefox/addon/colorblindext/…) which handled both, and got rave reviews from colorblind users. – process91 May 11 '12 at 1:25
• @Asaf: I'm sorry to hear about your CB. As for me, I'm fond of using colors in my beamer presentations but never used them in answers on MSE (perhaps, never wrote an answer where it would be of help). Indeed, the first impression of your post was that you encourage others not to use colouring (which I would dislike) but in the case you only ask people not to use colours whenever alternatives might work as well, please take my upvote. I also agree that those who use colouring on MSE are aware of existence of Meta and hence it's likely they'll see your post. – Ilya May 12 '12 at 11:00
• I use colors when I think it helps, and I also find special fonts features troublesome: underlining is unreadable with fractions and italics or bold might be unnoticed or taken as a mistake/typo. However, I understand this issue and indeed I would love to have commands like \distinguish{k}{text} which reader could customize in his profile: some would pick colors, others would pick font-features. – dtldarek May 14 '12 at 15:06
• I am quite shocked that none of you even considered posting alternative answers in Braille, this is obvious discrimination against blind mathematicians. – David Wheeler Jun 8 '12 at 9:48

I have made an extension that eliminates all color styling from MathJax output. It should work in all the major browsers, though some need plug-ins to handle it; see this post for links to those. For IE, you will need to download the file and change the name so that it ends in .ieuser.js rather then .user.js before the plugin will recognize it as a user script (silly but true).

It would be possible to make a similar extension that exchanges one color for another, but I don't really know what the right colors are that would be good for the visually challenged. A more sophisticated version would exchange ranges of colors based on their RGB or HLS or other color-space values.

Personally, I find that people use colors that are too saturated, and I think it helps to use less saturated versions. For example, I find $\color{#C00}{\rm\#C00}$ easier to read than $\color{red}{\rm RED}$, at least on a white background.

Anyway, hope this is useful. If anyone wants to suggest what would be appropriate for a color-exchanging extension to do, I might be able to come up with something for that.

• That is awesome. I'll give it a spin later today. – Asaf Karagila Jun 13 '12 at 16:18
• ${\color{Blue} T}{\color{Red} h}{\color{Green} a}{\color{DarkOrange} n}{\color{Yellow} k}{\color{Teal} s}!$ – The Chaz 2.0 Jun 13 '12 at 17:45
• @Davide How can one obtain such colors in MathJax? Doesn't definition of colors require some color libraries that aren't defaultly loaded? – Bill Dubuque Jun 13 '12 at 18:47
• @BillDubuque, no, MathJax's default \color macro allows you to use any valid HTML color value. So \color{#C00}{x+1} produces $\color{#C00}{x+1}$. (The default \color macro doesn't work the same as the $\rm\LaTeX$ one; if you want the $\rm\LaTeX$-compatible version, then you do need to load an extension). – Davide Cervone Jun 13 '12 at 19:17
• @Davide Thanks, that good to know. I had presumed it worked only with the named colors, not having seen any other usage. Perhaps I should really read the code. Where can it be found? – Bill Dubuque Jun 13 '12 at 19:24
• @BillDubuque, the MathJax code is on GitHub at github.com/mathjax/MathJax but you should look in the "unpacked' directory for readable code. – Davide Cervone Jun 13 '12 at 20:06
• The link just outputs the JS code, shouldn't the extension of the file be .user.js rather than just .js for GM to identify this as a script? – Asaf Karagila Jun 14 '12 at 6:29
• @AsafKaragila, thanks, you are right. I have fixed the link and the name on the server. How embarassing! – Davide Cervone Jun 14 '12 at 11:37

[Update:] This answer refers to the initial version of Bill's arguments in the comments. By the time I had posted this, Bill had already made a less extreme point about tweaking colour rendering rather than turning it off. I don't have enough experience with that to weigh in on the difference of opinion between Bill and Asaf about whether this would be an efficient solution of the problem.

I'm disappointed that four people downvoted this question, most of them apparently without making any constructive suggestions. All the more so because Asaf was merely alerting us to a problem and making a reasonable appeal to consider alternatives, not asking anyone not to use colours. If, as Phira suggests, people downvoted because they have trouble reading stuff emphasized by bold/italic/underline instead, then they should say so, so that we can find the best solution for everyone. So far, no-one has said so, and as long as no-one does, Asaf's suggestion seems like a good solution to me. I doubt that forcing monochrome rendering, as Bill suggested, is a solution, since it would turn off all colours, including in graphs etc., where content might not make sense without them, e.g. in this answer I gave today. I also don't see any contradiction between Bill's announcement that he will use colours when they "greatly aid discerning structure" and Asaf's appeal to consider alternatives, if "greatly aid" is interpreted as "aid more greatly than the alternatives without colour".

So I would ask everyone to take Asaf's appeal seriously, and then if you've considered the alternatives and still feel that those who can discern colours would benefit significantly more from your post if you add them, then by all means add them.

• Please note that I suggested to remap the colorspace. This includes various possibilities, both monochrome and multichrome. It would be interesting to know the reasons for the downvotes. Perhaps they simple boil down to, as Paul Simon sang "Mama, don't take my Kodachrome away". We need a new verse: MathJax, don't crash my bloated Chrome today! – Bill Dubuque May 10 '12 at 23:11
• @Bill: :-) Yes, I saw your more detailed comments only after I posted the answer; the first one was only about forcing monochrome rendering. – joriki May 10 '12 at 23:15
• I would like to thank you, I feel that without your answer more people would probably have misread and misunderstood me. – Asaf Karagila May 11 '12 at 15:48
• @joriki Could you please be so kind to update your answer to reflect more accurately what I wrote. Note that my comment about remapping the colorspace was in fact posted a half hour before your answer was posted. – Bill Dubuque May 11 '12 at 17:07
• I misread and misunderstood you, Asaf, and voted accordingly. The inflammatory comments about boycotts and limps/wheelchairs didn't do much to win me over, but joriki helped me to see it in a new light. Since the edit, I have changed my vote. – The Chaz 2.0 May 11 '12 at 21:13
• Again, it would be great if people could explain their downvotes. If they're because of my representation of @Bill's argument, please note that it correctly represented his argument up to a certain point. That I didn't reload the comments before posting and that Bill later made a better argument doesn't mean that I was misrepresenting him; the argument as I represented it was in fact all of what was there for a certain time. I'm updating the answer now. – joriki May 14 '12 at 9:46
• @Bill: I hope I've now represented the final form of your arguments more faithfully. You're right that I could have reloaded the comments before posting, but on the other hand, if you want to avoid your arguments being represented in a certain form, you shouldn't let them stand in that form before making more subtle points. – joriki May 14 '12 at 9:53
• @joriki I'm puzzled by your remark. It's only natural to propose simpler solutions first. As soon as Asaf made it clear that his problem was more complex, I proposed the remapping approach. Only 8 minutes passed between the two proposals. An additional 28 minutes passed before your answer was posted. Were I following an ongoing discussion, then before criticizing the viewpoint of another, I would be sure to be uptodate on the discussion. And when informed of a misunderstanding, I would fix it asap, not two days later. Else one risks coming across as purposely misrepresenting another. – Bill Dubuque May 14 '12 at 15:05
• (Wasn't me this time!) – The Chaz 2.0 May 14 '12 at 15:12
• @Bill: a) I'm sorry it took me two days to update the post. b) I'd already acknowledged that it would have been better had I updated the comments before posting. That's a mistake on quite a different scale than purposely misrepresenting someone. c) The "risk" of coming across as purposely misrepresenting someone wasn't as impersonal as you describe it; it came about because your comments created the impression that I was misrepresenting you. I still feel that I accurately described a certain state of the debate, then $47$ minutes later acknowledged in a comment that the debate had moved on. – joriki May 14 '12 at 19:49
• @joriki Could you please remove any mention of my name from your answer. Then we can remove these comments. I don't understand why it is so difficult for you to comprehend what I write, but my patience is quite exhausted. – Bill Dubuque May 14 '12 at 20:05
• @Bill: I don't like that combative style. I'm not going to do anything of the sort. This is a public record. – joriki May 14 '12 at 20:53
• @Bill: I moved the update to the top to make sure that no-one reads the rest without reading it. As far as I can tell that should take care of any issues of "misrepresentation". – joriki May 14 '12 at 20:56
• @joriki Again you baffle me. "Combative style"??? I hinted and asked (quite politely) for what I thought would be a simple gracious edit. Were I in your position, I would have quickly apologized for any misunderstanding. And if I lacked the time to edit the answer, I would have told you to feel free to do so. That would have avoided any friction, and resolved the matter quite quickly. But, alas, that's not what occurred. Now the mattter is completely blown out of proportion. – Bill Dubuque May 14 '12 at 22:08