I have created a survey about colors in Qubes, would you take a few minutes to do it? It’s completely anonymous.
Please don’t let the second page scare you away! The first one contains many questions which also give very valuable information to me/us.
I hope to learn a lot about you and what your thoughts are on colors in Qubes. Are there too many colors? Too few? What do you associate with the colors? what do you use them for?
I will run this for about a month. So there is plenty time to answer.
There is also qubes-issue for this survey for feedback and when I’ll post the findings there.
Interestingly, I found out that I should probably revisit my choice of colors while taking this survey.
Some inconsistencies have slipped in over time.
My need of color is just to differentiate VM of different clients, which rend more easy to use
@HackerNCoder What were the results?
Good catch; I didn’t realize that introduction would be relevant over here.
I might as well add that I use gray for trusted qubes like Vault (where keepass lives) as well as for the Keys qube i mentioned later in that thread–Owner ended up as purple, because he must talk directly to Ethernet. I’d have preferred white instead of gray, but there’s no white.
And black, of course, for low level system stuff in keeping with dom0. All of my templates, even for red-through-purple app vms, are black–the exception is the templates for networking qubes which are yellow (wifi) or purple (ethernet). The networking qubes themselves follow the same color convention. Cacher is black. Whonix is red. (Basically, the more I want my computer to be wearing a full body condom when it visits a site, the redder the color gets on the yellow-to-red continuum.)
Incidentally in my distribution, a lot of the qube icons for gray were identical to the black ones…but some of them weren’t–it depended on the actual size of the icon. I guess that was technically a bug, but it was very noticeable in qube manager. I was able to fix that with a lot of gimp and bmp work.
I am guessing the survey is long gone.
I’d like to see cyan added as a color. “WTF,” some people will say, “that’s just light blue.” (And I’d also like to see white, but the rest of this is going to be me ranting on the subject of cyan.)
No, not really–cyan is a distinct color. It’s just that English doesn’t recognize it as such. But Newton did; he deemed the rainbow to consist of red, orange, yellow, green, blue, indigo and violet. But there has been a language change since then, and people now think of “indigo” as a deep bluish-purple. Not in Newton’s day: What he called “blue” was very like the color of the sky, and “indigo” was the color that you see, for instance, in the US flag; not a bluish-purple at all but rather “dark blue” in today’s parlance. The dye named indigo produces that color. And if you actually look at a rainbow there’s a distinct “cyan” or “light blue” stripe next to the “indigo” or “dark blue” stripe. Which means, it is NOT just “light blue.” (Find me the light and dark orange stripes on a rainbow…why aren’t there any?)
Many languages (Russian and Italian for instance) make the distinction, with two totally separate color names for what we call “light blue” and “dark blue.” They wouldn’t think of describing the two colors as light or dark versions of each other, any more than we’d call green dark yellow.
English ends up with the following “basic” colors: white, gray, black, red, orange yellow green blue, purple, then brown (which is really a dark yellow or orange) and maybe tan or pink.
In fact, though, cyan is a separate basic color and should be recognized as such–furthermore it appears in the rainbow unlike brown and tan.
Pretty much. I don’t even know if I have the data anymore. I remember
sending it to someone else… nina?
Although! I did start on a blog post for it, which has just been sitting
in drafts for months. Here is the not-finished version, and what may be
all the data I still have: Qubes Colors — HackerNBlog
In my opinion, it is tragic that we have to do polls and write dozens of issues on GitHub for the creators to add a more of colors. It’s only colors, not the rebuild of the system and change everything. Around 20 lines of code and the topic has been discussed for two years. Users have asked for it so many times and the developers don’t care.
Well, it’s not that simple. There have also been several different arguments against simply adding more colors (which are detailed in the aforementioned GitHub issues), so saying that the developers “don’t care” seems disingenuous. It’s probably just that they have not yet been convinced (1) that it’s a sufficiently good idea, (2) that it’s a high enough priority relative to the hundreds of other open issues that are “just 20 lines of code,” (3) what exactly the best implementation would be (N additional hard-coded colors, custom hex-based color picker, add patterns instead of more colors, etc.), and/or (4) that there won’t be any unintended, unforeseen side effects on any other part of the entire complex system or the myriad ways users can interact with the system.
Stepping back for a moment, it’s not as if there’s a herd of contributors tripping over each other to volunteer for all the open “just 20 lines of code” issues. While we have wonderful community developers who make fantastic contributions, both core and community devs are rare and precious relative to the overall volume of enhancement requests from us users, so I think we should try to cut them a bit of slack and remember all they do for us.
Nina is no longer working on Qubes, but perhaps Marta might have the data now.
(Mentioning @marmarek since Marta isn’t on the forum AFAIK.)
To my mind, “more colors” is a “nice to have.” I’m certainly accumulating blue and yellow qubes right now and it’d be nice to be able to quickly distinguish between them (more so than the different titles already do, I mean)…but it’s not critical.
I would imagine the way to make the most people happy…ASSUMING someone in the relevant place has the time to do so…would be to have the color palette under the control of a config file that users could readily get to. That way users who think the current selection is good, can leave it be. And we haven’t even discussed the sorts of things the color blind would need, to even have additional colors be useful. So I can guarantee you that if “someone” makes color choices for everyone else, few will be happy about it. I can crusade for cyan; I ain’t going to get it and I know this.
But let’s say someone does all that, anyway: it wouldn’t solve the issue of generating all of the icons. (It’s error-prone already: as I’ve noted, some gray icons, in some sizes, in my distro were black…I was able to fix that, imperfectly, with a graphics editor.) I still have one icon, that at one size, shows me a padlock instead of template pair of qubes, and I can’t find out why; i’ve found multiple places where a padlock was still around as an icon and replaced it, but there must be another one, somewhere.
So even if the code were easy…the icons are NOT. And the code probably isn’t easy. So…well, it’s only a nice-to-have.
Sad to read @adw. Hope misunderstand.
Observe Nina still in
Should now be
Actually, I recently saw her tweet to @Demi that she will be back at some point, so I’m not sure.
If anyone is still following this…I’ve gone to a new color schema; thinking a little more rigorously about the colors.
It’s now based much more thoroughly on “what does it touch?” There are five basic cases for me:
- It touches nothing
- it touches wifi (thence the internet)
- it touches USB devices
- it touches ethernet (thence my home network)
- it touches SSDs (or parts of SSDs) not part of dom0’s file system, but physically located in the same box.
No qube is allowed to touch more than one of these four types of objects.
There are some qubes that don’t touch these directly; I am talking about using split-veracrypt to pass a decrypted container (that lives in one of those places) to a qube that otherwise doesn’t touch where that container is from, so we have:
6) Touches a container that’s hosted on the internet. (I don’t actually have this; that would basically mean putting veracrypt containers in the cloud, and I don’t put anything in the cloud.)
7) Touches a container on a thumbdrive
8) Touches a container on my home network
9) Touches a container on a local SSD.
OK, now to cases: Black is still used for templates, except for a few I color differently solely so they stand out. For example sys-net-wifi’s template gets colored red to match what I do with wifi-touchers.
Group 2 (touches the internet) is red. I’m typing this in a red qube.
Group 3 (touches a USB device) is orange.
Group 4 (touches the home network/ethernet) is yellow.
Group 5 (touches local SSD) is also yellow. I have to double up because there aren’t quite enough colors.
Group 6 doesn’t exist, no color assigned
Group 7 touches a container on a thumbdrive is green
Group 8&9 are blue. (In fact, they are the same qubes; on my desktop they’re group 8, on my laptop they’re group 9. So I don’t mind too much in this case having to double up on a color.)
Finally group 1, which touches nothing, is purple.
There is a somewhat closer correlation between trust and color here…but it’s still not perfect.
That does leave gray unassigned. I’m using it for oddball and special cases. For example, I have an email attachments viewer (i.e., a qube to send email attachments to to quarantine them). This qube touches nothing (so it should be purple), but it’s full of radioactive waste…so intuitively it should be not just red but glow in the dark. So I color it gray. Another use for gray is that although I won’t let a qube touch more than one thing directly, I might create one that touches a container and the internet; it could be both green and red if the container’s on a thumb drive. So make it gray.