Wikipedia:Reference desk/Archives/Computing/2016 March 4

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

March 4

I have a question about auto-correct.

I have a question about auto-correct. But I am not even sure that that is the right term. I have noticed the following happening. I type an email. I am absolutely certain there are no typo's or errors in it. I hit "send". Whenever I send someone an email, I always send myself a copy of that email. Then, when I get my copy, the email has some incorrect words in it (which I am absolutely certain I did not type). I have noticed this time and time and time again. So, my questions are as follows. (1) Why exactly is this happening? I am not even sure if this is an "auto-correct" issue. (2) If it is, when exactly does the auto-correction take place? I assume after I hit "send"? (3) And how can I avoid this issue? I don't believe that my email system has a "setting" to enable/disable auto-correct. Here is an example of something I might encounter. I will type: The candles are red. When I read the email, it will say The crumbles are red (or something like that). The words are "close" but still "off". Thanks. Joseph A. Spadaro (talk) 07:40, 4 March 2016 (UTC)

I think it would be helpful if you were to say what program you are using to send the emails.--Phil Holmes (talk) 09:13, 4 March 2016 (UTC)

AOL. America Online. Yes, I know it's ancient. But that has the email address that I use and everyone has used for the past umpteen years. Joseph A. Spadaro (talk) 21:15, 4 March 2016 (UTC)

Do you have a "sent" folder where you can compare the original with the copy. This might indicate whether the false "correction" was happening on your computer or whilst the e-mail was in transit. Dbfirs 12:41, 4 March 2016 (UTC)

With IMAP, depending on how your mail client is set up this won't necessarily help if the server is truly modifying emails. P.S. As an example of a malapropism although possibly just a temporary memory failure as I'm tired, when thinking of this answer I initially thought "with IMDB" before realising that was wrong.... Nil Einne (talk) 17:27, 5 March 2016 (UTC)

I have never ever under any circumstances seen a mail server that performs auto-correction on emails that pass through it. I have seen email programs (the program on your computer/phone/laptop/tablet) that perform autocorrection when you click "send." So, what program are you using? What device are you using it on? I know it is hard to believe, but nobody that answers questions here is omnipotent. 209.149.114.145 (talk) 13:07, 4 March 2016 (UTC)

... or omniscient? I agree that it is most unlikely to be the mail server, but I made that suggestion to get Joseph to check this strange effect. Autocorrection that changes candles to crumbles seems very strange. It sounds more like malware to me, but we can't be sure until we see some more "corrections". Dbfirs 14:29, 4 March 2016 (UTC)

Did you put the apostrophe in "typo's" - or are you going to blame that on Wikipedia? :-) Matt Deres (talk) 14:47, 4 March 2016 (UTC)

Are you sure you're not the victim of a practical joke? Some email clients (like Outlook) will allow the user to manually add words to the auto replace list. So it's probably worth having a look in the settings to see if there are any odd entries.2.120.121.251 (talk) 17:51, 4 March 2016 (UTC)

No. No practical jokes. No one has any access to my computers except me. Joseph A. Spadaro (talk) 21:18, 4 March 2016 (UTC)

The NSA disagrees. The Quixotic Potato (talk) 22:20, 4 March 2016 (UTC)

Try this for a while: when you're sure your message is correct, take a screenshot of it (or if it's easier for you, photograph the screen with a digital camera). Later, when you find a word that you're absolutely certain you did not type, compare it to the screenshot. I'm betting you'll find that you did type it and just didn't notice. Proofreading something you just wrote is hard, And if I'm wrong, at least you'll now have the evidence to prove it! --69.159.61.172 (talk) 01:13, 5 March 2016 (UTC)

No, I disagree. Because the words are so nonsensical. They are not spelling errors or typo's. They are completely different words. Like the example I gave with "candles" or "crumbles". There is no way my fingers would type "crumbles" when my mind is thinking "candles". Joseph A. Spadaro (talk) 03:56, 5 March 2016 (UTC)

You might be right. Then again you might be surprised. No harm in trying my suggestion. --69.159.61.172 (talk) 07:26, 5 March 2016 (UTC)

I agree with 69 here. Perhaps JAS's mind which differently from mind but I definitely type a completely different word that sounds similar (not necessarily a complete homonym) on occasion. I'm pretty sure there's a name for this, but I'm lazy to search since it's off topic.

P.S. Of course from what I can tell you haven't completely ruled out local autocorrect being involved anyway. Taking a screenshot won't differentiate between the two except for local autocorrect happening after you take the photo (unlikely unless you type more or your auto correct really happens locally but after you click send). Admitedly if it's webmail then local/remote isn't really different although it's still unlikely to happen after you click send. As has been said, proofreading is hardly infalliable. In fact if you're the only one proof reading it's quite easy to miss obvious errors. Errors which once you see them you think how the heck did I miss that. (The reason is at least partially because you knew what it was supposed to be.) Having someone else proofreading is far better although obvious errors can still slip through. You should at least test simple autocorrect like that mentioned above e.g. candles always being changed to crumbles. Preferably soon after the problem is noticed in case your computer is not as secure as you thought. But even if that's a negative, it could still be where you did make a typo and autocorrect guessed the wrong word. I'm presuming you're not using a touchscreen keyboard since if you are, it's going to be very hard to say what could have happened although crumbles is quite far from computer on a standard keyboard.

Nil Einne (talk) 17:43, 5 March 2016 (UTC)

Ah I remember now, I was thinking of Malapropism Nil Einne (talk) 17:14, 5 March 2016 (UTC)

As a bonus question, is there a term for when you type an antonym or something similar? I can't remember any personal examples offhand, but an example I found while searching was typing answer instead of question or vice versa. Double bonus for when you type a word which isn't really an antonym nor does it sound similar but was simply something you were thinking of at the time. Not a perfect example since I was thinking of words which weren't necessarily related to what you were typing, but above I initially typed "your your" instead of "how your".Nil Einne (talk) 17:27, 5 March 2016 (UTC)

Thanks, all. Next time I see this occur, I will dig deeper -- looking in my sent box, etc. Thanks. Joseph A. Spadaro (talk) 17:42, 5 March 2016 (UTC)

Turn off horizontal scrolling (magic mouse)

I recently upgraded my OS to El Capitan, and my settings went back to default, so I have horizontal scrolling again with my magic mouse. As I did a few years back, I searched, found this and followed the instructions: I opened terminal entered

defaults write com.apple.driver.AppleBluetoothMultitouch.mouse MouseHorizontalScroll -bool NO

and then turned the mouse off and then back on. But it hasn't worked. Any suggestions? This "feature" is maddening.--Fuhghettaboutit (talk) 13:42, 4 March 2016 (UTC)

For others: After further research it appears it is impossible to turn off in El Capitan and of course Apple cannot be bothered to fix this problem, turn it off as the default, or create a toggle to run it off on and off when, AFAICT, no one likes or would want horizontal scrolling and it has persisted for years with thousands of complaints. The solution I've found is the overkill program MagicPrefs, 99% of which I have no use for, but which allows you to turn off horizontal scrolling.--Fuhghettaboutit (talk) 14:41, 5 March 2016 (UTC)

Calculating 11 "consecutive" colours

I'm trying to plot a population map of the world. I've divided each country's population into 11 classes, so now I just need to pick 11 "consecutive" colours (in hex color format) of the same hue. How do I go about doing this? Johnson&Johnson&Son (talk) 14:36, 4 March 2016 (UTC)

Use a tool like this one to design a gradient. The theory behind it is here. It will give much better results than naive RGB or HSV interpolation. -- BenRG (talk) 02:30, 5 March 2016 (UTC)

Looks interesting, but note that the "Correct lightness gradient" option doesn't work well with multiple color stops (more than 2). StuRat (talk) 04:03, 5 March 2016 (UTC)

I think it works with any number of stops, but they need to be in ascending/descending brightness order; it won't rearrange them. -- BenRG (talk) 04:01, 6 March 2016 (UTC)

Consecutive isn't a very well defined term in colours. However, one way of getting colours of a similar hue but different saturation would be to start with, say, red (255,0,0) and increase the amount of green and blue for the next colour along: (255, 20, 20); (255; 40, 40); etc. Increasing by 20 each time may end you up with red that's too light for your taste, but you can adjust the increment to get what you want.--Phil Holmes (talk) 16:05, 4 March 2016 (UTC)

Agreed. If you are using the RGB color scheme and have 256 levels of each, then you want to change one, two, or all 3. With 11 colors, there are 10 gaps between them, so 256/10 = 25 or 26. So, I would add 25 or 26 at each step. For particular colors, see chart:

COLOR            Start         End
=====            ===========   ===========
Black to white     0,  0,  0   255,255,255

Black to red       0,  0,  0   255,  0,  0
Blue to purple     0,  0,255   255,  0,255
Green to yellow    0,255,  0   255,255,  0
Cyan to white      0,255,255   255,255,255

Black to green     0,  0,  0     0,255,  0
Blue to cyan       0,  0,255     0,255,255
Red to yellow    255,  0,  0   255,255,  0
Purple to white  255,  0,255   255,255,255

Black to blue      0,  0,  0     0,  0,255
Green to cyan      0,255,  0     0,255,255
Red to purple    255,  0,  0   255,  0,255
Yellow to white  255,255,  0   255,255,255

Black to yellow    0,  0,  0   255,255,  0
Blue to white      0,  0,255   255,255,255

Black to purple    0,  0,  0   255,  0,255
Green to white     0,255,  0   255,255,255

Black to cyan      0,  0,  0     0,255,255
Red to white     255,  0,  0   255,255,255
     

You can, of course, reverse the order and decrease from the end value down to the starting value by the same step size. Also, anywhere I have an unchanging value, you can set it to anything you want (I only showed 0 and 255). For example, if you wanted shades of brown, you would need to do this. And the values you vary don't always have to vary from 0 to 256, but the less variation you have the less obvious the transitions between colors will be. You could also increase one or two of the RGB values, while decreasing one or two others. So, you have quite a wide range of options. If you can tell me what starting color and ending color you would like, I can be more specific.

Also, do you need help converting the decimal values I used in the chart to hex values ? StuRat (talk) 17:13, 4 March 2016 (UTC)

To illustrate further, let's take the last case of red to white and expand out all 11 colors:

Decimal RGB   HEX RGB
===========   ========
255,  0,  0   FF, 0, 0  red
255, 26, 26   FF,1A,1A
255, 51, 51   FF,33,33
255, 76, 76   FF,4C,4C
255,101,101   FF,65,65
255,127,127   FF,7F,7F  pinks
255,152,152   FF,98,98
255,178,178   FF,B2,B2
255,203,203   FF,CB,CB
255,229,229   FF,E5,E5
255,255,255   FF,FF,FF  white

After you display the colors, you might decide that you want more difference between some shades and less between others, so you can tweak it to get it just right. Note, however, that 11 shades of one color will make each one fairly close to it's neighbor. You might instead want to take a "rainbow" approach, and go from red to orange to yellow to green to cyan to blue to purple to violet. That's only 8, so some in between colors would also be needed. Let me know if you want more details on this approach. StuRat (talk) 17:55, 4 March 2016 (UTC)

Based on the previous comment, you can begin with a start color and an end color. Lets say you want to start with 190,23,255 and end with 10,90,100. You will have 10 divisions to get 11 numbers (draw 11 dots and look at how many divisions there are between the dots). So, divide the difference of each color by 10. 10-190=-180 and -180/10=-18. So, you start with read=190 and subtract 18 for each color until you to the last one, which will be 10. Similarly, for green we have 90-23=67 and 67/10-6.7. So, I begin with 23 and add 6.7 for each color. I will likely want to round the result to get an integer. For blue, I have 100-255=-155 and -155/10=-15.5. So, I subtract 15.5 for each color. I will get 11 "consecutive" colors from the start color to the end color. 209.149.114.145 (talk) 17:03, 4 March 2016 (UTC)

It's usually better to reason, and calculate, about colours in a space like HSV, even if one has to use an RGB system for the final output. For example, this Python3 program lerps the Value parameter, holding Hue and Saturation constant:

#!/usr/bin/python3

import colorsys

hue = 0.2
sat = 1.0

def frange(x, y, jump):
    # from http://stackoverflow.com/questions/7267226/range-for-floats
    while x < y:
        yield x
        x += jump
                  
for v in frange(0.0, 1.0, 0.1):
    r,g,b = map(lambda x: int(x*256),
                colorsys.hsv_to_rgb(hue, sat, v))

    print ('<div style="width:200px; height:30px; background-color:#{:02x}{:02x}{:02x};">&nbsp;</div>'.format(r,g,b))

which produces a range like:

colours
The following discussion has been closed. Please do not modify it.

Tweak the H and S values, and the parameters to frange for different results. -- Finlay McWalter··–·Talk 17:02, 4 March 2016 (UTC)

Thank you all for the great help, guys. Johnson&Johnson&Son (talk) 02:21, 5 March 2016 (UTC)

Resolved

Program automation tools in Linux

Does Linux have well-developed tools like AutoIt or AutoHotkey (for programmatically reaching the GUI as an API)? Does it need them? That is, are there main Linux tools that cannot be scripted, nor can be accessed through the command-line or an API? Some tools have a GUI version and a command-line version, some only have a command-line version. But are there programs with only a GUI version? --Scicurious (talk) 18:48, 4 March 2016 (UTC)

I'm sure there are GUI tools that don't have CLI or API interfaces - it's hard to prevent that in a world where anyone can write and publish a piece of software. I'd say that the need for those gadgets for forcing scripting onto GUI tools is much less important in a world where there are so many good CLI tools.

That said, I believe Sikuli is available for Linux and it should do what you need.

SteveBaker (talk) 21:19, 4 March 2016 (UTC)

Google Chrome tab closure behavior

Let's say I have windows 1, 2 and 3 open. On tab 1, I then click on something which causes a new tab to open. So, at this point I have tabs 1, 1A, 2 and 3 open. I then close tab 1A by clicking on the X in the upper, right corner. I would then like to return to the tab which spawned it (1), but instead tab 2 becomes active. Any way to change this ? StuRat (talk) 19:05, 4 March 2016 (UTC)

Ehm, yeah, quite a few ways to change that. Not sure if you can find a plugin that sends you back to the tab that spawned the tab that you've closed, but there are many plugins that allow you to change Google Chrome's tab closure behavior. If you Google those words you'll find quite a few. Here an example. The Quixotic Potato (talk) 23:42, 4 March 2016 (UTC)

Thanks. So there's no setting to do this then, without a plugin ? StuRat (talk) 22:22, 5 March 2016 (UTC)

Not that I know of (and I couldn't find one with a quick Google search). Maybe there is a Google Chrome guru around here? The Quixotic Potato (talk) 12:10, 7 March 2016 (UTC)

I'm a little unclear since this is the way my Chrome works by default. Or, I should say, if I don't click on any other tabs, it goes back to tab 1. If I click on tab 2, click back over to 1A, and then close 1A, I get tab 2 again since that was my last viewed tab. Dismas|^(talk) 16:16, 7 March 2016 (UTC)