Welcome to the computing section
of the Wikipedia reference desk.

Select a section:

WP:RD/CWP:RD/C

Want a faster answer?

Main page: Help searching Wikipedia

How can I get my question answered?

Select the section of the desk that best fits the general topic of your question (see the navigation column to the right).
Post your question to only one section, providing a short header that gives the topic of your question.
Type '~~~~' (that is, four tilde characters) at the end – this signs and dates your contribution so we know who wrote what and when.
Don't post personal contact information – it will be removed. Any answers will be provided here.
Please be as specific as possible, and include all relevant context – the usefulness of answers may depend on the context.
Note:
- We don't answer (and may remove) questions that require medical diagnosis or legal advice.
- We don't answer requests for opinions, predictions or debate.
- We don't do your homework for you, though we'll help you past the stuck point.
- We don't conduct original research or provide a free source of ideas, but we'll help you find information you need.

Ready? Ask a new question!

How do I answer a question?

Main page: Wikipedia:Reference desk/Guidelines

The best answers address the question directly, and back up facts with wikilinks and links to sources. Do not edit others' comments and do not give any medical or legal advice.

February 1

Unspecified cellphone GUI

Please see File:Emergency alert.jpg, a cellphone (model not specified) displaying an alert sent out by a US federal government agency a few minutes after the 2018 Gulf of Alaska earthquake. Without knowing the model of cell phone, is it possible to determine how much of the display is the work of the cellphone's software's manufacturer, versus how much is {{PD-NWS}}? I assumed it was the latter, but someone submitted an OTRS that failed to provide permission, so clearly someone's expecting it to be the former. Nyttend (talk) 23:16, 1 February 2018 (UTC)[reply]

That looks like an Android status bar at the top, and as such it's probably under the Apache 2 License and conceivably ok for Wikipedia. See Android Open Source Project. 173.228.123.121 (talk) 08:54, 2 February 2018 (UTC)[reply]

I'm not sure any interface elements in there meet the threshold of originality to be copyrighted in the U.S. Of course the only way to know this for sure is to get a court ruling. Realistically, there's not the slightest chance that anyone who might own a possible copyright to any elements of the image would sue over it or even file a takedown request. If the "freedom or death" crowd make noises, just slap a fair use template on it. --47.157.122.192 (talk) 10:11, 3 February 2018 (UTC)[reply]

No way in hell that would pass under NFCC rules. It's also on commons, so you can't simply slap a template on it anyway. You'd need to upload it here and delete it on commons. Nil Einne (talk) 13:13, 5 February 2018 (UTC)[reply]

February 2

Power consumption of the same operation in two microchips

If you pick two microchips of different processing power, would they spend the same amount of energy for performing the same operation? --Hofhof (talk) 18:58, 2 February 2018 (UTC)[reply]

No. In particular, you've specified "the same operation" rather than "the same operation in the same time". It's very common that the same operation (in bulk) increases in power consumption with speed. Andy Dingley (talk) 22:23, 2 February 2018 (UTC)[reply]

February 3

What's the point of the "All Files" option for "Save link as..." in Chrome?

I'm using Chrome in Windows 7. When I right-lick on a link I'm given the usual options (starting with "Open link in new tab"). If I pick "Save link as..." Windows Explorer opens and I'm given the choice of saving the link either as a "Chrome HTML Document" or as "All Files". What is that "All Files" thing for? When I try it I sometime get a file with no file type. Now when I open it with Chrome then the result is the same as if I had saved it as a "Chrome HTML Document". So it is saved as an HTML Document no matter what. So what exactly is the point? Thanks. Basemetal 15:03, 3 February 2018 (UTC)[reply]

It allows you to choose file extension. Ruslik_Zero 20:21, 3 February 2018 (UTC)[reply]

You should be able to partially choose extension without it. If you give the file a name like 'save file.txt' or 'save file.com' or 'save file.jpeg' or something with a recognised extension (I think recognised in Windows Explorer), it will not add an extension. If you give the file a name like 'save file.meow' it will probably add the extension, so the final file name is 'save file.meow.html'. If you choose the 'All files' option, it should never add an extension. The all files option will also show all files (except those hidden due to Explorer settings) in the target location. Otherwise you will only see those with the .html extension unless you put the * wildcard into the file name box and hit enter. Nil Einne (talk) 04:28, 5 February 2018 (UTC)[reply]

Thanks guys. I have another question: When you're displaying a page in Chrome and you right-click on it then the options you are given for saving (that is when choosing "Save as...") are "Webpage, complete" and "Webpage, HTML only". On the other hand when you right-click on a link in a page you get (when choosing "Save link as...") what I said above (namely "Chrome HTML Document" and "All files"). So my question is: Why the difference? Why not behave the same when right-clicking on a page and when right-clicking on a link in a page? Thanks. Basemetal 19:03, 5 February 2018 (UTC)[reply]

The laziness of Chrome's programmers. 78.0.213.2 (talk) 04:32, 6 February 2018 (UTC)[reply]

Not necessarily. The are obvious implications for changing the behaviour. 'Save link as' just saves the link whatever it is (e.g. a 7zip, PDF, txt, aac, jpg) with minimal processing or interpretation. I believe it follows stuff like HTTP 303 but not much more. When you use 'save as' with a loaded page, and choose the 'complete' option, it saves the extra loaded files that are necessary to display the page, generally including images, css, fonts and possibly even some scripts in a directory with the same name and modifies the HTML file to refer to these local copies. It's able to do this because the page has already been interpreted (and rendered). It would be possible to do this without displaying the result, but I'm not sure whether that's something which really already coded for i.e. it may be more than just a few lines of code. Further while I wouldn't recommend you actually do it with any untrustworthy page, simply saving a link should generally be fairly safe with fairly limited risk from browser vulnerabilities. By adding the 'Complete' option to the save link and interpreting the page, you are increasing the security risk whenever that is used whether by accident or on purpose. Note although AFAIK it would be possible to 'save complete' without properly rendering the page (e.g. interpreting the images and fonts), I wouldn't be surprised if this is simply something Chrome, as with any browsers, does not do. There are plenty of spidering type programs which can do this if it's desired. While it may seem weird considering how much people encounter what they consider feature bloat, I'm fairly sure developers do actually consider whether they should bother to add a feature for reasons besides laziness or time, such as the risks of adding more complexity to the code. Nil Einne (talk) 05:12, 7 February 2018 (UTC)[reply]

When to use what programming paradigm

Besides [1] I am having trouble finding more precise examples about when to use what programming paradigm. --Hofhof (talk) 19:34, 3 February 2018 (UTC)[reply]

You can look at Comparison_of_programming_paradigms. Ruslik_Zero 20:25, 3 February 2018 (UTC)[reply]

I already took a look at it. I was hoping for something more specific — a concrete example about a project that decided for this or that paradigm. --Hofhof (talk) 20:43, 3 February 2018 (UTC)[reply]

Often it will be guided by the system on which the end program will run. eg interactive systems Microsoft Windows, iOS will be event driven as the user does things, eg types, clicks, touches, turns the phone. A program on a mainframe to pay workers may be procedural, ie print out the check, move money, update account. But this may also be time driven- once a fortnight. A building automation system might be declarative, eg keep the tempeerature between 22 and 25°C, have lights on when there are people in the office. Graeme Bartlett (talk) 22:21, 3 February 2018 (UTC)[reply]

Modern desktop applications, and the operating systems that run them, are invariably object-oriented. This includes Windows, Linux, MS Office, Firefox, etc., and is because these programs all use large portions of existing code: the object-oriented paradigm is the only way to do this at scale. As such, it's really more of a super-paradigm, in that the other paradigms listed will also be object-oriented. That said, the only real difference I see today is between functional languages such as Python and Scala, and imperative languages like C++, Java, C#, etc. Even then, other than a few esoteric languages like Haskell, programming languages are a blend of paradigms, because programs have to do multiple things, and this gives the language (and the programmer) flexibility to choose the paradigm best suited to the task at hand.OldTimeNESter (talk) 14:45, 4 February 2018 (UTC)[reply]

In 40 years of programming, I've never had the choice of a paradigm (or even a language) unless it was something I was doing on my own. Professionally, management chooses everything. The programmer is just a code monkey. So, how does management choose? They invest in a paradigm or a language. Once invested, they need to show return on the investment. To show return, the programmers must use the paradigms and languages (and packages and tools, etc...) provided. It is a choice of what is best. 71.85.51.150 (talk) 19:36, 4 February 2018 (UTC)[reply]

February 5

Another HTML question

I've noticed something curious on a site I've been visiting (let me call it "www.site.com"). Going to www.site.com/directory/ gives (on that site) precisely the same result as going to www.site.com/directory/index.html as long as there is indeed in that directory an "index.html" file. On the other hand when such an "index.html" file does not exist in that directory, then going to www.site.com/directory/index.html gives a 404 page not found error and going to www.site.com/directory/ gives a directory listing. Is that completely standard HTML behavior? What I'm asking is: Does by default any "index.html" file in a directory become synonymous as a link to that directory? Can that be changed? For example is it possible to make say a "contents.html" file (instead of a "index.html" file) behave the same? What if there is an "index.html" in the directory but I want to force the link www.site/directory/ to give a directory listing nonetheless? Does the "index" file have to be an "html" file (not say an "htm" file)? Thanks. Basemetal 20:05, 5 February 2018 (UTC)[reply]

This behavior has nothing to do with HTML. It is just your web server. Ruslik_Zero 20:44, 5 February 2018 (UTC)[reply]

What you are describing is default Apache behavior. Apache is one webserver program. There are others. Some behave similar (though not exactly like) Apache. Some behave in an entirely different way. By default, if you do not declare a file, it will look for index.html and show it. If index.html does not exist, it will show a directory listing using default Apache icons for folders and files. If you declare a file that doesn't exist, such as index.html, you will get a 404 error. All of that can be customized. You can customize the default page to be a different file name or even a set of files. You can say "look for index.html or index.htm or index.php" if no file is given. You can customize the 404 error to go to an actual page. You can turn off directory listings or redirect directory listings to another page. A lot of people don't do that, but it is possible. 209.149.113.5 (talk) 20:45, 5 February 2018 (UTC)[reply]

Why can't I play embedded YouTube videos from webpages I've saved?

I've saved a webpage which contains an embedded YouTube video. That video plays fine on the site I've saved it from. However when I try to play it from the copy saved to my hard disk, it never works. All I get is that stupid TV static screen containig the YouTube error message: "An error occurred. Please try again later. (Playback ID: <16 character playback ID>). Learn More". And the "Learn More" link is useless. Can someone help, or at least explain? Thanks. Basemetal 20:49, 5 February 2018 (UTC)[reply]

The webpage doesn't "contain" the Youtube video, it links to it. As you didn't actually download the video and saving the webpage to your hard drive probably broke the embedding mechanism, it doesn't work. You cannot download a Youtube video simply by saving the page. 78.0.213.2 (talk) 04:34, 6 February 2018 (UTC)[reply]

Despite the implication of the answer to my query just above, I of course know that the actual video data of the embedded YouTube video doesn't reside on my saved webpage or on the original webpage for that matter and that I'm not actually downloading a YouTube video simply by saving a webpage that has an embedded YouTube video. My use of the word "contain" did not mean to imply anything of the sort. I didn't say "contain a video" but "contains an embedded video". That might be ambiguous but I've seen that phrase used everywhere and everybody understands what it means. If anyone has a better way to say it, I'm open to suggestions. Again: I know it is just a link to the YouTube video which resides on the YouTube server. Whew. I hope that's clear and out of the way now.

However the error message comes from YouTube. YouTube is incapable or unwilling to play the link when the request comes from my saved page as opposed to when it comes from the original webpage and sends me an error message. Why? Is there a way to fix that? The statement that "saving the webpage [...] probably broke the embedding mechanism", whatever it means, is too vague to be useful, as it is basically tautological. Yes, I know something "broke". That's why I'm asking questions here. The "embedded video" is still there (you know what I mean, the YouTube screen) and the link to the YouTube video is there too (because I can save the link with "Copy video URL" and then play it fine if I open another window and use that URL) so what "broke" and why? . Again the link to the YouTube video is still there on my saved page because the error message obviously comes from YouTube and because I can play the URL saved from the "embedded video". If anyone can help please ignore the previous non-answer and let's start again.

Thanks a lot.

Basemetal 05:22, 6 February 2018 (UTC)[reply]

So are you saying that your saved webpage contains a link back to YouTube, and when you click that link, YouTube tells you it doesn't work? And if you copy the link URL and paste it to a web browser to go to YouTube directly, it works? In such a case, one thing that comes to my mind is HTTP referers. Most browsers use a HTTP header denoting the referer, meaning which site the link that caused the request was located at. It might be that YouTube expects the referer value to be something different than what gets sent when you click the link from your saved webpage. JIP | Talk 10:07, 6 February 2018 (UTC)[reply]

How can one check that? Could you check you get the same problem when you save a webpage with an embedded video? Just to confirm everybody's got the same problem, which would sort of confirm your guess. If you've got the time, could you try with this Memsaab's Story page which has this Daily Motion video embedded. First check the video plays fine on that page when it's on the Memsaab's Story site, then save the page and try to play the same video from the saved copy. If it works for you then it's gotta be something else. In this case the error message I get is "Whoa, hold up" and "Yikes! There's been a server error". Contrary to the YouTube case the error message appears immediately, before I even try to play the video. Thanks. Basemetal 14:10, 6 February 2018 (UTC)[reply]

PS: The YouTube video I was having problems with was on this page. But as I said, it seems to make absolutely no difference whatsoever if its YouTube, Dailymotion, Vimeo, or whatever. Basemetal 14:21, 6 February 2018 (UTC)[reply]

Yes, it's not working when I save the page as "Web page, complete". I get the video's start picture and a button on top of it, but the button does nothing. If I save as "Web page, HTML only" it works. The same behaviour as you have. It appears that the video is linked from a <video> element and not from a normal hyperlink, so I don't know if it even uses HTTP headers or HTTP in the first place. I tried saving a YouTube page as "Web page, complete" and it appeared completely blank in my browser. So I think it might be something more complicated than HTTP headers. JIP | Talk 19:35, 6 February 2018 (UTC)[reply]

This is because the world hates you and Donald Trump doesn't listen to Tim Berners-Lee.

In the TB-L web, "information wants to be, if not free, then at least neutral about it". YouTube would serve you a page of content, and links to content. If you saved that content (and the content that was at the far end of those links) then you could replay it offline. Clearly one can't do this.

One reason is streaming. There is no "content" at the end of those links, instead there is a server which dynamically serves portions of the content, as your playback client requests it. There just isn't any single monolithic downloadable resource for you to store. This is broadly a good thing - streaming allows us to watch the start of videos, or to skip to the middle, without first having to download and store the whole piece of content. With plausible values for bandwidth and storage space, this is much more useful, to more people. There are ways round this: the site may offer a different link, to a downloadable copy. Or you may use a client-side video downloader tool (there are many browser plug-ins) which can stream the video down, but also store it to such a format.

In the YouTube case though, it's because YouTube does not want you to download the videos. You might play them again without having to watch YouTube's ads. You might even upload them to a different site, taking away YouTube's business.

So the YouTube "web page" is far from a simple web page. It doesn't simply "link to" content (as TB-L's vision of an open web had it) it hides those links as closely as it can. It wraps them up in a client-side video player (maybe coded in Flash, or something equally opaque) to "improve the video watching experience" but mostly to keep them away from you. This player layer can impose DRM, location restrictions, age restrictions, or pay-per-view.

Trump's recent support of abandoning net neutrality means that this is the direction the web and internet are moving in. Expect to lose some YouTube content altogether - valuable content will instead retreat inside walled garden sites, where access needs to be bought. We're heading back to the days of AOL and Compuserve. Andy Dingley (talk) 10:31, 6 February 2018 (UTC)[reply]

Although Youtube does I believe use MPEG DASH, it is generally possible to get a link to a single file albeit not simply from the page. If you use something like KeepVid for example, you should normally be able to get a simple link which should be downloadable with any browser or even simply tools like Wget without plugins. (The site has Javascript which your browser runs to get the appropriate link.) The Youtube T&C (including for the API) do forbid downloads of this sort so, apps on Android, Windows Store, iTunes which allow such offline playback other than via the YouTube Red function do generally find themselves in trouble particularly if they are prominent e.g. Microsoft, Amazon. Plugins and third party software may do the same thing or they may have support for MPEG DASH or HTTP Live Streaming which may mean more quality options. Nil Einne (talk) 05:25, 7 February 2018 (UTC)[reply]

The browser's "save as" function saves a bit too much stuff. "Save as" fetches and saves a file from youtube.com. That file contains a bit of JavaScript that loads the embedded player. The JavaScript can't find the player because it doesn't know that it needs to go to youtube.com - it searches for it on your hard drive and fails. The JavaScript tells the browser "load the player from the same place you loaded this file", which is wrong when saved.

To fix, load the saved html file in a text editor. Find the line that starts <span class="embed-youtube". Open the original web page, do "view source" on it, and find the same line. Copy and paste the original line into the saved file. You'll notice there is a src="blah blah" - that is the bit that broke. 91.155.192.188 (talk) 16:40, 6 February 2018 (UTC)[reply]

I should have made clear I was saving the webpage on a computer running Windows 7 as "Webpage, Complete" because that's what caused the problems. When I save as "Webpage, HTML Only" then, as I've just realized, everything is fine and the video does play. So it is saving it as "Webpage, Complete" that breaks things. I discovered that by trying to follow your instructions and realizing I couldn't see the code you said. So when I save the page as simply an HTML file it works. Now the question is, is it possible to fix things so that even when I save as "Webpage, Complete" it will play the video? Because there are advantages to saving the page that way. For one thing if I save as an HTML file only then all the pictures are gone. The options "Webpage, HTML Only" and "Webpage, Complete" come from Explorer. I don't know what OS you're running. The JIP might be surprised to discover that on his Linux system everything works fine... Basemetal 17:21, 6 February 2018 (UTC)[reply]

Question: (regarding to Andy Dingley) this thing about streaming vs. loading the whole video. Could it also depend on the browser? I used to watch a video with the Task Manager window open. When using IE, the under Performance, the physical memory usage history (green line) goes uphill direction, until the computer slowly freezes, whereas when I view a video using Google Chrome, the memory usage line stays the same. So it seems like Chrome knows how to stream the video, but this is prolly with an older version of IE. Thanks. 12.130.157.65 (talk) 14:40, 8 February 2018 (UTC).[reply]

Clicking on links in e-mail goes to my home directory

My home computer's Linux system (Fedora 26) developed a strange problem just today. Whenever I click on any HTTP or HTTPS link in my e-mail, the browser goes to my home directory on my local computer (file:///home/jip/). The links themselves are OK, as copy-pasting the link to the browser works OK. Stopping and re-starting both the e-mail client and the browser didn't help. I use Evolution (3.24.6) as my e-mail client and Firefox (57.0.1) as my browser. Does anyone have any idea what is causing this? JIP | Talk 21:23, 5 February 2018 (UTC)[reply]

I use Thunderbird now, but when I used Evolution, I used xdg-settings to set the defaults. Try "xdg-settings --list" in your terminal to see if you have a default-url-scheme-handler and a default-web-browser set. There is no reason that the settings should have broke just to break. So, that is the confusing part. The goal is to track down the setting that borked and fix it. This is the starting point. My xdg default-web-browser is "Default web browser", which is the web browser set in my Xfce defaults. 209.149.113.5 (talk) 18:36, 6 February 2018 (UTC)[reply]

It appears that the problem might have started when I made Firefox my default browser in the Firefox preferences. But then I went to "Preferred applications" in GNOME preferences and explicitly checked Firefox as the preferred web browser (it originally showed "Preferred web browser", literally) and the problem went away. I went to the same preferences again, and the problem came back. It appears explicitly unselecting Firefox from the "Preferred applications" preferences and then explicitly selecting again fixes this. "xdg-settings --list" shows the same output every time, listing "Default web browser" as the default web browser. JIP | Talk 19:31, 6 February 2018 (UTC)[reply]

February 6

Web design survey data

Background: Long ago, the standard for professional web design was to have a splash page with nothing but a logo. You had to click through that to get to the website. Everyone did it because everyone did it. When it was clear that everyone hated splash pages, that practice stopped. In recent design, every professional page had a small header, a huge wide-screen image that really had nothing to do with the website, and a set of blurbs at the bottom. To get to the content, you had to click a blurb or hunt down the menu. It is clear that the huge pointless image isn't helping, so that is vanishing. Modern web design is going a step further. You hit a website and you get a full-screen image (sometimes a video). You swipe it up to get another full-screen image and, possibly, a very tiny blurb. Swipe that up and you get another full-screen image and, if you are very lucky, another very little blurb. After a dozen or so screen swipes, you might figure out what the company does and what it has to offer. My gut feeling is that people don't prefer the current cool web design any more than they liked the older bad designs. But, how can I quantify that?

Question: I'm looking for surveys on how users like design choices. I know about the 3-click rule surveys and about the F-layout surveys. I'd like something more general. I'm hunting through scholar.google.com, where I always start, but I thought someone here might know of a good resource I could look at. 209.149.113.5 (talk) 20:02, 6 February 2018 (UTC)[reply]

February 7

Getting statistics from Apache Subversion

Is there a way to find out what programmers actually do? I'd like to see who doesn't submit much anyway, who's submitting but whose code gets changed soon afterwards by someone else, etc. There's not much I can find. Joepnl (talk) 01:24, 7 February 2018 (UTC)[reply]

The usual approach goes a bit further than this. svn will tell you who did what, but it won't tell you why. So it's common practice to link this to your issue tracking system, by embedding the ID for the issue tracker into the svn commit comment as "[1234] Fixed the exploding foobar problem". You can then use svn log on a build tag to get a changelog between builds. Use some scripting to process all that up, link it to your issue tracker, and you can slice and dice it every which way. Some people do this with external scripts, other push it into a database (sometimes permanently, better just doing it temporarily) and then query across both the issue tracker and changelog databases.

You'll also need a commit script which checks that the coders have given a validly formatted issue ID on each commit, or else they get an electric shock and no food pellet. Andy Dingley (talk) 03:04, 7 February 2018 (UTC)[reply]

I wish I thought about the electric shock before! I don't really care why someone's code got reverted or changed afterwards. One change may or may not have been their fault. After 1000 submits in a year, of which 800 were reverted or changed within a month, I think that's a nice metric to seperate the men from the boys. I'm looking for a tool that can show just that. There's certainly more to that than a simple script. A diff between two files isn't easy. A diff over 50 versions showing whose code was actually changed is something that needs a tool that must exist but I haven't found. Joepnl (talk) 04:15, 7 February 2018 (UTC)[reply]

The usual algorithm is to use svn log to generate a set of issue/bug IDs which were "affected" by commits between those two build tags. Then to link the current status of those issues. That's two linear lists, not hard to do.

If you want to look at history per issue, then you need to produce a list-of-lists of the commits involved for each issue (scanning the log comments - a commit may involve multiple issues). You might then filter that to particular issue(s). You now have a list of commit IDs. svn can then give you a list of changed files for those, and you can then use svn to get the file histories (between the build tags) for each file. The only hard bit are the file content changes, and svn diff does the legwork for those. Andy Dingley (talk) 11:01, 7 February 2018 (UTC)[reply]

February 8

63-pass compiler?

At 22:24 into this video about FORTRAN II on an IBM 1401, he says that it is a 63-pass compiler. I know about multi-pass compilers, but why so many passes? Bubba73 ^{You talkin' to me?} 00:48, 8 February 2018 (UTC)[reply]

Well, this is addressed in the IBM 1401 FORTRAN section of FORTRAN. Bubba73 ^{You talkin' to me?} 01:23, 8 February 2018 (UTC)[reply]

When memory was measured in bytes, and programs that run instantly on today's PCs would take hours to finish, you went to some lengths. See also The Story of Mel; although it recounts an extreme case, such techniques were not that rare in those days. --47.157.122.192 (talk) 06:37, 8 February 2018 (UTC)[reply]