Wikipedia:Reference desk/Archives/Computing/2016 April 11

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.

April 11

Internet usage

Sometimes my son close the broadband connection and shutdown the computer, the internet usage goes up by 200-300 mb. How can that happen? Then it may be possible somebody use the network even when he is connected to it. How can we stop that? Thank you.175.157.35.52 (talk) 03:00, 11 April 2016 (UTC)

How are you checking your usage? Most "usage meters" I've seen aren't really an instant snapshot. It's quite possible if you check your usage before you turn off your computer, that the usage meter hasn't finished "counting" all your usage until a few hours after you've turned off your computer, which you will only see once you've turned on your computer again. Vespine (talk) 04:27, 11 April 2016 (UTC)

You could check this by looking at the meter after a few hours with the internet off, then shut down the computer again for a few more hours before checking again. Do you have any other devices (mobile phones, tablets, TVs, security devices ...) that might be using your broadband connection? Do you have a WiFi router that is not protected by a secure password? You could try turning that off and connecting only by wire. Does your computer do automatic upgrades when it is switched back on again? There are lots of possibilities. Dbfirs 08:49, 11 April 2016 (UTC)

Is it possible that you're seeing a bunch of automatic updates being loaded? Noting the usage stats just before you shut down - and then again just after you're fully rebooted - would reveal the amount of stuff that was uploaded by your operating system as a part of that automatic update cycle. Sometimes it's not much - sometimes it's nothing at all - and recently, if you have Windows 7, it may be Microsoft downloading the whole of the Windows 10 update onto your machine without asking you. Another possibility - if you have WiFi - is that your son, or one of his friends is using a cellphone with WiFi enabled. If he habitually uses his phone to get to the internet while his computer is shut down - then that might explain those elevated numbers while the computer is offline. SteveBaker (talk) 19:57, 11 April 2016 (UTC)

What is Pocket Windows?

I came across a redirect from 2003: Pocket Windows→Microsoft Windows. I'd like to point this to something more specific, but I'm not sure what it actually refers to. Pocket PC? Windows Mobile? Windows CE? Something else? Reach Out to the Truth 04:31, 11 April 2016 (UTC)

I can't find evidence that any Microsoft product was named or nicknamed "Pocket Windows". There are no hits in the microsoft.com or msdn.com domains, for example. Wikipedia has a lot of meaningless redirects and I suspect this is one of them. -- BenRG (talk) 05:40, 11 April 2016 (UTC)

I have listed the redirect Pocket Windows at WP:RFD. The discussion is at Wikipedia:Redirects for discussion/Log/2016 April 11#Pocket Windows. --Guy Macon (talk) 12:54, 11 April 2016 (UTC)

Why do directories have a tree structure?

Couldn't directories be organized in a grid? Or allow to put a file in two directories?--Scicurious (talk) 14:09, 11 April 2016 (UTC)

Well you can put the same file in two directories, right? ./MyDir1/MyFile1.txt and ./MyDir2/MyFile1.txt causes no problems. We even have different ways of doing it. One could be a symbolic link to the other. As for the bigger question, see Directory_(computing), file folder, Filesystem_Hierarchy_Standard.

At an even higher level see Interface_metaphor and desktop metaphor, with a healthy does of skeumorphism. So, sure, there are conceptually other ways to organize files (see e.g. Freenet for an interesting early example of distributed storage), but we are where we are largely because of history and metaphor. SemanticMantis (talk) 14:46, 11 April 2016 (UTC)

There's also the tendency of the human mind to classify and categorize things - we have natural instinct for for taxonomic/hierarchical information. Even very young children have the ability to sort objects into classes based on shared features. Roger (Dodger67) (talk) 15:25, 11 April 2016 (UTC)

You can develop a filesystem in which files (or directories) have multiple parents. The reason that isn't done is simplicity. Two brief examples. When implementing a filesystem, each file/directory has one parent. It is a one-to-one relationship, that is easy to program. All I have to do is account for the very top parent that has no parent and then everything is easy. When using a filesystem, there are many "get the parent" commands such as pwd or cd.. that expect there to be one and only one parent. If there was more than one, what would they do? They could spit out more than one directory if that is the purpose of the command (such as pwd). For cd.., it is moving a directory. Does it ask you which directory to move do? Regardless, it add complication for both the programmer and the user. 199.15.144.250 (talk) 17:55, 11 April 2016 (UTC)

Files can be linked from multiple directories in both *ix and Windows (though some filesystems don't support it, notably FAT/FAT32/exFAT). This is not a symbolic link, it's not some sort of trick, it's literally the same file appearing in more than one place in the directory hierarchy.

And directories can be symlinked, so the problem of which parent directory you want exists in real life. In *ix, the interpretation of .. in these situations is inconsistent. Usually it means the one true parent directory, but most shells special-case cd .. to strip off the last part of the current directory path (unless you use the -P option). So if /fake/foo is symlinked to /real/bar, and you're cd'd to /fake/foo, then ls .. and even echo ../* will show /real/bar, but cd .. will put you in /fake. In Windows, .. always refers to /fake. Since the Windows (NT) kernel has no concept of the one true parent directory, there seems to be no reason why NTFS couldn't support (non-cyclical) multiply linked directories. But in fact it doesn't. -- BenRG (talk) 20:22, 11 April 2016 (UTC)

"each file/directory has one parent. It is a one-to-one relationship" -- no, it's a many-to-one relationship. jnestorius^(talk) 09:31, 12 April 2016 (UTC)

What difference would organizing files in a grid have anyway? If you're simply talking about DISPLAYING them in a grid - then I guess that's an application question - and tools like Windows Explorer or the Desktop icons already let you do that. If you're talking about representing a directory as a two-dimensional structure in code - then I find it hard to imagine the direct benefits of that. I suppose that if you had something that was a grid in the real world (such as a map where each 'cell' of the grid was a separate JPEG file) then rather than naming them cell_001_001.jpg, cell_002_001.jpg, etc you might find some tiny benefit from having the directory do that - but then you'd have to concern yourself about slots in the directory that don't exist. Generally that's better handled either by having one folder representing each row of the grid - or by having a completely file that contains that stuff.

In very early versions of UNIX (and possibly other OS's too) - the idea was that the "file" was represented by a small structure on disk called an "inode" (information node) that simply had a unique number and contained the permission bits, ownership, reference-count and pointers to the actual blocks of space on the hard drive that the file occupied. Then a "directory" was just a file that happened to contain a list of inode numbers and their corresponding file names. The directory was more or less just like any other file - and programs could read and write it indiscriminately. That soon acquired layers of protection and such - but the idea is still there.

In a system like that, the file structure and the names of files are just a "convenience" layer built on top of the basic inode structure. The ability to have the same file in two directories just happened as a result of that. You just have two directory files - each referencing the same inode - possibly using the same name - but working OK with different names. You could also have the same file exist twice in the same directory, with different names - or even (confusingly) with the same name!

It would have been fairly easy back then to make a new kind of "directory" file that had a 2D array of inodes and names instead of a 1D list. The operating system wouldn't have cared too much - but just about every application program that cared about directory "files" would need to be rewritten (including the shell).

The idea was that "everything is a file" (including directories) - and that a directory was merely a convenient way to map human-readable names onto inode numbers. Even the "." and ".." entries were simply inode numbers - you didn't have to have a ".." or even a "." file if you didn't want to - and you could do ungodly weird things by point the ".." file to someplace other than the usual parent. You certainly could make directory structures that were rings or had other weird loops. The idea for the file system to be a "tree" was merely a convention.

To do this in modern system would certainly require you to rewrite the code for an ungodly number of programs that expect directories to be organized in trees and to have simple linear lists of files in them.

But if you were prepared to have your "grid" directories work in just those few programs that you'd be prepared to rewrite - then why not just make your own "directory files" that list the files in the current directory in a grid. You'd need your own equivalent of Windows Explorer (or whatever) in GUI land - your own File-Selector tool - your own special way to name a complete path as a string - and possibly your own command-line tools to implement stuff like 'cd' and 'dir'.

I believe that the reason that the old UNIX systems didn't implement structures other than trees was simply that those other things never seemed necessary - so although the OS - and 90% of the tools were more than capable of handling other structures - nobody ever felt that they were necessary.

SteveBaker (talk) 19:50, 11 April 2016 (UTC)

Microsoft Word not using available italic font, substituting a slanted version of the Roman style instead

I have a problem with Word 2013 involving a font that I installed. The font comes with a real italic variant. For some reason, when I try to use the italic style of the font, Word substitutes an automatically-generated slanted version of the Roman style instead. Other programs on the same computer don't have the same problem, and Word doesn't do that with other installed fonts. What could be the cause?

I've seen the same problem with an older version of Word years ago. In that case, the problematic font was one of the standard ones that came with Word. Thanks. --134.242.92.97 (talk) 17:44, 11 April 2016 (UTC)

Some additional info may help: How did you install the font? What type of font (e.g. perhaps TrueType?), and what type of file? Here [1] is a general overview for troubleshooting font problems in Word 2013. SemanticMantis (talk) 19:04, 11 April 2016 (UTC)

And for general reference, this is a problem where oblique type is being substituted for italic type. (PSA: LaTeX is a powerful, free word processing system that is largely bug-free, and has a very helpful support community [2] :) SemanticMantis (talk) 19:08, 11 April 2016 (UTC)

The font in question is a TrueType font. I don't remember how it was installed, but it was probably by using the "Install" option in the context menu in Windows Explorer. --134.242.92.97 (talk) 19:47, 11 April 2016 (UTC)

This reddit thread shows a user complaining about the same problem [3]. It seems he found a work-around by renaming some files, but not a true fix. It seems likely there is something off in the font ID registration properties, but that's all the help I can give on this topic. SemanticMantis (talk) 20:57, 11 April 2016 (UTC)

Windows 10 battery icon for a desktop computer

I have a new desktop computer with Win10 Home pre-installed. In its notification area is the battery icon usually present for a laptop. Mouseover shows "Fully charged (100%)". I first thought perhaps they had started putting battery packs in desktops, to reduce the need for UPS's. So I shut it down, pulled the plug, and pressed the power switch; nothing happened; ergo no battery. Is this just a Win10 bug? Anybody know anything about this? ―Mandruss ☎ 23:51, 11 April 2016 (UTC)

I've never heard of a desktop PC with an internal battery, however EVEN if it did have an internal UPS battery, being unable to turn the computer on without a power lead connected is NOT a good enough test to rule out the possibility. Many UPS systems are JUST enough power to shut down a system gracefully, they are not designed to run or boot up a computer. The battery icon is a feature of power settings in windows, go to control panel and find the power settings, you might be able to disable the battery icon from there. There IS actually a small battery in most desktops that just keeps the clock set while the computer us unplugged, but I doubt that's what the tray icon in windows is referring to. Vespine (talk) 03:56, 12 April 2016 (UTC)

→"pre-installed" I guess we can't know if the manufacturer used a laptop or tablet for preparing the hard drive image. Do You use an UPS or has anybody connected one to this computer? --Hans Haase (有问题吗) 10:08, 12 April 2016 (UTC)

Some wireless devices, especially mice and game controllers, have the ability to tell the host computer their battery level - so the computer can warn the user if the battery is getting depleted. If you click on the 100% indicator, in that event I'd expect it take you to the mouse/controller properties sheet, rather than the computer's general power property sheet (the one that talks about turning the monitor off after so many minutes of idleness). -- Finlay McWalter··–·Talk 10:38, 12 April 2016 (UTC)