Wikipedia:Reference desk/Computing

From Wikipedia, the free encyclopedia

The Wikipedia Reference Desk covering the topic of computing.

Welcome to the computing reference desk.
Want a faster answer?

Main page: Help searching Wikipedia

How can I get my question answered?

  • Provide a short header that gives the general topic of the question.
  • Type '~~~~' (that is, four tilde characters) at the end – this signs and dates your contribution so we know who wrote what and when.
  • Post your question to only one desk.
  • Don't post personal contact information – it will be removed. All answers will be provided here.
  • Specific questions, that are likely to produce reliable sources, will tend to get clearer answers.
  • 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 are not a substitute for actually doing any original research required, or as a free source of ideas.

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.
Choose a topic:
See also:
Help desk
Village pump
Help manual

March 12

Unique visitors to Wikipedia and popularity by region

Where can I find the stats on unique visitors to Wikimedia/Wikipedia etc.? Please don't direct me to Wikipedia:Statistics, I am familiar with many of those tools, but I couldn't track dawn that particular stat. Also, any chance anyone here has access to Alexa stats on Wikipedia? I asked here but that's a low traffic forum and anyway, that's a low-popularity database... --Piotr Konieczny aka Prokonsul Piotrus| reply here 04:14, 12 March 2018 (UTC)

Since we don't do browser tracking, we don't count unique visitors. Our best metric is 'unique devices'. —TheDJ (talkcontribs) 09:11, 14 March 2018 (UTC)

What is the difference between DDR3 and GDDR3 and how to differentiate between them?


I already know DDR3 is the memory for the CPU and GDDR3 is the specialized memory designed for the GPU. However, even graphics card manufacturers usually mix these 2 terms together and cause a lot of confusions. I have a few questions:

  1. Can DDR3 be used as VRAM? Or maybe all video cards with "DDR3" are actually GDDR3?
  2. I've seen quite a lot of cards with 128 bit DDR3 VRAM. Can DDR3 bus with be 128 bit wide? As far as I know the DDR3 bus should be 64 bit.
  3. What is the performance benefit of GDDR3 over DDR3? The formulas to calculate memory bandwith seem to be the same. (GDDR5 on the other hand has higher clock speed and 4x multiplier when calculating effective memory clock, instead of 2x.)
  4. Is there any reliable way to tell DDR3 from GDDR3? GPU-Z sometimes shows a DDR3 card as GDDR3. I am really confused.

Thank you for your time reading these questions. -- Livy (talk) 20:46, 12 March 2018 (UTC)

From our article GDDR3: It has much the same technological base as DDR2. So, there is little in common between DDR3 and GDDR3. Ruslik_Zero 21:03, 12 March 2018 (UTC)
DDR3 can be used as VRAM (or maybe more accurately RAM connected to a discrete GPU). DDR3 is much more likely than GDDR3 nowadays (and probably for the last 6 years or so) as GDDR3 is a specialised product and provides limited advantage over DDR3. (As said above, GDDR3 is based on DDR2.) If you wanted faster memory, you'd use GDDR5 or 5x (which are based on DDR3). Frankly, I wonder if anyone is still even producing GDDR3 and in any case I strongly suspect the GPUs don't support GDDR3. There is no reason why you can't have 128 bit bus for DDR3. After all nearly all CPUs since the beginning of the x86 IMC (I think possibly only the Socket 754 CPUs had single channel, unless I'm missing some of the Atoms or other specialised products) have had dual channel or better so a minimum of either a 128 bit buses or 2 independent 64 bit buses, and this even predated the IMC or even DDR2. See also [1] and List of interface bit rates#Graphics processing units' RAM. As for GPU-Z, well what is this card or at least how old is the specific GPU and are you sure it's using DDR3? More like though GPU-Z is just wrong. While it's a decent product for certain things, as with many products like that it's far from perfect. And frankly given the amount of nonsense out there about DDR3, GDDR3, GDDR5 etc there may not be much demand to fix this particular aspect. Nil Einne (talk) 11:53, 13 March 2018 (UTC)
I have a lot of cards around and switch between them quite often. It is good to know their differences, but it is not really important anymore. Thank you for your answer. -- Livy (talk) 14:18, 15 March 2018 (UTC)

March 13

Wall mounting a TV

I'm helping someone move and there's a 32" flatscreen TV that we want to mount at a specific location on a wall. It's normal drywall and because of proximity to doorways and stuff, the TV must be placed in a fairly precise spot, within 1cm or so. I think the TV has the usual cutout for a VESA mount, but the usual mounts if I'm not mistaken usually want to be screwed into a wall stud, so the location of the stud controls where the TV can hang.

Is there such a thing as a mount that allows adjusting the horizontal placement of the TV? I see a lot on newegg with tilt, swivel, extension arms, etc. All I'm after is a stationary mount where the TV just stays in one exact spot, no adjustments after installation and tightening things down, that hopefully doesn't stick out too far from the wall like those extension arm things would. Thanks! (talk) 04:27, 13 March 2018 (UTC)

Perhaps you can just screw a couple of horizontal pieces of timber into the wall studs, and then attach the normal mounting frame to that at the position you want. Graeme Bartlett (talk) 04:59, 13 March 2018 (UTC)
D'oh! That's a great idea, thanks very much! It's probably better than any special purpose gadget so I'm likely to use that approach. Thanks again! (talk) 05:26, 13 March 2018 (UTC)
A good VESA wallmount adapter can be adjusted. --Hans Haase (有问题吗) 11:19, 13 March 2018 (UTC)
Any good extension arm wall mount should barely stick out of the wall when you don't want it to. Mine is only about 1 cm. That said, if you're sure you have no use for the feature, I agree it's better to get one without. Nil Einne (talk) 11:55, 13 March 2018 (UTC)
  • Typical large TV mounts consist of a couple of horizontal rails (which may be DIN rail or similar) and a mount which attaches easily to those. The long rails can span wherever the studs are. The mount can be moved along the rails.
There are also fixing like the Grip-it [2] which provide strong connections to plasterboard or drywall. However plasterboard itself has its limits. In particular, it's surprisingly strong for a simple pure shear load, but the safe reduces reduces to about a quarter of this if it's at all offset from the plane of the board. Andy Dingley (talk) 12:04, 13 March 2018 (UTC)

Gaming Console Cooling System Working All The The Time

I've installed a game on my console since three weeks, and every time I play that game I hear the sound of the cooling system all the time during the game, while in other games played by a blue ray disk, I don't hear the sound too much, does that sound makes an indication of a horrible end for my device? (talk) 05:41, 13 March 2018 (UTC)

If it's really the cooling fans, (as opposed to the hard-drive), I wouldn't worry about it too much. It could just indicate that the new game is more graphically or CPU intensive. Some compressed air to clean out the dust probably wouldn't hurt, though.
If the noise is coming from the harddrive, that could be serious. Obviously, physical hard-drives make noise when they run, but usually it's pretty quiet. If it's making noticeably more noise than normal, you might consider backing up your data, just in case. Here are PS4 instructions, and Here are XBOne instructions. ApLundell (talk) 23:01, 13 March 2018 (UTC)

Thanks alot. (talk) 18:31, 17 March 2018 (UTC)

ignoring changes to the .gitignore file

I'm working on a public git repo where there is a .gitignore file, per standard practices. On my local repo I have some personal customization files, let's say at "mueller/", and those files should never ever be checked in. So I added a new line, mueller/, to the .gitignore file so that that mueller/ never shows up in git. That's all working well and good.

But the problem is now git is reporting that the .gitignore file has been changed (with the extra line), and is bugging me about it constantly. Is there a way to ignore changes to the .gitignore file?

I know I can gitignore the .gitignore file itself, but that involves untracking .gitignore, which affects other people unfortunately. Mũeller (talk) 07:06, 13 March 2018 (UTC)

  • As far as I understand the question, this StackOverflow question (and the accepted answer) solve it. TigraanClick here to contact me 16:31, 13 March 2018 (UTC)
Thank you very much. That solved it perfectly. Mũeller (talk) 07:25, 14 March 2018 (UTC)

March 14

YouTube video id

I've noticed YouTube video ids are always 11 characters long and contain only letters (upper and lower case), digits and the characters - and _.

But is it really an absolute rule? How can one be certain there isn't and there won't ever be a YouTube video id containing the characters +, = or , or . or #?

Thanks. Basemetal 07:37, 14 March 2018 (UTC)

It is a practice, not a rule. They are encoding a 64 bit integer in Base64. 11 characters gives you 66 bits. So the end result is 11 characters. But, what if they opt for a different ID in the future? The API for YouTube says that the 11 character rule should not be hard coded in any application because it may change at any time. (talk) 09:46, 14 March 2018 (UTC)
The ID is currently 8 bytes, which is 64 bits. They use Base64 encoding with two minor changes. They don't pad the output. They changed the standard + and / characters to - and _. You can easily reverse the Base64 encoding to get the original 8-byte ID, with two extra 0's added to the end. (talk) 12:08, 14 March 2018 (UTC)

Another YouTube video id question

Sometimes YouTube changes the last character of the video id I supply. For example when I try the video ids ___________, -----------, xxxxxxxxxxx, nnnnnnnnnnn, YouTube changes them to __________8, ----------8, xxxxxxxxxxw, nnnnnnnnnnk respectively. Could the last character be an error correcting character? If yes, does anyone know how to calculate the last character from the first ten to get a valid YouTube video id? Thanks. Basemetal 07:52, 14 March 2018 (UTC)

The actual ID is 64 bits. The 11 digit Base 64 encoding gives 66 bits. The extra 2 bits at the end are set to zero, which limits what the last character may be. (talk) 09:47, 14 March 2018 (UTC)

March 16

Using special symbols in Microsoft Word

When I want to use special symbols in Microsoft Word, I hit "Insert" and then scroll through a lot of symbols until I find the one that I want. I had asked a somewhat related question on the Math Reference Desk. (Here: Wikipedia:Reference desk/Mathematics#Symbol for "Real Numbers" in Microsoft Word.) Many of the replies refer to a "Unicode". What does that Unicode have to do with Microsoft Word? In other words, if I do know the specific Unicode for a specific symbol, how exactly do I use that information in Microsoft Word to actually insert the symbol? Thanks. (talk) 18:33, 16 March 2018 (UTC)

Unicode is the encoding used to represent characters internally in almost all modern software. In Microsoft Word, you can enter a specific Unicode character by typing the hexadecimal code (for example, "211d" for the double-struck R), then press Alt-X. CodeTalker (talk) 19:07, 16 March 2018 (UTC)
Thanks! Never knew that! (talk) 04:51, 17 March 2018 (UTC)
Is that trick also available in the Mac version? The keystroke would differ. —Tamfang (talk) 23:51, 17 March 2018 (UTC)
  • Note that Unicode is not an encoding. It is a single, very large, character set intended to be useful for all characters needed, for all situations, and simultaneously. It is a list of code points (the numbers we discuss) which map to the glyphs needed.
To store Unicode characters in a file, it is necessary to encode them into an encoding such as UTF-8 or UTF-16. Other encodings are possible and there are detail variations within this (such as whether a "BOM" is used), but these are the main two. For Windows it is usual to use UTF-16 within programs, but UTF-8 when stored in files. These are mostly invisible to the user, certainly if you're using Word, and most programs hide these details well. However when exchanging files, it's not enough to specify "Unicode", but also that the encodings are the same, or at least compatible.
If in doubt, use Unicode with a UTF-8 encoding using a BOM. Andy Dingley (talk) 00:07, 19 March 2018 (UTC)

March 17

Android Question Again

I asked this question about two weeks ago, but will ask again because I still haven't solved the problem. I have a Samsung Galaxy J7 running Android version 7.0. I am still having the problem when driving, that, under circumstances not under my control, the phone pings and tells me to log on to Cox WiFi. Cox is a major WiFi service provider in the area, and I assume that the message means that I have driven into the service area of a home or commercial wireless router that is providing Cox WiFi service. I don't have a Cox account, and I don't want to get these pings. How do I suppress them? Please don't give me accurate but useless advice such as turning the phone off (when I am using its GPS) (since I know that some editors here will give technically accurate but useless advice such as to turn off the phone). Robert McClenon (talk) 22:39, 17 March 2018 (UTC)

Maybe try going to the screen where you select a WiFi network. Press the ... button then pick Advanced, and there deselect "Hotspot sign-in". I think either that should work or go to the nut icon in the same WiFi selection screen and deselect "Network notification". I'm not sure what's happening from your description, looks like either you're getting a notification that there's a free Cox Wifi hotspot around, or your phone has already connected and wants to sign in on the hotspot page. Another possibility is that you logged on a Cox WiFi hotspot somewhere before and your phone kept its settings in memory, and now you've come accross a hotspot with the same name, password etc. I don't know how often that happens. Another solution is to turn off WiFi connection when driving. That's what I usually do, since I have no idea what use could WiFi possibly have in a moving car, but you might have some gadget I don't (e.g. I don't have a GPS). (talk) 02:56, 18 March 2018 (UTC)
I tried the Advanced settings, but it doesn't have a Hotspot sign-in thing. It has a Hotspot 2.0, but that is already turned off. As to name and password, the prompt that I get (that I don't want) asks me for my name and password. It doesn't appear to have them saved. (I have my name and password saved for four bands of my home wireless router, but that is unrelated.) I don't see a nut icon, not down at this level. (I get into WiFi settings from the overall Settings icon on the main screen, and that is a nut icon, if you mean nuts-and-bolts, not seeds with shells, but I assume you mean down inside the Settings.) What nut icon? (I hope that it isn't a peanut, because peanut allergy can be a medical emergency.) Robert McClenon (talk) 03:11, 18 March 2018 (UTC)
I meant the nut-without-a-bolt icon in the main Wifi selection menu, next to the ... icon for the Advanced settings :) However, what you write is a little strange. You're talking about the usual prompt that you get when you select an unknown network from the Wifi menu, or the internet browser thingy when you have to log on to a hotspot? My phone only asks me for a password when connecting to a home WiFi. (talk) 03:39, 18 March 2018 (UTC)
At the top, I see WI-FI in black (a heading), WI-FI DIRECT (blue, clickable), and a ... thing. I agree that the phone only asks for a password when connecting to a home WiFi, but what is happening appears to be my phone asking for a password to connect to someone's home WiFi, someone whose house I am driving by. Robert McClenon (talk) 14:57, 18 March 2018 (UTC)
Last time you could not find your list of known wifi networks. It sounds as if this is one of those. You will have to delete (forget) it from the list. Graeme Bartlett (talk) 06:03, 18 March 2018 (UTC)
User:Graeme Bartlett - I don't understand. You tell me to delete or forget "it" from the list. Delete or forget what? I don't see a network that could possibly be Cox. You are evidently telling me to bring up a list of saved networks that will include it, but where do I bring up the list to delete or forget Cox? Robert McClenon (talk) 14:57, 18 March 2018 (UTC)
In Android setting select WiFi. Then touch the ... symbol to get extra actions, select "saved networks". Then find your cox network entry, touch that and then touch "forget". (that is on nougat). Graeme Bartlett (talk) 21:57, 18 March 2018 (UTC)
Doesn't that forget it only for the current session? If wifi is turned off and back on again later, the whole thing starts again. Or if the phone is turned off, it starts again when next turned on, no? To me, the problem is that he's driving around with wifi turned on, so of course the phone is going to prompt him when a free network comes in range. Why does he have wifi turned on? Wouldn't it be better to turn off wifi and use mobile data instead for his GPS? Assuming he's on a data plan of course. Akld guy (talk) 19:29, 19 March 2018 (UTC)

March 18

Connecting an Amiga to a modern monitor

I just found my old Amiga 500 and 1200 computers in my father's basement. I'd like to see if they still work, but I no longer have an Amiga monitor. Can they be somehow connected to a modern monitor with a VGA and DVI input? JIP | Talk 22:29, 18 March 2018 (UTC)

The distinction between "monitor" and "TV" is blurred these days, but for old computers it's usually easier these days to connect them to something TV-like, rather than something like a modern monitor. Most old TVs like this can be connected fairly easily to a SCART connector, and even SCART can then be connected to HDMI with a fairly common adapter. SCART is RGB (each colour gets its own wire) which gives a notably better image than composite video. Note that PAL / NTSC conversion might be needed if you're using US computers in Europe, but TVs are fairly flexible about that nowadays. Here's some YouTube on doing this for Amigas, but most computers of that vintage will be similar. Video Out on YouTube Andy Dingley (talk) 00:13, 19 March 2018 (UTC)
Our articles on the Amiga 500 and the Amiga 1200 say they came with an RF adaptor port that was PAL and NTSC capable. You should be able to find a suitable cable to allow connection to the antenna input of a TV. I don't know whether current TVs have such capability, but old pre-digital TVs shouldn't be hard to find. The Atari 520ST and up models had the same RF adaptor and I successfully used my 520 with a standard TV (audio as well). All that was necessary was to tune the TV to channel 36. Akld guy (talk) 19:54, 19 March 2018 (UTC)

March 19

Amdahl's law variant, for when sequential portion excludes only itself?

Does a variant of Amdahl's law exist for the case where the parallel-before-sequential or sequential-before-parallel dependency only applies to each work unit, so that one instance of the sequential portion can run concurrently with any number of instances of the parallel portion (and, unlike with Gustafson's law, the amount of sequential work is asymptotically proportional to the total amount of work)? An example would be a Java program like this:

public static void example(int i) {
  final int sequentialOutput;
  synchronized {
    sequentialOutput = sequentialPortion(i);
public static void main(String[] args) {
  for (int i=0; i < SCALING_PARAMETER; i++) {
    final int thisI = i; // necessary in Java
    ForkJoinPool.commonPool().submit(() -> example(thisI));
  ForkJoinPool.commonPool().awaitQuiescence(); // wait until all tasks are done

I believe that in the above case, the maximum throughput will be achieved (neglecting overheads and memory constraints) if the number of threads in ForkJoinPool.commonPool() is at least the total running time of example divided by the running time of sequentialPortion, and each thread has a processor core. But under Amdahl's law, maximum throughput could only be approached asymptotically. NeonMerlin 04:09, 19 March 2018 (UTC)

How does space complexity affect empirical time complexity?

As a program's memory usage increases, so will its memory access times, as it begins to run out of registers, then to miss the L1 cache, then miss the L2 cache, then miss the L3 cache, then hard-fault to disk, then hard-fault to a "disk" that's an abstraction of tertiary storage. As well, AFAICT, RAM stick specifications involve a tradeoff between access speed and capacity, even on an unlimited budget. Are any models available of how an algorithm's space complexity affects the difference between its theoretical and empirical big-O time complexity? NeonMerlin 04:47, 19 March 2018 (UTC)

In modern usage, all programs are swapped all the time. Therefore, the rule is that you decrease time in an algorithm by using more space. You don't increase time by using more space. (talk) 12:53, 19 March 2018 (UTC)

Freenode banning me and the possibility I am a zombie

IRC banned me a few hours ago. A report dated 2013 or something said something like code 17 and that their automated open proxy scanner was able to exploit software running on my machine. I just got back onto IRC with no problem. Odd. Am I a zombie? Is this some sort of false alarm? Anyone? Anna Frodesiak (talk) 08:13, 19 March 2018 (UTC)

This is fairly common if you're on a dynamic IP range, in certain locations. Freenode uses a blacklist to ban IPs, so it probably doesn't refer to what's currently happening on your computer. It just means you've been assigned an IP address that was once assigned to a computer that was compromised in the past. Just wait to be reassigned a new IP and you'll be good. -- zzuuzz (talk) 08:36, 19 March 2018 (UTC)
Phew. Thanks, zzuuzz. I just paid my annual net bill. Maybe they gave me a new IP. Is there a way to tell what my old one was and if it was static and the same about my new one? Anna Frodesiak (talk) 08:40, 19 March 2018 (UTC)
Probably the easiest thing to do is visit and note your address over a period of time. I'll tentatively wager that it changes on its own within a day. If you restart your router it may change again. -- zzuuzz (talk) 08:48, 19 March 2018 (UTC)
Wonderful. I will check and keep a record. I just have a computer and am not sure what a router is. I don't think I have one. Search engine pics show a box with rabbit ears and wires. They look fancy and important. Maybe I should get one. Then again, these things always give me trouble. I'd probably end up putting balls of tinfoil on the rabbit ears to make it work properly, and then it still wouldn't. Thank you kindly, zzuuzz. Anna Frodesiak (talk) 08:59, 19 March 2018 (UTC)

What kind of card is this?

What kind of card?

I bought this card on ebay and it was stated to be an IBM Solid Logic Technology card. The card is very similar to the card pictured in that article, but the components are different. Reading the article section "Later developments", it sounded like a Solid Logic Dense card. But it has no resistors (or any components) on the other side. So what kind of card is it? Bubba73 You talkin' to me? 14:30, 19 March 2018 (UTC)

It looks like a transistor array used as a digital buffer. The top transistor is the controller or clock. On a complete circuit, you'd expect a ground with a capacitor and regulator power, but this is a card. So, that could be on the main board and removed from the card. To get a real idea of what is going on, you need to look at the opposite side to see how the transistors are connected. (talk) 19:49, 19 March 2018 (UTC)
The other side of the card has solder spots like the other side. You can see the three leads from each transistor soldered there, but there are no connections between the transistors. Perhaps this is memory or registers? Bubba73 You talkin' to me? 20:03, 19 March 2018 (UTC)

March 20

Retrieved from ""
This content was retrieved from Wikipedia :
This page is based on the copyrighted Wikipedia article "Wikipedia:Reference desk/Computing"; it is used under the Creative Commons Attribution-ShareAlike 3.0 Unported License (CC-BY-SA). You may redistribute it, verbatim or modified, providing that you comply with the terms of the CC-BY-SA