Wikipedia:Reference desk/Archives/Computing/2012 December 22

= December 22 =

Is the data rate of a connection the same as that of the slowest device in the path?
Consider two (simplified) networking systems. In one, a high-speed device sends data over fiberoptic cabling to a device with a low upper limit of bitrate (eg, 56K). The second system is the same, but the cabling is telephone line rather than fiberoptic. Would the download speed of the recieving device be any different between the two systems, given that fiberoptic is "faster" than phone lines?

Is it like a car changing speeds at different parts of its journey (the total average speed of the car is not the same as its slowest speed)? Or is the total average speed generally the same as that of the slowest device or technology (computer, NIC, UTP, hub, switch, modem, phone line, server, etc) that the data travels over? ± Lenoxus (" *** ") 01:18, 22 December 2012 (UTC)


 * Think of the bits of data as being like cars then and your telephone line being like a road through a small town. Just because a multi lane highway brings cars to the outskirts of town does not mean that any more cars are going to be able to go down the small road in town every hour. The time for all the data to download is normally far longer than it takes for the first item of data to arrive after being sent, even a small web page consists of tens to hundreds of thousands of bits, so the rate going through the slowest part is what is important. Dmcq (talk) 01:49, 22 December 2012 (UTC)


 * That's a good analogy, and it carries through to the case where the only roads are narrow over the whole distance. There are likely to be delays further back along the route, so even the last (56K) narrow road might be unused at times if it is waiting for traffic that was delayed further back.  The analogy doesn't quite carry through to average speed, because it's congestion that causes the main delays further back, and it's the number of cars that get through per unit of time that matters.  Your 56K is the main limiting factor, but congested telephone lines further back might mean that download is even slower than the 56K narrowest part.  Good buffering (car parks, but with engines running ready to go again) at the right places can mitigate this effect.    D b f i r s   09:01, 22 December 2012 (UTC)


 * Buffering doesn't help, it actually makes things worse by hindering TCP's ability to adapt to the available bandwidth. See bufferbloat. I think the OP is simply asking whether fiber-optic cable is really faster than a phone line, and the answer is not much if at all. Optical fiber says that a typical index of refraction is 1.6, meaning the signal travels at about 5/8 of the speed of light, while speed of electricity says that it can be as much as 97% of the speed of light in unshielded copper, though I suspect it's slower than that in phone lines. Optical fiber also needs fewer repeaters, which add delay. However optical fiber is "faster" in that you can pack more bits longitudinally in it, meaning you can transfer more data in a given time, though each bit takes just as long to arrive. -- BenRG (talk) 00:07, 23 December 2012 (UTC)


 * Thanks for clarifying. I'd assumed the OP was asking about download speed.  Also, I wasn't aware of the buffers' effects on TCP.  I suppose one could design a protocol where buffers helped, but if that's not what the internet uses, then it's pointless.    D b f i r s   08:02, 23 December 2012 (UTC)

Egyptian Arabic Wikipedia

 * 1) Why accept the establishment of Egyptian Arabic Wikipedia (Meta) with it is known that the Egyptian Arabic dialect?
 * 2) Can this version is canceled? --41.130.174.90 (talk) 10:31, 22 December 2012 (UTC)
 * Our article about that version of Wikipedia is at Egyptian Arabic Wikipedia. It has nothing to do with this English version and it has nothing to do with this reference desk which is about computing questions. Why not ask your questions at that Wikipedia instead? In general all that is needed for starting a new Wikipedia is a demonstration that there would be sufficient support, that obviously was shown. Basically the attitude I believe you'll find anywhere on Wikipedia to your idea of stopping something that has support is that if you don't like it that's your problem and you can go and lump it. Dmcq (talk) 12:21, 22 December 2012 (UTC)

My laptop speakers have become crackly & tinny - how to restore the sound myself?
Hello, the left speaker doesn't put out at all, and the right one, as of 15 minutes ago, started putting out a severely ruined sound that is more tinny and crackly than true sound. I now can't make out the words 3/4 of the time.

Instead of buying external speakers, how do I get in there and actually repair the workings of the speakers so that they put out correctly again?

The computer is a Dell Inspiron 1720 and went out of warranty in September. I hope to get a killer replacement next month (preplanned for quite some time) but would like to know how to fix the speakers in order to tide me over until then. Thanks kindly. --70.179.167.78 (talk) 17:29, 22 December 2012 (UTC)


 * I'd suggest just using headphones until then (or external speakers, when home). Laptops are quite difficult to repair, and you're likely to damage something else in the process.  Plus, you'd likely need to buy new speakers to install into the old laptop, and, unlike external speakers or headphones, those can't be reused elsewhere, so it's just throwing money away on something you will only use for a month (less than that, including delivery and installation time).  There's also some possibility you just have some loose wires to deal with, but, unless you're an expert at soldering, that's not easy to repair, either. StuRat (talk) 17:44, 22 December 2012 (UTC)

Video calls
How do video calls work on phones. Do they work over 3g data? Im confused as many mobile networks don't class video calls as mobile data. Clover345 (talk) 17:35, 22 December 2012 (UTC)


 * It's probably less about the specific route data takes and more about whether one can reasonably say that a wireless provider for a phone with a video camera is not assumed to provide video phone calls. Some providers are picking nits about this however (and some aren't). ¦ Reisio (talk) 06:42, 23 December 2012 (UTC)


 * Bear in mind MMS also uses data (3G if you're using 3G) but few carriers charge for the data directly (they charge per MMS, in many countries only for sending). As I understand it, push to talk also uses data, but I believe any mobile provider offering it doesn't charge for data. Ultimately a provider can choose to set-up their data charging however they want, those offering Facebook Zero don't charge for using it, and there are or were plans to offer a similar service for wikipedia. However according to this link,, real video calls over 3G don't use what most people would call 'data' they use a circuit switched connection similar to real voice valls. 3G-324M seems to confirm this. If you're using something like Skype or Facetime, then this would be using data, I'm not aware any provider offers these free but perhaps they do in the US since evidentally they don't offer real video calls there. Nil Einne (talk) 07:48, 23 December 2012 (UTC)


 * There's little stopping people in the USA from making video calls, except that they aren't interested in doing so. ¦ Reisio (talk) 17:31, 24 December 2012 (UTC)

Ping hosting provider
I am deciding on a hosting provider and am considering companies that are located in another country (than most expected visitors to the website). I would like to compare ping and download time for various providers. Is there a tool to do that, and more importantly, which server/ip should I ping? bamse (talk) 17:42, 22 December 2012 (UTC)
 * Ping time is measuring the delays in the route to the hosting server. Download speed also depends on this route, and also on congestion at the hosting server and their link to the internet. I don't know of a tool, but if you use these tests, make sure that the slowest link is not your own internet connection.  I would also want to look at data on server downtime.    D b f i r s   20:54, 22 December 2012 (UTC)
 * Thanks for the reply. But which server should I ping? E.g. for OVH, should I ping just their website at www.ovh.de for instance, or rather something else, since my server might be at a different location than the server hosting their website? bamse (talk) 09:17, 23 December 2012 (UTC)
 * Yes, OVH have servers in several data centres in different countries, so you can't know in advance which server you will be allocated. I would have thought that it would, in general, be faster to use servers in the country of your customers, but there are so many other variables to consider, so neither location nor ping time need be a major factor in your decision.  As you probably know, you can't really be sure that your hosting provider will provide all of your requirements efficiently until you have tried them for a while.  I'm not an expert on this topic, so perhaps someone else can help?    D b f i r s   11:23, 24 December 2012 (UTC)

What other Chrome Apps are like Session Buddy?
Just an hour ago, in the last crash-restart, my Session Buddy lost all of its sessions that it has ever saved, and I have no idea how to recover them. I had 44 tabs open, and would prefer to find them anyhow.

But for the future, in case Session Buddy goes wrong again, I'd like to run another app like Session Buddy that will hopefully have all the previous sessions if Session Buddy no longer does, or vice-versa.

So what apps are like Session Buddy that can hopefully save my sessions if Session Buddy doesn't? Thanks. --70.179.167.78 (talk) 21:11, 22 December 2012 (UTC)

how do I prevent text from being part of a link? (esp. Google Docs)
I always have to do really funky things, like click on the next line, write some text and then backspace so the text can be adjacent to the link but not actually part of it. It's so annoying! When I use "clear formatting" on the part that I don't want to be part of the link, it doesn't work at all! How do people explicitly specify that a link should stop? 72.229.155.79 (talk) 22:33, 22 December 2012 (UTC)