Computing desk
< November 18	<< Oct \| November \| Dec >>	November 20 >

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.

November 19

Why do uploads slow down downloads

When I upload a video on to YouTube, I can't surf the web anymore, downloads slow to a crawl. Why is this? ScienceApe (talk) 00:19, 19 November 2012 (UTC)[reply]

Whenever you're uploading or downloading, there is always two-way communication between your computer and the server. When downloading your computer will send a verification message every few packets or so, so the other computer knows you got them. If there's a problem with some of the data, your computer will send a request to resend the corrupted packets. This upstream communication is almost impossible when the Youtube uploader uses 100% of the upload bandwidth.

ADSL differs from the less common symmetric digital subscriber line (SDSL) in that bandwidth (and bit rate) is greater toward the customer premises (known as downstream) than the reverse (known as upstream). This is why it is called asymmetric. Providers usually market ADSL as a service for consumers to provide Internet access in a relatively passive mode: able to use the higher speed direction for the download from the Internet but not needing to run servers that would require high speed in the other direction.

The solution is to restrict the uploadspeed for the YouTube uploader. I use NetLimiter. Trio The Punch (talk) 00:34, 19 November 2012 (UTC) p.s. If you open the Windows Task Manager and go to the Performance tab you can see if your CPU is fast enough to keep up. If the CPU usage is constantly at 100% while uploading then that is the bottleneck, but that should only happen on really old or busy computers.[reply]

In principle, a full-duplex communication system can transmit uploads at full speed with absolutely no effect on download speed. In practice, both your hardware and your software are not-quite-full-duplex. Ethernet switches and routers, and other telecommunication gear between your computer and the server, all have hardware limitations. Software protocols at many different levels of the OSI model provide limited support for full-duplex communication. As a general rule, any protocol that exercises a "handshake" - like TCP - is designed such that any data flow in either direction has a corresponding data acknowledgment flowing the other direction. In this particular instance, we don't have enough information to concretely identify the source of the slowdown. Some possibilities include hardware limitations of the networking gear; software limitations of the application software; Quality of service (QoS) implemented by the service provider(s); and as always, potential observer bias. Nimur (talk) 12:26, 19 November 2012 (UTC)[reply]

I don't think there's any observer bias... I just recently switched to a new ISP, Comcast. We had DSL before and we got 1.5Mbps down about 400Kbps up. Back then when I uploaded videos to YouTube, it slowed down my browser sure, but I could at least surf the net still. Now we have 20-25Mbps down and about 4Mbps up but whenever I upload videos I can't browse the web at all. Whenever I try it just says the site is not responding. ScienceApe (talk) 15:41, 19 November 2012 (UTC)[reply]

Try slowing the youtube uploader down a bit (with NetLimiter or something similar). Does that fix the problem? If it does not the CPU may be the bottleneck (on old or busy computers). Check if the CPU usage is 100% while uploading. Trio The Punch (talk) 16:40, 19 November 2012 (UTC)[reply]

raspberry pi stuff

I still need to get a mains wire to plug my raspberry pi in, but I don't know what sort I want. As well, is it possible to get a battery for it, so I can use it out and about like a laptop?

82.132.244.24 (talk) 14:53, 19 November 2012 (UTC)[reply]

The Raspberry Pi is powered by a micro USB connector (like a lot of phones), but the mains->microUSB power adapter has to supply at least 700mA, which is a lot more than the adapters for mobile phones supply (e.g. the adapter for my phone only supplies 200mA). The same suppliers that sell Raspberry Pis retail will also sell you an appropriate adapter. -- Finlay McWalterჷTalk 15:02, 19 November 2012 (UTC)[reply]

You will find a lot of useful information at RPi_VerifiedPeripherals. As for the power supply, assuming you bought from Farnell (or the USA equivalent site), there are power supplies in the accessories section. Alternatively see the link I gave earlier for compatible supplies, which must be ideally rated at least 1000mA (1 Amp) at 5 volts. As Finlay has pointed out you also need a micro-usb cable. It is also possible to back-power from a powered USB hub, see the aforementioned link. As for using it out-and-about, it draws about 500mA with no peripherals, so you'll need to size your battery pack accordingly. I believe it will run off 6 volts, ie 4 * 1.5v or 4 * 1.25V NiMH rechargables. Be very careful with the polarity and pin selection if you power via the GPIO header as there is no protection and applying 5 volts to any pin other than the supply pin will destroy the board. Kram (talk) 15:33, 19 November 2012 (UTC)[reply]

I'm going to roll back a little more on this. The consensus on the Pi forums seems to be 5V ±0.25V (ie the old TTL spec), I had rather naively expected a slightly wider tolerance more in keeping with HCMOS, but things have moved on somewhat since my earlier electronics hacking days, so I defer to more recent voices. Sigh, perhaps I'll go back to lurking.Kram (talk) 19:31, 19 November 2012 (UTC)[reply]

Please don't do that: your answers have been informed and informative. 87.112.60.60 (talk) 20:34, 19 November 2012 (UTC)[reply]

There is more on power supply at RPi_Hardware. I take back what I suggested earlier about using the GPIO pins. Unless you are an electronics expert, just buy an approved external battery pack which will connect via the micro-usb power connector. Kram (talk) 15:42, 19 November 2012 (UTC)[reply]

a .pdf for an intel board

Some one knows where I can find the Intel server board S4600LT2 Technical Product Specification??

Iskánder Vigoa Pérez (talk) 19:13, 19 November 2012 (UTC)[reply]

Is this what you are looking for, or do you need more detailed specifications? Yetta Bronstein (talk) 20:46, 19 November 2012 (UTC)[reply]

The Technical Product Specification doesn't seem to be linked on the Ark yet, so I suspect there isn't a version published for customers yet. I recommend contacting your Intel distributor for more details on the board. The launch date is listed as Q4, so I wouldn't be surprised if they could get you a decent amount of documentation and possibly a few sample boards. 209.131.76.183 (talk) 20:55, 19 November 2012 (UTC)[reply]

yeap... no way I can find the .pdf

thanks for answer

Iskánder Vigoa Pérez (talk) 19:11, 20 November 2012 (UTC)[reply]

HELP... C code keeps asking for an identifier

 int i=0;
 int count=0;
 
 // we don't know how long the string is
 // but we do know it ends with '\0'
 while (s[i]!='\0')
   {
     // count the e's
     if (s[i]=='e')
       count++;

     i++; // if you don't have this line, infinite loop!
   }

 return count;

I keep getting the errors "[i]expected identifier or '(' before 'while'[/i]" and "[i]expected identifier or '(' before 'return'[/i]". What's the typo? Im pretty sure I don't need a parenthesis there. 128.111.43.37 (talk) 21:47, 19 November 2012 (UTC)[reply]

Have you declared s? Rojomoke (talk) 22:10, 19 November 2012 (UTC)[reply]

(ec) Is this your entire program? If so then you are missing a main() function, vis

char s[] = "Hello World";  // Thanks Rojomoke
int main()
{
 int i=0;
 int count=0;
  
  // we don't know how long the string is
  // but we do know it ends with '\0'
  while (s[i])
    {
      // count the e's
      if (s[i]=='e')
        count++;

      i++; // if you don't have this line, infinite loop!
    }

  return count;
}

87.112.122.5 (talk) 22:15, 19 November 2012 (UTC)[reply]

I'm doing this lab (I'm not in the class, doing it on my own) http://www.cs.ucsb.edu/~pconrad/cs16/10S/labs/lab05/

(ec)Well, we don't do homework here, but as a hint, check the parentheses of your enclosing function (I got exactly the same errors as you did by omitting the main() function). Kram (talk) 22:33, 19 November 2012 (UTC)[reply]

What I meant by I'm not in the class, I'm literally not taking Computer Science 16. I'm just going through the labs in order. So I'm not 'cheating' on my homework. Sorry I didn't make that clear 128.111.43.37 (talk) 22:39, 19 November 2012 (UTC)[reply]

So I don't want s to be a defined string, I want it to be wahtever string is being tested.128.111.43.37 (talk) 22:25, 19 November 2012 (UTC)[reply]

You only copied a program fragment. The variable s has to be declared somewhere, even if the value of s is not specified there. Note what your link says above the code you copied: "Next, replace the stub with this code". You didn't replace anything but just copied the code by itself. It's a fragment and doesn't work by itself. Earlier on the page is a link to http://www.cs.ucsb.edu/~pconrad/cs16/10S/labs/lab05/code/ where you can see the original function http://www.cs.ucsb.edu/~pconrad/cs16/10S/labs/lab05/code/howManyEs.c. PrimeHunter (talk) 22:42, 19 November 2012 (UTC)[reply]

So I need to take the code I quoted at the very top of this question, and paste it in for return -42; and that's it? — Preceding unsigned comment added by 128.111.43.37 (talk) 23:16, 19 November 2012 (UTC)[reply]

That's it for that file. The exercise relies on other files. Try to follow the page from the start. It says "To compile, use this command":

cc howManyEs.c tdd.c testHowManyEsDriver.c -o testHowManyEsDriver

You need all three c files, and they include tdd.h so you also need that. That's it. The files are in http://www.cs.ucsb.edu/~pconrad/cs16/10S/labs/lab05/code/. PrimeHunter (talk) 23:58, 19 November 2012 (UTC)[reply]

What happened to Framing (World Wide Web)

At the end of the 90's, Framing (World Wide Web) was extremely popular, not just for programmers but also designers seemed to like it. The menu, the logo, etc would remain visible all the time. A typical remark as I can remember is something like "This is clearly a major invention by Netscape and such a shame IE followed much later. Users love it, let's just make a page 'you have an outdated browser so you cannot access this site. Click here to download Netscape'." The article mentions "In late 2011 frame-like user interface features began appearing on various well maintained websites." The vast majority of professional sites however still don't and would probably be laughed at as if they were using an animated GIF. Does anyone know what event led to everyone suddenly stopping using frames? Joepnl (talk) 23:02, 19 November 2012 (UTC)[reply]

You seem to have missed a whole chunk of the history of the WWW. Perhaps frames were the cool new feature and were popular among end users for a short time after introduction but they quickly developed a poor reputation among end users for various reasons that our article hints at. See also [1] [2] Nil Einne (talk) 02:00, 20 November 2012 (UTC)[reply]

Framing was replaced by AJAX and CSS. 190.60.93.218 (talk) 16:18, 20 November 2012 (UTC)[reply]