Computing desk
< October 17	<< Sep \| October \| Nov >>	October 19 >

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.

October 18

How does a PC find a specific web page on a server in another location?

When I type www.amazon.com and press enter, do electronic signals search the entire web. This may sound naive, but does a pulse go to China, Germany and look for it? I know the words are a pointer to a number but I just don't understand the path the electonic car takes to deliver the correct page. Thank you. This has been on my mind for a while.

Your computer sends "www.amazon.com" to your nameserver. You don't know what your nameserver is. It is provided to you by whoever is providing your internet access. The nameserver send back the real name for "www.amazon.com", which is 72.21.206.5. Then, your computer sends the request to the real name, which is easy to find. It is on the "72" network. From there, it is on the "21" subnetwork. From there, it is on the "206" subnetwork. From there, it is computer "5". --Kainaw ^(talk) 02:02, 18 October 2006 (UTC)[reply]

Uh not exactly. The IP address is a unique identifier that's composed for four octets that happen to be separated by dots, they're not subdomains. See my answer to a similar question here. This is an extremely complex question and internet archetecture is a pretty mind-boggling topic. the ISOC has a lot of good articles about it if you're a member. --⁪froth^{T C} 05:37, 18 October 2006 (UTC)[reply]

It appears you think subnetwork is a typo for subdomain. They are two completely different words. --Kainaw ^(talk) 16:10, 18 October 2006 (UTC)[reply]

Thanks for thinking!

It happens “automagically”, using complex technology to produce miraculously simple results. The first part of the process involves changing a name to a number, which involves domain name resolution via the domain name system. In a common case like amazon.com, it is likely that the name can be resolved without external requests, because it has been seen before. The resulting IP address is like a street address for a postal service, an exact destination on the Web. Ever wonder how a physical package reaches its destination? Tracking sites for, say, UPS or FedEx show a series of intermediate steps along the way. The Internet uses packet-switching technology to accomplish much the same thing with electronic “packages”, sending the data along a path that is dynamically determined at each step by target and traffic. A large message will be broken into smaller packets, and each packet may take a different route; thus out-of-order reassembly is needed at the destination. At each step, local software uses large tables of connections, speeds, and reliabilities — all constantly updated — to choose where next to transfer the packet. If an accident or congestion takes out one path, another will be used.

Fortunately, all this resolving and routing can be done without any step requiring information about the entire Web. By keeping only local data about a small piece of the whole, each step can be quick and reliable.

In the famous words of Arthur C. Clarke, “Any sufficiently advanced technology is indistinguishable from magic.” --KSmrq^T 07:49, 18 October 2006 (UTC)[reply]

I think the article on router is a good starting point to understanding network traffic. Vespine 05:25, 19 October 2006 (UTC)[reply]

ipod

Can you put video found on DVD on an ipod?

Yes, there are several tools to do this. With Windows, I tend to use the Videora iPod Video Converter. -- Consumed Crustacean (talk) 01:22, 18 October 2006 (UTC)[reply]

thankyou

Can't find mpdat.exe

Every time I turn my PC on, it gives this message. Is mpdat.exe an important file? Should I be worried it's not there? And how can I get it to stop saying that if not? Cheers --212.100.250.212 01:58, 18 October 2006 (UTC)[reply]

Google says that mpdat.exe is bad stuff; more specifically, it's an IRC/botnet trojan. I'd run a full virus/adware scan, just incase. If you do have a virus scanner, it could be showing that message because it caught and removed the file. It could also be because the botnet operator decided to nuke the network, for whatever reason. -- Consumed Crustacean (talk) 02:02, 18 October 2006 (UTC)[reply]

It's done this for quite a while now and my scans haven't turned anything up. If the virus has gone then how could I get this message to stop coming up. It's the standard 'Windows can't find this file, you may want to search for it' box. I guess it's not going to be easy...? --212.100.250.212 02:09, 18 October 2006 (UTC)[reply]

It's showing up because Windows is still trying to run it. Try running msconfig.exe in the run box (Win+R). I'm still betting something is there though, or this wouldn't be suddenly showing (unless you had it all along and the botnet operator caused it to self-destruct). Is your AV up to date and everything? -- Consumed Crustacean (talk) 02:29, 18 October 2006 (UTC)[reply]

Yeah AV is up to date. Ok thanks for that command, I'm going to go through what is causing it. Thanks! --212.100.250.212 02:49, 18 October 2006 (UTC)[reply]

Might want to try another AV if it's not catching it try Avast!. I switched over to Avast! about a year ago and it picked up 9 trojans that Macafee said I didn't have. Whispering 20:36, 19 October 2006 (UTC)[reply]

Lanczos Resampling

Hi, I'm trying to put together a program that can resize images using the Lanczos kernel. I already have the code to convert the image to an array of RGB values. However, I'm not sure how to convolve this two-dimensional image with the Lanczos kernel. The article seems to indicate that there exists a two-dimensional formula for kernel. Does anyone know what it would be? Or I might be going about the problem in the wrong way. I would greatly appreciate any advice. Thanks. Austboss 03:57, 18 October 2006 (UTC)[reply]

A bit of guessing here. The one-dimensional thing was

L(x)={\begin{cases}{\frac {2\sin(\pi x)\sin \left({\frac {\pi }{2}}x\right)}{\pi ^{2}x^{2}}}&-2<x<2,\,x\neq 0\\1&x=0\\0&\mathrm {otherwise} \end{cases}}

.

Is it unreasonable to believe that we would get the two-dimensional version if we interpret

x

as the distance in the two-dimensional plane, rather than just on the one-dimension axis? —Bromskloss 09:48, 18 October 2006 (UTC)[reply]

Oh, and another thing. Is this method considered to be the best for resizing images? We don't seem to have an article comparing different algorithms. —Bromskloss 09:50, 18 October 2006 (UTC)[reply]

Thanks for the responses, Bromskloss. I've been looking into this for several hours now, and I think I've figured out what I was doing wrong. Just in case anyone cares, I'll explain it here. To resample and resize a two-dimensional (or more, actually) image using a filter like this, you need only apply the one-dimensional version multiple times. That is, first resize one dimension of the image, and then resize the resulting "intermediate" image in the other dimension. Also, we don't have an article comparing them, but there has been extensive work done comparing various sampling algorithms. The Lanczos kernel, while not always the best, is generally considered to be a good choice when both computation time and image quality are considerations. Most image processing software uses this filter as the default. I'll probably try to update the article with this information soon. Austboss 11:40, 18 October 2006 (UTC)[reply]

Ah, one dimension at a time. Convenient. Why is it that one talks about two- and three-lobed versions but not higher numbers? I mean, wouldn't it be better to have as wide a window as possible? And which algorithm should one choose if computation time isn't a problem? —Bromskloss 12:34, 18 October 2006 (UTC)[reply]

The proposal to process one dimension at a time means working with a separable filter. That is, the impulse response H(x,y) factors as F(x)G(y). A Gaussian filter is both radially symmetric and separable, but it's special that way. Replacing a 1D Lanczos filter in x with a 2D filter in radius does not give a separable filter. However, for resizing purposes, use of two 1D filters gives a very good result quite a bit more cheaply. The education links at Rice University might be helpful. --KSmrq^T 15:36, 18 October 2006 (UTC)[reply]

Ah, I think I see it. The circular symmetric function I proposed, needs to be shaped square for us to get the separable version, right? —Bromskloss 15:48, 18 October 2006 (UTC)[reply]

Questions reguarding the Task Manager in Windows XP

One, whenever I turn on my computer, and I log onto Windows as the owner, and I press ctrl+alt+del to bring up the task manager, a window pops up saying "THE TASK MANAGER HAS BEEN DISABLED BY YOUR ADMINISTRATOR" with a red circle and an X in it. So I looked on Google to see if there were anyways to fix this, and I found that by running REG add HKCU\Software\Microsoft\Windows\CurrentVersion\Policies\System /v DisableTaskMgr /t REG_DWORD /d 0 /f in the "run..." in my start menu, it enables me to view my task manager, but only for that current session of windows. If I logout or restart my computer, the task manager becomes disabled again. Please note that I am the only user on this computer, and I am the administrator, so obviously it's an error message that shouldn't come up. My question is, is there a way to permanently enable the task manager so I don't have to run that application to enable it every time I turn on my computer? Thanks.

And then my second question is, I use the task manager to terminate "owner" (as opposed to SYSTEM or LOCAL/NETWORK SERVER) programs that begin at startup. Is there a way to permanently stop these programs from running themselves whenever I turn on my computer? My computer has low specs and if I don't stop those programs immediately at startup then my computer functions very slowly. Thanks for answering my questions. --Ķĩřβȳ ♥Ťįɱé Ø 05:12, 18 October 2006 (UTC)[reply]

For your first problem try this. For your second question, run regedit and browse to HKCU/Software/Microsoft/Windows/CurrentVersion/Run. These items on the right are startup items, just delete ones you don't need (but dont delete anything that you don't know what it is). Also look in the Startup folder in your start menu (both for you and for the All Users pseudo-profile) --⁪froth^{T C} 05:28, 18 October 2006 (UTC)[reply]

Or, for a somewhat less daunting way of managing startup programs, run msconfig (enter it into the run box (Win+R)), and click Startup. -- Consumed Crustacean (talk) 06:26, 18 October 2006 (UTC)[reply]

Regedit seems to be disabled too (!) "Registry editing has been disabled by your administrator". Will the fix for the first problem fix this one too? --Ķĩřβȳ ♥Ťįɱé Ø 07:19, 18 October 2006 (UTC)[reply]

Also, I ran that program at symantec.com and it said it didn't find anything. --Ķĩřβȳ ♥Ťįɱé Ø 08:31, 18 October 2006 (UTC)[reply]

It's definately malware, google around for a fix --⁪froth^{T C} 14:15, 18 October 2006 (UTC)[reply]

Shareaza's Advanced Settings

I'm trying to tweek them to get the most powerful searches, fastest downloads, and anything, and everything that fits along these lines, etc.. For example, I want to get as many possiblities in my bittorrent downloads, so do it make it 100% in BitTorrent.BandwidthPercentage. Like, what does that even mean? Help would be greatly appreciated. You can also reply on my talk page. Thanks! Danke!100110100 10:10, 18 October 2006 (UTC)[reply]

It means that you shouldn't use shareaza because 1) It's horrificly bloated and 2) Because it has poor bittorrent support. Many trackers will refuse to serve clients running shareaza; you're better off using utorrent. Oh and I guess that BitTorrent.BandwidthPercentage is probably the percentage of your bandwidth that is allocated to torrent traffic rather than other "networks" like gnutella which shareaza also supports. --⁪froth^{T C} 14:13, 18 October 2006 (UTC)[reply]

html forms

i was making a form field bit for a web page and i got the

form action="MAILTO:email address"
method="post" enctype="text/plain"

i was wondering can i add anything/edit something that will allow me to control what the subject line of the email will be?--Colsmeghead 14:38, 18 October 2006 (UTC)[reply]

Yes, you can. The body text too, in fact.

mailto:user@example.com?subject=This_is_the_subject&body=This_is_the_body

Remember to percent-encode any characters that need it. For example, if you want spaces (instead of the underscores I used), you use "%20". —Bromskloss 15:39, 18 October 2006 (UTC)[reply]

Oh, wait. You're using a form? Hmm, are you sure you don't just want an ordinary link? Like this:

<a href="mailto:user@example.com">Link text</a>

—Bromskloss 15:43, 18 October 2006 (UTC)[reply]

I think he wants a form so you can type the message in the web page. First - that is a very dumb thing to do unless you really desire getting spam-injection in your web-based forms. If you want to do it, drop the "method='post'" so it is all on get. Then, add the subject as a hidden input tag and keep the textarea you are likely using as the body. --Kainaw ^(talk) 16:13, 18 October 2006 (UTC)[reply]

DFT bins

I often read terms like transform bin or DFT bin in papers concerning DFT and other transforms.. Can you explain me the meaning? --Ulisse0 16:44, 18 October 2006 (UTC)[reply]

Version numbers

I've read the version article, but I have two questions:

What happens if I am at version 1.9 and want to release a new version with a minor update? What should the next version be called? Is version 1.10 understandable? Or would I have had to start with v.1.01 (the extra '0') all along?
The article says that the major version number only gets changed for major upgrades. But what happens if I keep releasing minor upgrades? No one upgrade is enough to say "this is a major upgrade", but version 1.7 (say) is by now really different than version 1.0. Should one arbitrarily say "now this version is different enough than 1.0 that the next minor upgrade will get the number 2.0"?

I'm not so interested in the marketing forces behind version numbers, just the general practice.

Thanks! --Sam

"1.10" is perfectly okay, and has been used in some circumstances. It should be pronounced "one point ten". As for major upgrades, it's entirely up to whomever releases the program. The convention on what constitutes a major upgrade varies wildly, but generally major UI changes constitute major versions. —BorgHunter (talk) 17:52, 18 October 2006 (UTC)[reply]

Personally, I always thought if you have 10 updates at one level, then you don't have enough levels. I always used a three step version, MAJOR.MINOR.FIX, to avoid this problem. After I had 9 fixes, then I would put out a minor version. After I had 9 minor versions, then I would put out a major version. StuRat 18:00, 18 October 2006 (UTC)[reply]

But why would the random fact that we have ten fingers (i.e. that we have a base-10 counting system) have any bearing on when a program should have a major release? Capping a maximum of x updates at one level is something that ought to be decided by the specific program, not fixed by our counting system, I would have thought. No? --Sam

If I were using hex numbering, then I would stop at version F, I suppose. StuRat 21:56, 18 October 2006 (UTC)[reply]

Add another level, [version 1.9.1] is quite legitimate. Vespine 05:18, 19 October 2006 (UTC)[reply]

NPROTECT folder

When I use various scanning programs, it scans a folder (apprently hidden) NPROTECT. So what is in there and why is it hidden? - Tutmosis 18:32, 18 October 2006 (UTC)[reply]

Are you kidding me... --⁪froth^{T C} 19:39, 18 October 2006 (UTC)[reply]

Yea this was a crazy joke I thought up. Man that korean company sure explains why I have a hidden folder. Thanks. - Tutmosis 19:32, 18 October 2006 (UTC)[reply]

Check the link again --⁪froth^{T C} 19:39, 18 October 2006 (UTC)[reply]

Oh... okay thanks. I tried google but my search wasn't as detailed and it just showed info on nprotect.exe. - Tutmosis 19:52, 18 October 2006 (UTC)[reply]

Extending wireless range?

I live in Corpus Christi, where free wireless is being provided by the city. The problem I have is that there are four wireless access points surrounding me, but all four of them are just a little too far away for me to get access on my laptop with built-in wireless. Is there any way that I can extend my range so I can access them from inside my house? 69.154.49.244 19:43, 18 October 2006 (UTC)[reply]

You can get signal boosting replacement antennae quite easily from computer stores/websites, but they are usually for desktop PCs (the antennas themselves are detachable even if the wireless network card isn't). I think Laptop antennas are built in; if not, i'd like to know where to get one too. CaptainVindaloo ^{t c e} 19:49, 18 October 2006 (UTC)[reply]

Laptops with built in wireless cards do, indeed, also have the antenna built in, so you're kind of stuck there. There is, however, nothing preventing you from disabling the built in adapter and using a Pcmcia or Usb adapter, such as this one. Livitup 04:01, 19 October 2006 (UTC)[reply]

One possibility would be to use some kind of wireless bridge? TERdON 23:54, 21 October 2006 (UTC)[reply]

Some registry keys

Got a virus or something a while back that screwed around with my registry. Although I removed the virus, the registry stands broken. An issue is that when I hit start, then {my documents, my computer, my network places, my pictures, control panel, my music}, nothing happens. So I navigated to an area in the registry: HKLM/SOFTWARE/Microsoft/Windows/CurrentVersion/Explorer/StartMenu/StartPanel. I think the virus messed with the entries in there. Specifically, each folder (ControlPanel, MyPics, etc) has three subfolders (Hide, Menu, and Open) that might be malicious. Seeking guidance before I go and delete something. Hyenaste ^(tell) 20:03, 18 October 2006 (UTC)[reply]

Do a Google search for "Tweak UI" ... install it, run it ... there should be repair functionality within it that might help with this issue. If I remember correctly, that is. :) A Google search for "repair Special Folders" might yield some results as well. Good luck! JubalHarshaw 20:17, 18 October 2006 (UTC)[reply]

TweakUI was fun, but the Repair Regedit didn't fix anything. Still checking Google for "repair Special Folders"... Hyenaste ^(tell) 20:41, 18 October 2006 (UTC)[reply]

(edit conflict) I've got Norton SystemWorks, which has a tool called WinDoctor (love CamelCase, don't they?). That would fix something like this. You could borrow a CD from someone and run it off the disc, but I think there is a similar standard Windows tool that would do the job, only I can't remember what it is called... CaptainVindaloo ^{t c e} 20:21, 18 October 2006 (UTC)[reply]

Making money from web design

I'm reasonably handy with web design and CSS, and good at Photoshop too, and I'm sure there's some way to make money out of this but I'm not sure how. I designed a website for a friend's business which netted me a cool £80, but friends with businesses are not thick on the ground and so I need clients. Problem is, I have no idea where to start: I don't have the time or inclination to set up a fully-fledged web design company. Worth1000.com has corporate Photoshop logo contests, where companies pay you money to design a logo for them. Is there a similar thing for web design? Or have you any other ideas? Thanks.

(also, speaking of Worth1000's contests: these seem too good to be true to me. Not that I don't think they're genuine, but it just seems to me that only a small percentage of Photoshoppers are good enough to design acceptably professional corporate logos, otherwise everyone could do it. How easy is it?) Sum0 20:15, 18 October 2006 (UTC)[reply]

There isn't much money in "web design". The money is in "web maintenance". Get a hosted server that can easily hold 10-20 websites. Then, hit up a few small businesses that need cheap websites. You'll get a big payment for the initial design and then charge them every month to host it (and make minor changes). I have many clients who pay every month that haven't asked for even the smallest change to their site in years. Over time, you'll have a rather stable income that allows you to think about getting some kid to do your web design work and then you charge money and do nothing at all. --Kainaw ^(talk) 20:22, 18 October 2006 (UTC)[reply]

Freezing and Jumbled Screen

I recently built a new computer. First build and is going horribly bad.

After finding out the RAM was bad I bought a new pair of sticks and now after getting the system booted randomly the system freezes and the screen jumbles/scatters. Also when I try to reboot the system via the front reset button the system is unresponsive. I am forced to turn off the system via the psu switch. When I turn the switch back on and try to turn the computer on it will not turn on unless i play around witht the motherboards atx connector at which point the system boots but their is no display and the system is unresponsive again. IF i leave turnthe main power off for 20 minutes then the system will boot.

I believe it may be a motherboard problem because my motherboard (asus m2n-e) is riddled with bios problems. I currently have bios 0402 installed and have maxed out the vdimm's to 1.95v. My RAM runs at 1.8v so there is alot of overhead.

Ram timings have been manually set to 4-4-4-8 2T. CPU frquency is at 220 mhz and cpu multiplier is at 11x. In addition the cpu voltage is set to 1.5625 V and is offset by 50mV. These settings have given me longest up time but the system does still crash.

The system specs include:

Amd 4200+ X2, Asus m2n-e, Asus en7600gs (256mb), Kingston valueselect 2X512 pc4200, Enermax noisetakeII 420W, 250 gig WD SATAII hd.

Thanks in advance Rahul.

Your problems sound somewhat like a PSU problem (though it could be a motherboard problem). You can test it with a paperclip and some duct tape as seen here. Unplug it from everything except for a fan or a light, then leave it running for a while and see if anything goes wrong. Safety first: make sure everything's unplugged when you insert the paperclip, cover it with duct tape for safety, and then turn everything on. I've no idea if this will definitely show up any problems, but it might help. Sum0 21:07, 18 October 2006 (UTC)[reply]

I just tried another psu. ThermalTake TR2- 430 W. Same problem but now the jumbled screen wasn't as bad. The original display on the monitor wer scattered and the text was readable. Could my problem be that my psu is too weak for the job? The one i just tried is 430W and the last one was 420w.

Well, I'm no expert but I think that would be enough. If you get the same problems then it sounds like the problem is somewhere else... If you have spare unused components you could try replacing the components in your PC with them and seeing if that solves the problem, in which case the component you replaced would be faulty. I'm not really sure what the problem might be, so you might be better off trying another site like Guru3D. Sum0 09:47, 20 October 2006 (UTC)[reply]

nesting in css

when you nest in css, for example

body table { width:100% }

where that would make the width of all tables in the body 100%, can you do something like:

#myID table.myClass { ...; } thanks.

--66.193.5.99

Sounds like you need to read about specificity of selectors. A more specific selector (which that ID would provide) will dominate a less specific one.

Incidentally, it's (usually) silly to say "body table", since a table could appear nowhere else. --KSmrq^T 08:17, 19 October 2006 (UTC)[reply]

If you're asking about ~~cascading and~~ selectors, then yes, they work that way. --Kjoonlee 09:28, 19 October 2006 (UTC)[reply]

LaTeX formatting

I've got a pretty nice-looking LaTeX document, with just a handful of tables and figures. So few, in fact, that it seems really stupid to have the List of Figures on one page and List of Tables on the other. Is there any way to stop them automatically creating a new page? Confusing Manifestation 20:51, 18 October 2006 (UTC)[reply]

You could probably use \renewcommand to redefine the \tableofcontents not to include a \clearpage (or whatever does the page break) at the end. This might be what the original looks like (found in the file book.cls:

\newcommand\tableofcontents{%
    \if@twocolumn
      \@restonecoltrue\onecolumn
    \else
      \@restonecolfalse
    \fi
    \chapter*{\contentsname
        \@mkboth{%
           \MakeUppercase\contentsname}{\MakeUppercase\contentsname}}%
    \@starttoc{toc}%
    \if@restonecol\twocolumn\fi
    }

Hmm, I don't know how to modify this so as not to make it start on a new page. :-( I see no \clearpage, for example. —Bromskloss 12:20, 19 October 2006 (UTC)[reply]

I almost have an answer, but unfortunately I don't know what makes LaTeX generate thoses lists in the first place. Can you specify more about your file? Is it two-column, for instance, so that the automatic returns to two-column mode force a page break? Are you using some particular package and/or command to generate the lists? --Tardis 20:27, 19 October 2006 (UTC)[reply]

I decided I could live without them, but the commands I used were \listoffigures and \listoftables. Confusing Manifestation 10:10, 21 October 2006 (UTC)[reply]

I use both of those commands for my thesis. I am using the article class and they do not make new pages between them for me. I use \listoffigures before \listoftables if that makes a difference. Ansell 10:39, 21 October 2006 (UTC)

Wikipedia:Reference desk/Archives/Computing/2006 October 18

Contents