Wikipedia:Reference desk/Archives/Computing/2007 March 29

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.

March 29

How does a keyboard work?

I was wondering if someone could tell me how computers are able to take the words entered on a keyboard and translate them into digital information? —The preceding unsigned comment was added by 68.31.231.229 (talk) 00:19, 29 March 2007 (UTC).

There are a few articles you may want to look at: Computer keyboard#How it works, Keyboard technology. --Transfinite (Talk / Contribs) 00:34, 29 March 2007 (UTC)

Bittorent Trackers

Are Bittorent tracker computers that run special server software, computers seeding the torrent with a normal client, or something else? Please explain!--Ryan 01:10, 29 March 2007 (UTC)

A tracker is a server that coordinates the torrent - it is special software, but it is open-source. There is an article here: BitTorrent tracker --Transfinite (Talk / Contribs) 01:13, 29 March 2007 (UTC)

The BitTorrent specification (if you can call it that) is open, and there are open implementations of the BT tracker and client. There are also closed implementations. -- mattb @ 2007-03-29T01:36Z

data recovery

I am writing a paper on how the FBI recovers evidence from hard drives even after the files have been deleted/erased. 72.40.60.1 02:39, 29 March 2007 (UTC)jc

Sorry all, can anyone please lead me in the right direction on this topic72.40.60.1 02:41, 29 March 2007 (UTC)jc

See data recovery, although it's emphasis is on data accidentally lost. Also see undelete. StuRat 02:52, 29 March 2007 (UTC)

Computer forensics is good, although much of the content needs to be interwikied. Splintercellguy 03:34, 29 March 2007 (UTC)

It may be interesting to note that if someone erases something from a hard drive with a secure erase procedure, the data is not recoverable (at least not from that media). -- mattb @ 2007-03-29T04:03Z

Like # dd if=/dev/urandom of=/dev/hda? (from dd (Unix) --antilived^{T | C | G} 08:51, 29 March 2007 (UTC)

That may not be enough, given the right tools. —The preceding unsigned comment was added by 149.135.25.65 (talk) 09:16, 29 March 2007 (UTC).

Data remanence is a great article that descusses how to get rid of data properly. Capuchin 11:57, 29 March 2007 (UTC)

/dev/urandom only gets about 6MB/s, so you better ask the Feds to give you a few hours notice. :) --TotoBaggins 18:00, 29 March 2007 (UTC)

Don't quote me on this, but I'm sure I read somewhere (probably a webpage for a 'deletion' program that securely removes your data) that writing over data once doesn't always mean the data isn't recoverable - and they said in fact that when formatting your drive for this purpose you'd in fact have to do it 'several times'. What was great about this program is that it would overwrite the data several times with gobbledy-gook and then erase it. Rfwoolf 12:02, 29 March 2007 (UTC)

This is paranoia. A single overwrite of all data on the hard drive with ones or zeros or gibberish renders the original data unreadable (assuming of course that the data is actually, physically overwritten). You may be thinking of the infamous Gutmann paper which has been wholly misinterpreted by many an internet fellow. If data on a hard disk has been overwritten, the best you can hope to do is get a probability that a given bit was different before, and this only by statistical analysis and an MFM. Given that analysis of an entire disk platter with an MFM would take about a year of continuous scanning and generate a metric crapton of data, this isn't exactly a feasible recovery method. Secure agencies use much more stringent methods (multiple overwrites), but this is just extra paranoia (the NSA also physically destroys hard drives in a big hard drive shredder). -- mattb @ 2007-03-29T12:42Z

It isn't exactly pure paranoia. Because disk read/write heads have a certain width and because they are only positioned to a certain accuracy, it is possible that a single over-write will be slightly mis-positioned compared to an earlier write, leaving behind a thin "stripe" of the original data. It won't have any effect on the normal read operations of the disk because the new data will certainly overwhelm any remnant "stripe" of old data, but if you are willing to use other methods to read the disk, the "stripe" may, in fact, be readable. Multiple write operations are assumed to have enough distributed positioning error that they will obliterate any remaining "stripes" of good data.

Another term for this whole process is "Data Security Erase".

Atlant 13:52, 29 March 2007 (UTC)

No, it's not pure paranoia. I acknowledged that it is theoretically possible to recover some "truly" overwritten data from hard disks (I just noticed we've been ignoring other media here) with SPM techniques. Perhaps it's more accurate to say that it's "practically paranoia", since it is not (yet) practically possible to recover significant amounts of data using these methods. The best kind of paranoia has a dash of hard science mixed in for flavor. :) -- mattb @ 2007-03-29T17:22Z

Is this the Peter Gutmann paper? It details several techniques which easily (that is, less than $1500 and a day) recover data from several previous write cycles. —EncMstr 17:38, 29 March 2007 (UTC)

It is. $1500 can get you an MFM setup, not the ability to recover overwritten data. See [1]. -- mattb @ 2007-03-29T18:00Z

Is there a way to rip Zoomify images?

I'm looking at this piece of artwork here. It's PD (author died over 70 years ago, published c. 1890), but there's no link to download the full image. Is there anything I can do to get the whole thing short of screenshotting each piece and stitching them together with hugin? (Which I can do... it's just a bit labor-intensive.) grendel|khan 03:02, 29 March 2007 (UTC)

Do they sell any cds of artwork? I'd try the legal commercial route first. In short, yeah, I suspect with the proper tools that it wouldn't be too hard to grab the incoming images as they are being downloaded to the flash player unless there's some sort of encryption/data format obfuscation involved. But Jeez, lets support our public art museums, why don't we? These pieces of art have to be kept up, atmospheres have to be conditioned, etc. Root⁴(one) 04:21, 29 March 2007 (UTC)

I'd just be uploading the image off the CD to Commons, instead of doing it this way. What did you think I was going through this effort for? grendel|khan 05:27, 29 March 2007 (UTC)

And there are other ways for museums to make money other than spurious and false copyright claims. Might as well consider advocating that the rob banks, since after all art has to be kept up, atmospheres have to be conditioned, etc... spurious copyright claims on public domain artwork do not really aid museums in the long or short terms and they certainly don't aid culture as a whole. --24.147.86.187 14:18, 29 March 2007 (UTC)

Also, I wouldn't be so quick to assume that because you think the "copyright" of the image has expired automatically gives you the right to download and print off a copy that someone went to the trouble of meticulously scanning and presenting online. The museum may not technically own the rights to the original work, but they probably OWN the original work and they certainly OWN the presentation you are viewing. Vespine 05:06, 29 March 2007 (UTC)

Why the scare quotes around 'copyright'? I don't "think" the copyright on the image has expired; it has. It's not an opinion and it's not a gray area. Why does it matter for my purposes that the museum owns the original work? How is this different from pulling a plain JPEG off their site? Are you aware that Zoomify is a simple Flash app that lets the user view an impractically large image, and that the presentation to which you refer is a drag and drop affair to set up? I'm not saying they didn't take the effort to scan or photograph the work in the first place, but I am saying that pulling an image from that site is no different from downloading a plain JPEG or, say, scanning photos of public domain paintings from a book. grendel|khan 05:27, 29 March 2007 (UTC)

Sorry, haven't seen the scare quotes article before but I'll pay it, bad habit... Anyway, I still wouldn't be so sure that "no different to downloading a plain JPEG" means it's legitimate way of obtaining it either. If it was that simple, what's the point of the "Copyright © 2006 Fine Arts Museums of San Francisco" at the bottom then? Vespine 05:42, 29 March 2007 (UTC)

The copyright notice is partly because they own the copyright to their site design etc., partly because some of the works in the ImageBase might be photos of sculptures and such (which are copyrightable) and part copyfraud. (Here, for instance, a Canadian library claims that an 1854 treaty is copyright "Government of the United States of America", managing to be wrong in two ways at once.) Now, museums in Britain, for instance, can copyright their reproductions of public domain works (and by prohibiting photography in the museum, effectively remove the work from the public domain!), but in the United States, museums can't do that. To be intimidated by yet another scattershot claim of copyright constitutes an unacceptable chilling effect, saying that we ourselves need to scan any art we use. It's definitely not our policy at the present time; consider this enormous set of images which rely on the noncopyrightable nature of copies of two-dimensional works. grendel|khan 06:03, 29 March 2007 (UTC)

For the record it is not at all clear that they are copyrighted in the UK; it's just that they haven't had a direct ruling on it yet on the level of Bridgeman Art Library v. Corel Corp. (to my knowledge, anyway). However museums in both the UK and the US have all the right in the world to prohibit photography — it is their private space and private property — the question of the intellectual property is totally separate, of course. --24.147.86.187 14:22, 29 March 2007 (UTC)

Another example: Privat-Livemont-Absinthe Robette-1896.jpg. The site it came off of states at the bottom, "No pictures or text may be reproduced or used in any form without written permission of the site owner." On the other hand, I can't nick images from ARTstor here, because I have to agree to a license to get to anything, and part of what I'm agreeing to is not to reproduce the images. I didn't sign/agree to anything to access either the poster I pulled from the Zoomify viewer or that Absinthe Robette poster. If using the latter is kosher, so is the former. grendel|khan 11:28, 29 March 2007 (UTC)

Also, see a more thorough explanation of copyfraud (this is the article that the above-linked abstract goes to), and see a discussion on Commons about similar claims made by the New York Public Library. grendel|khan 11:50, 29 March 2007 (UTC)

As for Zoomify — it uses a Flash player to accomplish said zoomification. Look in the source of the page for the word "zoomifyImagePath", and it will have a path after it. Add ".jpg" to the path and you have your high-res image. For this one, it is http://www.thinker.org/media/images/3/332820130166/images/3328201301660070.jpg. Enjoy. --24.147.86.187 14:25, 29 March 2007 (UTC)

Excellent answer. StuRat 17:37, 29 March 2007 (UTC)

Ha. And this after I'd already gone through two hours of stitching together essentially uniform background grain. I should be lazier. At least it's good practice for stitching together other sorts of images. I cropped it losslessly; it's now uploaded at Privat-Livemont-Ameublement-1890.jpg. grendel|khan 18:13, 29 March 2007 (UTC)

Not really that excellent an answer. Although it is true you can access a large image at the www.thinker.org url given above (actually for an image on www.famsf.org/) using the method described, I think this must be because there are separate large images uploaded in those directories as well as the zoomify content. I checked this on some sites I developed that use zoomify and it doesn't work - as I expected since when you create a zoomify image there is no single large image! You get a "directory listing denied" for the folder path and ""page cannot be found" for the .jpg, at the folder level or in the folder. --Rorymwiki (talk) 11:23, 22 December 2007 (UTC)

composing music

Please do not double post questions, section moved from Entertainment desk.

can you tell me some good user-friendly and intuitive programs to make my own electronic music... orrr if i want to compose midi... or if i want to sample stuff and then compose a midi-like track that uses the sampled new instrument... or if i record myself singing then record myself playing flute or harp how to put them one on top of another or slow one down or synchronize etc.?--Sonjaaa 04:05, 29 March 2007 (UTC)

Look no further then Ableton Live, it does all that and more. I believe there is a free trial you can use to see how you like it.Vespine 04:59, 29 March 2007 (UTC)

My favorite is Sibelius, for scorewriting, but I know a lot of people that like Finale 2007 too, which is kind of the standard. [Mαc Δαvιs] (How's my driving?) ❖ 05:12, 29 March 2007 (UTC)

Hmm...Sibelius is the one of the best, and easy to use, but it's more than $400 last time I checked. Don't be fooled into buying the cheap ones like I did, though (MusicWrite is especially horrific). bibliomaniac15 05:31, 29 March 2007 (UTC)

Really? I swear it was $100 when I last checked! Perhaps I was looking at the academic version? In any case, for scorewriters or recording, or editing audio, if you are going to do anything professional and don't have any money, the best bet is to download it :D [Mαc Δαvιs] (How's my driving?) ❖ 07:41, 29 March 2007 (UTC)

GarageBand. Cernen Xanthine Katrena 11:27, 31 March 2007 (UTC)

question

I was wondering if there is anyway to disable the math-co proccesor on a computer so that if i tried to divide 0 by 0 on a custom made dev C++ program i could cook something on my cpu. —Preceding unsigned comment added by 76.11.5.113 (talk • contribs) 06:36, 29 March 2007

0/0 will always return an error; you can't get around it, you can't cook something with just an illegal op. If you want, run a tight loop instead (while(1);). —The preceding unsigned comment was added by 149.135.25.65 (talk) 09:15, 29 March 2007 (UTC).

A somewhat related article would be Halt and Catch Fire. --cesarb 09:38, 29 March 2007 (UTC)

Back in the days of core memory, it was rumored that tight loops (for example, a "Branch self" instruction) could toast the affected few cores, but I never saw it happen myself and I ran quite a few branch-self programs over the years. They never even seemed to heat to the point of causing data errors.

And for divide by zero, the division algorithms aren't written to simply do subtraction over and over again; they're much more clever than that so they'll catch your error.

Atlant 13:56, 29 March 2007 (UTC)

mediawiki

on a new install of mediawiki, what is the admin's password? 195.194.74.154 11:31, 29 March 2007 (UTC)

I believe the admin's password (or ROOT user) is set during the installation when you create a database for your wiki.
You can find more information here: Help:Installation
You can also try "sysop"
Rfwoolf 11:56, 29 March 2007 (UTC)

Mediawiki does not have a "root". I believe you are referring to the root user for the database (either MySQL or PostgreSQL). By default, no root password is set during installation. For example, you connect to MySQL using mysql -u root mydatabase. For security reasons, the standard MySQL and PostgreSQL installers now ask you in big bold letters to please set root's password after installation. --Kainaw ^(talk) 12:15, 29 March 2007 (UTC)

How to do a particular image manipulation

In the image shown here, notice that the wherever there isn't "something", there is a gray background. The darkness of this backgroung varies in the "y" direction. What I would like to do is take a samples of this varying background at any given y and subtract that value all the way across the image at that y. Did I explain that clearly? Notice down the right hand side there is a wide blank space just for this purpose of sampling the background.

Any idea how I can do this? I have GIMP but am not particularly skilled with it and I am not a programmer. The original images I want to process are relatively large tiffs, if that matters. Thanks!! —The preceding unsigned comment was added by Ike9898 (talk • contribs) 16:53, 29 March 2007 (UTC).

It sounds to me like you want to (in Gimp) go to Layers→Colors→Brightness-Contrast. Increase both brightness and contrast. The lightest gray will become white and the darkest gray will become black. --Kainaw ^(talk) 18:54, 29 March 2007 (UTC)

You could probably get a reasonably good result in the Gimp by applying a mask that was created with a vertical gradient. I thought it was a fun problem, so I wrote some code that does it with Perl. It's not fast, but it does seem to do what you want. You run it with:

 perl -w normalize.pl REFERENCE-COLUMN SOURCE-IMAGE

You can figure out the reference column (x-pixel) on which to normalize using the Gimp's measuring tool. The tool puts the resulting image (with a red line showing the reference column) in out-SOURCE-IMAGE. Here's an example output. Enjoy! --TotoBaggins 19:54, 29 March 2007 (UTC)

#!/usr/bin/perl -w

use Image::Magick;
use strict;

use constant MAX_COLOR => 0xFFFF;

my $err;

my $fname = shift() or die 'usage';
my $reference_column = shift() or die 'usage';

my $src = Image::Magick->new();
$err = $src->Read($fname);
die $err if $err;

my $height = $src->Get('height');
my $width = $src->Get('width');

my $dest = Image::Magick->new();
$err = $dest->Set(size => "${width}x$height");
$err = $dest->ReadImage('xc:white');
die $err if $err;

$" = ',';
my @colors = qw(R G B);
foreach my $y (0 .. $height)
{
    my @ref_pixel = split ',', $src->Get("pixel[$reference_column,$y]");
    foreach my $x (0 .. $width)
    {
        my @pixel = split ',', $src->Get("pixel[$x,$y]");
        foreach my $color (0 .. 2)
        {
            $pixel[$color] += (MAX_COLOR - $ref_pixel[$color]);
            $pixel[$color] = MAX_COLOR if $pixel[$color] > MAX_COLOR;
        }  
        $err = $dest->Set("pixel[$x,$y]" => "@pixel");
        die $err if $err;
    }
}
my $box_right  = $reference_column + 1;
$dest->Draw(stroke      => 'red',
            strokewidth => 1,
            primitive   => 'line',
            points      => "$box_right,0 $box_right,$height");


$err = $dest->Write("out-$fname");
die $err if $err;

Well don't know if this can help fix it but what if you were to select the region and right clinck and then go into 'layer/colours/invert' and that might help.

Mix Lord 00:33, 30 March 2007 (UTC)

I know of something that can help. ImageJ. It allows you to do some math operations (like subtraction of some constant) over entire images (including 16bit gray level Tiffs). Real simple too. Root⁴(one)

I love you guys! I haven't quite figured out how to use your suggestions yet, but at least I am on my way now. ike9898 15:05, 30 March 2007 (UTC)

Strange issue with Quicktime

Hello!

I have a Quicktime .mov (3,4GB, DV at PAL resolution) created with iMovie on a G4 mac. Now I have the problem: On my WinXP SP1 the movie won't play with any version of Quicktime,even not the newest. On my Debian, VLC, ffmpeg,mencoder and MPlayer can't view it. Only Noatun/Kaffeine can.
So: How can I convert this movie (as losslessly as possible) to a "normal" format like mp4?

Regards,84.56.25.16 17:16, 29 March 2007 (UTC)

Would PSP Video 9 do it? You should be able to choose the quality and everything there and it converts to MP4.

Mix Lord 00:36, 30 March 2007 (UTC)

Try Zamzar (www.zamzar.com). It's free! DebateKid 19:56, 31 March 2007 (UTC)

My favourite converter AND viewer is VNC. It's open source and waaaay cross-platform--71.195.124.101 20:23, 31 March 2007 (UTC)

Chatbot Website Creation

I have a chatbot and need to know how to create a website that takes in your input and sends an output, I looked up html and basically ended up with this through editing source code of websites like Google:

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <title>Chatbot</title>
</head>
<body>
<form action="test.txt" method="get"> You say: <input name="input"
 type="text"> <input name="id" value=""
 type="hidden"> <input value="Say!" type="submit"> </form>
</body>
</html>

I think it sends the information to the server but I'm not sure. Even if it does send information to the server how do I use it. (my chatbot's written in C# if that helps) What now? Thanks for all the help. —The preceding unsigned comment was added by 80.109.79.136 (talk) 20:29, 29 March 2007 (UTC).

The form's "action" attribute needs to be set to the actual page or program that will receive the input, and of course that page or program must be on the server. Take a look at this lesson and see if that helps. --LarryMac 00:34, 30 March 2007 (UTC)

Document management system for a private person

I have always more and more documents (scanned, in pdf, in html, pics). I would like to better keep track of them and the a directory with folder is somehow too little. I would like to be able to comment them (without changing them), keep notes of what relates to what, etc.

Is there some simple DMS out there for this kind of thing?

Mr.K. (talk) 21:21, 29 March 2007 (UTC)

PS: I am actually using a free, open-source tool. Mr.K. (talk) 21:21, 29 March 2007 (UTC)