Wikipedia:Reference desk/Archives/Computing/2008 October 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.

October 20

Xbox 360-Ubuntu tethering

after watching this video [1], i was able to 'tether' my 360 to my Mac so they can share internet connections and I won't have to resort to buying a 100$ adapter for the 360. The problem is that my mac and 360 are on opposite sides of the room and i'm not going to move any of my desks anytime soon. I do, however, have a computer with Ubuntu on it on the same desk as my 360. Sure it's a really old piece of junk, but it'll do the job without lag right? So, how can I tether my ubuntu pc to my 360 the same way i tether my 360 to my mac? --Randoman412 (talk) 02:43, 20 October 2008 (UTC)

Since both the Mac and the XBox have a limited expansion potential, I would look at running your internet into the ubunto, and then running the mac from that, and then the XBox from there...--99.185.0.29 (talk) 02:49, 20 October 2008 (UTC)

No, I can't do that. Remember, my mac and 360 are on other sides of the room, and i'm not moving stuff. besides, i don't have the ubuntu on all the time (that whirring would keep me up all night). --Randoman412 (talk) 11:27, 20 October 2008 (UTC)

data compression

how they shrink data or compress a large volume data to a small volume

what is the algorithm —Preceding unsigned comment added by 59.92.92.15 (talk) 10:00, 20 October 2008 (UTC)

Data compression is achieved by physcially crshing the data for hours to produce denser data. They estimate between 10 and 15 miniture people inhabit every Personal Computer purely for the purpose of working your data down to the tightest archive they can. In large IBM mainframes, you can find habitats of thousands who work between the numerous disks and processors. It's the real reason for the distaste of Microsoft, they force this slave labour upon manufacturers, who then have no choice but to distribute the colonies to you. - Jimmi Hugh (talk) 10:05, 20 October 2008 (UTC)

There are lots of different algorithms to do it. One of the simplest ones is to replace long strings of the same character with a count, so instead of writing "aaaaaaaaaaaaaaaaaaaa" you would write "20a", which compresses the 20 characters down to just three. See data compression. --Sean 13:39, 20 October 2008 (UTC)

It depends on what you're compressing. For English text, you could (for example) use a dictionary - give every word in the book a number and compress text by replacing every word with it's number. There could be one special "magic" number that is used to precede words that aren't in the dictionary (eg because they are spelled wrong). There are far less than 60,000 words in most people's vocabulary - and you can store a number up to 65,536 in two bytes. Hence any word with three or more characters would be squashed into two bytes instead of three or more bytes. That's a rather crude scheme - but this kind of "dictionary" compression is a very common approach. You can even go one step further and eliminate the need to keep the dictionary by taking the original document - and give each UNIQUE word in that document a number. The first time you encounter a new word, you store it normally - and give it a number. Each time you find another instance of that word, you replace it with the number. Since most documents will have a lot fewer than 60,000 unique words - the numbers can be stored in less than two bytes.

This approach is pretty much how we compress photographs in file formats like JPEG...except that we cheat! We break the picture up into small tiles (perhaps 4x4 pixel squares) - and make a list of all of the unique ones...but the odds of the exact same 4x4 pixel square showing up several times in the same photo is rather small - much less than the number of times the word "THE" appears in a large text document. So we go further and instead of requiring the tiles to literally match - we let them more-or-less match. Nobody much cares if the color is 'off' by a little bit in a photo - so we find a set of 4x4 pixel tiles that most accurately represent the actual pixels in the image. The more tiles we use - the more accurately we reproduce the image - but the bigger the 'dictionary' of tiles is - and the bigger the number you need to use to represent one tile. That's how the JPEG image compression system allows you to dial in the quality setting you want - the higher the quality you demand - the more tiles in the dictionary and the bigger the file becomes.

Approaches like JPEG that don't PERFECTLY reproduce the original data are called "lossy" compression schemes - there are other compression schemes (such as PNG) that do perfectly reproduce the original (at the price of much bigger files) that are "lossless". However, it is a fact that lossless compression is a hit-and-miss process. Some files simply don't get any smaller when you try to compress them losslessly. That's what prevents you using ZIP to compress a file - then compressing the ZIP file to make it smaller - then compressing THAT ZIP file to make it smaller still. What happens is that the file that ZIP produces is typically impossible to compress any more - so if you ZIP a ZIP file, it probably gets a little bit bigger.

SteveBaker (talk) 02:05, 22 October 2008 (UTC)

I have used PCA to compress pictures. Its pretty sweet! Brusegadi (talk) 04:46, 23 October 2008 (UTC)

Shareware

Is there a program which turns a photograph into a line drawing?? Kittybrewster ☎ 14:24, 20 October 2008 (UTC)

That depends both on what kind of photo and on what kind of line drawing you want. Gimp and Photoshop both have a "find edges" filter which, with a bit of tweaking, can turn some continuous tone pictures into something similar-ish to a line drawing. Inkscape's trace functionality will turn a photo or scan of a diagram or similar line-ish drawing into an editable vector diagram (to a limited extent). Gimp's find edges followed by inkscape's trace can, in some cases (and when used with skill and care) turn a photo into a useful SVG drawing. -- Finlay McWalter | Talk 14:29, 20 October 2008 (UTC)

Also, Gimp and Photoshop have filters to make it somewhat appear that the photo was sketched with a pencil or pen, if that is what the OP means my "line drawing". -- kainaw™ 15:45, 20 October 2008 (UTC)

Vector Magic is a very funky online tool for "tracing" bitmaps to vectors, if that's what you were after. - IMSoP (talk) 17:17, 20 October 2008 (UTC)

Vector Magic is pretty dang good on the whole — better than similar functions in Inkscape, Photoshop, and Gimp by far. Adobe Illustrator CS3 also does a pretty good job these days as well (it has some pretty neat features for testing out all sorts of different types of vector tracing on the fly, and is very fast), but if you don't own it already, it's not worth getting just for this, though you used to be able to download a fully-functional demo that conked out after a month or so, which might be a way to what you want.

The key issue here is that "photograph to line drawing" is very ambiguous. What do you want it to look like? What type of photo is it? As usual, the more specific you can be in your question, the more specific we can be in our answers. --98.217.8.46 (talk) 22:40, 20 October 2008 (UTC)

Thank you. I have a jpg of four people and dogs. The hope is to produce a sketch from it. Kittybrewster ☎ 10:17, 21 October 2008 (UTC)

What do you mean by a "sketch"? Black and white line art? Vectorized art? Are you trying to make it look like it was done by hand? Painted? Drawn? Made by a computer? Can you give us an example of what you would like the final product to look like? Because the ambiguity here is what is really making this hard to answer. -98.217.8.46 (talk) 19:22, 21 October 2008 (UTC)

I am hoping to make the product look as if it had been hand drawn. Kittybrewster ☎ 20:23, 21 October 2008 (UTC)

This is what I got with a simple Sobel filter in GIMP in about 30 seconds of work:

File:Sarah and oliver.pngFile:Sarah and oliver line.png

With some fiddling around, you could do much better.

SteveBaker (talk) 01:44, 22 October 2008 (UTC)

Can you subscribe by email to a RSS Wordpress blog?

Is it possible to get an email subscription to a Wordpress blog that has an RSS feed, but does not give you the option of subscribing by email? --Gary123 (talk) 17:25, 20 October 2008 (UTC)

You might be able to do that via Google Reader, I think. HardDisk (talk) 18:53, 20 October 2008 (UTC)

Hotmail account possibly hacked?

Yesterday I received a message from a Swedish friend, saying he had received spam originating from a Chinese IP address, which had been faked to use the Hotmail address of our mutual friend in Finland. This alone was not unusual, because the two have been members of the same Internet forum for a long time, and it is not unheard of for spammers to harvest e-mail addresses from there. But what is strange is that the advertising tagline that Hotmail automatically adds was in Finnish. There is no way for the spammers to tell from the Hotmail address, or from our friend's IP address, that she uses Hotmail in Finnish - for example, I use it myself in English, even though I speak Finnish natively. Is there a reasonable risk that her Hotmail address was hacked into? JIP | Talk 18:15, 20 October 2008 (UTC)

Maybe, but probably not. If the forum was in Finnish, or contained a bunch of other addresses in .fi, then they could infer that sending in Finnish might be a good idea. Similarly if they harvested an address using a virus or trojan that reads someone's Outlook address book, and they see a bunch of .fi addresses. But then I get a lot of spam in Russian (in Cyrillic) but I don't speak Russian, don't know any Russians, don't have anything to do with any Russian forums, and I don't believe anyone in my address book does either (what an insular lot). -- Finlay McWalter | Talk 18:38, 20 October 2008 (UTC)

The forums is Nordic. It has three sections, one in Scandinavian (meaning Swedish, Danish and Norwegian - they are mutually intelligible enough), one in Finnish, and one in English. Finns are the second biggest group of users, after Swedes. So yes, there are a bunch of .fi addresses. So it might very well be that the spammer just learned from statistical analysis that using a faked Hotmail advertisement in Finnish would be a good idea. Anyway, our friend has changed the password of her Hotmail account, just to be on the safe side. JIP | Talk 19:09, 20 October 2008 (UTC)

You can just look in the headers to see whether it actually came from Hotmail or not. --71.106.183.17 (talk) 21:10, 20 October 2008 (UTC)

Nero 8 Cover Designer

Hi all,

how do I produce Slimcase DVD stuff in Nero8? It only supports the standard-sized stuff :(

HardDisk (talk) 18:41, 20 October 2008 (UTC)

Scissors? Just keep your artwork simple and trim it down. Astronaut (talk) 20:48, 22 October 2008 (UTC)

Virus

I have a proxy-trojan which my KAV cannot delete/neutrlize. I tried my best but any solution has not been produced yet. Please tell me what I can do to delete it and how. Thanks, --87.68.158.69 (talk) 20:23, 20 October 2008 (UTC)

Usually, when i'm in windows and i get a virus, i usually do a reinstall(of windows). Though this depends on the severity of the virus and how much it spread. Mile92 (talk) 23:46, 20 October 2008 (UTC)

I'm amazed to hear you say "usually". How many times has this happened to you? I've used Windows (and MS-DOS) for about 15 years, and I've never once had a computer infected with any kind of malware. I stopped using antivirus software years ago because it didn't seem to do anything except slow my system down. I can't even figure out what people do to get their machines infected. Are you sure you've been infected by viruses many times, or does you system just do mysterious things that you assume are due to a virus? -- BenRG (talk) 00:44, 21 October 2008 (UTC)

Can you give any more details? Does Kaspersky detect the trojan, and if so, what does it tell you about it? Does it interfere with your web browsing in a specific way? If you post a HijackThis log here (inside Template:Show or similar) I'll see if I can glean anything from it. -- BenRG (talk) 00:44, 21 October 2008 (UTC)

Yes KAV does detect it. It is called "Trojan-Proxy.Win32.Saturn.a".... --87.68.158.69 (talk) 08:33, 21 October 2008 (UTC)

Big O notation

Can someone provide a concise introduction of this for someone who finds the main Wiki article a touch too academic? —Preceding unsigned comment added by 82.42.62.245 (talk) 22:45, 20 October 2008 (UTC)

First, a warning... Big O in computer science is not Big O in mathematics. In computer science, Big O merely an approximate number of iterations an algorithm will require. For example, if you have a group of items and your algorithm requires you to check each item once and only once, it is a O(n) (where n is just the number if items you have - you are generalizing with 'n' because the next guy may have a different number of items and he'll use 'n' for his number). Now, assume you have to compare every item to every other item. For every one of the 'n' items, you need to do pretty much 'n' comparisons. That is n times n or O(n²). All in all, computer science Big O is much simpler in computer science than math because in computer science it is much further generalized. If your algorithm does a divide and conquer method, you know you'll see a 'log n' in your Big O. If you look at code and you see three embedded for loops, you will expect to see n³ in the Big O. Does that introduce it enough to read the article? -- kainaw™ 22:53, 20 October 2008 (UTC)

I don't know what you mean when you say "Big O in computer science is not Big O in mathematics." It's the same notation, and it means the same thing in both fields. Big O notation is often used in computer science to express an approximation of the number of operations required by a particular algorithm, but it can also be used to express other things, such as space complexity. —Bkell (talk) 17:37, 21 October 2008 (UTC)

The math in the Big O notation article was hard for me to follow, but in the section about The family of Bachmann-Landau notations, I found the following example which reminded me how Big O notation is used in computer science.

• Example: The running time $T$ for a newly developed algorithm with input size $n$ is $T(n)=73n^{3}+22n^{2}+58$.
  $T(n)$ is $O(n^{3})$, which means $T(n)$ grows asymptotically as fast or more slowly than $n^{3}$.

Big O notation is usually used when analyzing running times of algorithms. Although the example $T(n)$ was specified exactly, you usually analyze the code and determine just the fastest growing term. Big O notation is used to indicate the fastest growing term of a function. --Bavi H (talk) 01:24, 21 October 2008 (UTC)

Actually, that's not entirely correct. $f(n)=O(g(n))$ means that, for sufficiently large $n$, some multiple of $g(n)$ is always greater than $f(n)$. So $n^{2}$, $100n^{2}$, and $n$ are all $O(n^{2})$. Granted, no one would write $n=O(n^{2})$ instead of $n=O(n)$, but it is still correct. « Aaron Rotenberg « Talk « 07:21, 21 October 2008 (UTC)

In the simplest possible terms (for software) - if you imagine a piece of software that processes some units of data. It has to crunch through N units. If the time it takes to do that is proportional to N - then it's an O(n) process. If you double the number of units of data - the program takes more or less twice as long to process it. However, some algorithms need to do things like compare every item of data to every other item. For each of N things - it has to do N operations. The time it takes to run is now proportional to the SQUARE of the number of data items - if you double the amount of data, it take FOUR times as long to do the work. This is an O(n²) algorithm. Knowing what ORDER (that's what the 'O' stands for) of complexity you have lets you predict how well your software will scale when it has a lot of work to do. It's NOT true to say that an O(n²) will always be slower than an O(n) algorithm for any given amount of data. If you have one algorithm to do some job that works at O(n) but takes 8 seconds per operation - and another that is O(n²) that only takes 1 second per operation - then for less than 8 items of data, your O(n²) algorithm turns out to be faster. What the O-notation tells you is that if the amount of data gets large enough - then EVENTUALLY, the O(n) algorithm will be the most efficient. SteveBaker (talk) 01:16, 22 October 2008 (UTC)

It can be a bit confusing. You might like to read this bit on an alternative to make teaching easier Donald Knuth: Calculus via O notation. Dmcq (talk) 12:31, 23 October 2008 (UTC)

I've been disconnected all week, so this is a tad late... In math, Big O of f(n) is any function g(n) such that for any large value of n, g(n) > f(n). So, if f(n) = n², you can use g(n) = n¹²³⁴⁵ as Big O of f(n). In computer science, the Big O of n² is O(n²). It is the smallest g(n) such that g(n) > f(n). In math, that is usually called Big Theta, not Big O. Of course, there are many mathematicians who venture into computer science and many computer scientist who venture into mathematics. So, they drag along their own definitions and you will find the Big O used where Big Theta should be used in some areas of math and see Big Theta used instead of Big O in some areas of computer science. Overall, when discussing Big O, you really need to ask which definition is being used. -- kainaw™ 19:33, 25 October 2008 (UTC)