Computing desk
< September 27	<< Aug \| September \| Oct >>	September 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.

September 28

ISP big brother?

Does my internet service provider know all the sites I visit, dirty or otherwise? Can it follow what I am doing, peering over my shoulder seeing what Wikipedia articles I visit and which natural titty model I'm leering over? And if so, do they have any public policy on purging my browsing history, or, say, if I was running for U.S. Senator could a bit of well slipped cash allow someone to get dirt on me? I find the prospect that they can see what I'm doing kind of scary.--162.83.168.103 (talk) 03:37, 28 September 2010 (UTC)[reply]

It is technologically ~~easy~~ trivial for them to monitor all of your unencrypted communications. As long as your encrypted communications links are not tampered with (man in the middle attack), and are otherwise done properly, those are not viewable. There are laws, federal and state, that protect against blatant monitoring without consent, however you need to review the terms of your service to know what you've consented to. There are also exceptions in the law for monitoring used for some maintenance and of course court orders, but I know of no large ISP in the U.S. that openly monitors their user's ingress/egress data as a general rule. They do store some information for limited times, and websites you visit may store information as well, but again, as a general surveillance, I doubt any large ISPs do. Shadowjams (talk) 03:45, 28 September 2010 (UTC)[reply]

Note that even if your communication with the website is encrypted, they can still probably guess what sites you're visiting by your DNS lookups. One option would be run some sort of programme which randomly visits sites, I think those exist. If done properly, it may be hard for them to guess which sites you're actually visiting. Another is of course to ensure your lookups themselves are encrypted and use a DNS service you trust or alternatively use something like Tor (but if you do, bear in mind exit nodes have far less qualms abount monitoring what you do, and depending on how careless you are i.e. how much unencrypted traffic you do and whether you give away details which can identify you in said unencrypted traffic they may be able to work out who you are.) Nil Einne (talk) 03:57, 28 September 2010 (UTC)[reply]

Well that sucks! I guess it's pretty paranoid and maybe a bit daft to think anyone really cares what I am doing in particular being random internet guy, but still I don't like it. But I'm not very tech savvy. I wouldnlt kno where to begin with setting up a randomizer or a "tor" Also, I don't think a randomizer would fool anyone if they were actually looking because what they'd see is random, random, random, random, and then twenty related sites that are linked by topic because that's what I'm doing so they'd know that was really me.--162.83.168.103 (talk) 04:01, 28 September 2010 (UTC)[reply]

Your ISP publishes their privacy policy, which should contain all the details about what they do and don't monitor. Usually it's available on their website. My ISP sent me a copy with one of my bills. It said they do not log any websites I visit, but then there was a whole section on all the stuff they can do if they get a government subpoena. Read yours and plan your Internet-security measures appropriately.--el Aprel (^facta-_facienda) 04:28, 28 September 2010 (UTC)[reply]

You could also only visit sites you don't want to be tracked visiting via access points owned by others, such as the wireless network of a coffee shop, or McDonalds, or even some public libraries. You may find disagreeable content filters though. The Masked Booby (talk) 04:45, 28 September 2010 (UTC)[reply]

Anonymity is a hard technological problem. The tor project talks about this quite a bit, which may be of some interest to you. Bruce Schneier has a good blog post about the distinction between anonymity and privacy. The former is not at all private, but the identity is hard to link to you. The latter is secure, but who you're talking to is clear. Think about interactions with random strangers... you buy a cup of coffee in an airport. You're anonymous, but you're not private... it's clear to everyone around you bought a cup of coffee. The internet's a little bit like that. Shadowjams (talk) 06:07, 28 September 2010 (UTC)[reply]

Well I'm not going to do anything. I'm not important enough for anyone to really want to follow or log or bribe or expose through my ISP. It's just the principal I don't like. Thanks for the info.--162.83.168.103 (talk) 08:48, 28 September 2010 (UTC)[reply]

You could use a VPN or some other secure Tunneling protocol...but otherwise, yes, your ISP can "see" all the sites you go to...and depending on your ISP, this gets logged for up to a few years. There are plenty of other programs designed to monitor you including Perfect Citizen, Echelon, the Information Awareness Office, ADVISE, Carnivore (FBI)...not to mention more recent efforts.Smallman12q (talk) 00:14, 29 September 2010 (UTC)[reply]

submask

how doesa submask work????????please explain with an example...... —Preceding unsigned comment added by Naveenkumarrocks (talk • contribs) 12:16, 28 September 2010 (UTC)[reply]

I assume you are referring to a "subnet mask" by merging the two words into "submask". Subnet is short for subnetwork. It is a part of a network. In most cases, the network is all of the addressable internet. Using IPv4, every node on the internet is addressed with four numbers in the form 1.2.3.4. Each number can be up to 8 bits long (00000000 to 11111111, which is 0 to 255). There are reserved addresses, but that isn't important here. Due to the organization of the internet, addresses with sequential numbers are often on the same local network. All of the computers on a local network are considered a subnet. A subnet mask is a way of identifying which binary digits of the address are used to define the subnet. Lets assume that all of the first three numbers are requires. So, we need 1.2.3.* since every node on our local network has the address 1.2.3.something. The last number has 8 bits, but the first two never change for our local network. That means that the digits that never change for our local network are 11111111.111111111.11111111.11000000. The 1's never change. The 0's do change depending on which node you are addressing on our local network. That is the subnet mask. Converting binary to decimal you get 255.255.255.192. Now, by doing a simple masking operation on an address, I can see that 1.2.3.16 is on the same local network as 1.2.3.4. I can also see that 1.2.3.49 is not. -- kainaw ™ 14:05, 28 September 2010 (UTC)[reply]

And the point of subnet masks is that networking devices (e.g. PCs) know that they can reach other PCs on the same subnetwork using (normally) Ethernet addressing. So they perform an ARP function to find the Ethernet (MAC) address and send the frame direct to that address, rather than sending it to their local gateway for forwarding to the wider world. --Phil Holmes (talk) 16:36, 28 September 2010 (UTC)[reply]

If your subnet mask is 255.255.255.192, then 1.2.3.49 is going to be on the same local network as 1.2.3.4 and 1.2.3.16. However, 1.2.3.149 (for example) is not. JIP | Talk 16:55, 28 September 2010 (UTC)[reply]

Correct. I was sure I had at least one major typo, but I didn't catch it on a read-through. -- kainaw ™ 17:02, 28 September 2010 (UTC)[reply]

Article CIDR has more information on this. --antilived^{T | C | G} 03:35, 29 September 2010 (UTC)[reply]

More on facebook login

I asked a question a few days ago. Since then I've worked out that href="http://www.facebook.com/login.php?email=myname%40mydomain.com" gets halfway to doing a login. I tried adding &password=abcdefgh but that didn't fill in the password field. Anyone know how to achieve that? -- SGBailey (talk) 13:00, 28 September 2010 (UTC)[reply]

That can only answered by people who know about how that page is implemented; there's no way to see what parameters a particular page can take. My bet is that no such parameter exists because there's no secure way to use it. Publishing such a link would involve publishing the account's password. Using such a link without publishing it is probably unsafe, too, since the contents of URLs are not treated with as much care as form variables in POSTed forms, and for that, there's the "remember me" option and browser-saved passwords. I bet there are browser extensions for people who want to juggle multiple saved passwords. Paul (Stansifer) 13:49, 28 September 2010 (UTC)[reply]

If it has to be a HTTP POST, you could take this code and wrap it into a bookmarklet. --Sean 15:27, 28 September 2010 (UTC)[reply]

I'll try that. Thanks. -- SGBailey (talk) 21:53, 28 September 2010 (UTC)[reply]

DPI in Irfanview

I've been looking at some jpg images in Irfanview and pressing the 'I' key for information. I've noticed that they have different values for DPI, which I assume is dots per inch. After changing the DPI, the image does not appear to change in size. Why do images have different DPI, what is its significance? Does this mean that images with high values for DPI could be enlarged yet still have the same amount detail compared with low DPI images? Thanks

Supplementary question - how do I get images to open by default in Iranview rather than in the "Windows Picture And Fax Viewer"? I use WinXP. Thanks again. 92.28.249.130 (talk) 14:16, 28 September 2010 (UTC)[reply]

DPI is used primarily for printing (I can't think of any other use). It is printing size preference. Do you prefer the image to be printed at 30 DPI or 300 DPI? For print media, DPI is important. When sending images around in JPG format, DPI is used to quickly decide how large an image can be before it gets pixelized. If you aren't in print media, you can ignore it since it doesn't affect the image on the computer screen. As for the default application in XP, I remember it being: hold down shift and right-click on the file. You will see the new option "open with". Find the program in the open with dialog and check the "always open with this program" box when doing so. -- kainaw ™ 14:42, 28 September 2010 (UTC)[reply]

DPI in the sense that you are seeing there is an internal value that relates only to a calculation of what you'd ideally want the output size of a given set of pixel dimensions to be. There are lots of attempts to explain it in more practical terms in the Reference Desk archives, other than the Dots per inch article we have. Explaining it can be a little tricky if you are not used to thinking about images for the purposes of printing quality (which you presumably are not, since you are asking about DPI in the first place).

The short story is that metadata DPI settings don't have anything to do with the pixel dimensions of the image. They have to do with how it is rendered on an output device. Monitors generally reproduce images at a 1:1 pixel ratio (which can vary in real-world DPI, but 72 and 96 dpi are usually the values you use in estimating), so changing the DPI setting won't change how it looks on screen. What matters in the end is the purpose of your output image. If you have a 300 pixel by 300 pixel image, and you print it out on something that requires 300 DPI to look "good", it will only print out "good" at 1" by 1". If you try to print it out at 2" by 2" it will be twice a poor (150 dpi). What you need to do to figure out how much "detail" an image would have when printed out, you first figure out how big you'd want the printout to be (e.g. 5 inches across), then figure out backwards what an ideal DPI would be for the device (300 dpi is pretty standard as a minimum threshold for things looking OK, so that would mean your image would need to be at least 1500 pixels across. Depending on your output device, you might want many more pixels than that). You can set the internal DPI of an image to any arbitrary amount, but it doesn't affect the total pixels. So our 300 pixel by 300 pixel image might have an internal DPI setting that says it is meant to be 150 dpi (and thus could be printed out at 2 inches by 2 inches), or it could have an internal DPI setting that says it is meant to be 3000 dpi (and thus could be printed out at a maximum length of .1 inches on each side). None of that would change the amount of pixels in the image, just how it is processed by a printer. The value of the DPI setting of a particular image does not, by itself, tell you anything about the amount of detail in the image; that's still always going to be in the pixel count. --Mr.98 (talk) 17:39, 28 September 2010 (UTC)[reply]

So in short, the DPI figure only affects the image when printed, and is not used when displayed on a computer screen? 92.24.188.89 (talk) 18:10, 28 September 2010 (UTC)[reply]

Basically... except it doesn't really affect anything. It's the number you get when you say, "I have this many pixels, and I want to print it out at this size." It's just an expression of that ratio. --Mr.98 (talk) 18:58, 28 September 2010 (UTC)[reply]

The DPI setting could be used for screen display. Modern versions of Windows, Linux and Mac OS are aware of the physical size of screen pixels and will report it to any application that asks. If your image says it's sampled at 300 DPI and your screen has 100 pixels per inch (a typical value), then scaling the image down by a factor of three horizontally and vertically will cause it to appear at the correct size. For the most part, image viewing software doesn't do this. I'm not sure why, but I think it just isn't seen as an especially desirable feature. Rescaling degrades the image quality, and most images don't have a "correct" size anyway. It's more useful to display them pixel for pixel, or scaled to fit on the screen with a minimum of wasted space. -- BenRG (talk) 20:10, 28 September 2010 (UTC)[reply]

mshome.net

What is mshome.net? The remote desktop program is connected to that site, but there doesn't seem to be anything there when I browse to it 82.44.55.25 (talk) 18:35, 28 September 2010 (UTC)[reply]

A whois lookup shows that Microsoft owns the domain, for what it's worth. You can grab it yourself if Microsoft forgets to renew it by the expiry date of May 8, 2012. Comet Tuttle (talk) 20:39, 28 September 2010 (UTC)[reply]

Word count in Publisher 2010

Hi there. Does anyone know how to get a word count in Microsoft Publisher 2010? Or if there's an add-on for this? the wub "?!" 19:31, 28 September 2010 (UTC)[reply]

See the following pages.

—Wavelength (talk) 19:39, 28 September 2010 (UTC)[reply]

Well yes, there's loads of word counters about (heck, I could roll my own in Python easily enough), but I'm really after something that can be used in-program. I should add that it's not for me but a customer, I wouldn't touch Publisher with a ten-foot pole. the wub "?!" 19:58, 28 September 2010 (UTC)[reply]

Googling around, it seems there is no built-in function. Apparently you have to do an Edit > Select All > Edit Story in Microsoft Word. Which is about the same as just copying and pasting it. Not very impressive, Microsoft... --Mr.98 (talk) 22:30, 28 September 2010 (UTC)[reply]

Matlab Array Column Misalignment - Help!

Hi there,

I recently asked the following matlab question - http://en.wikipedia.org/wiki/Wikipedia:Reference_desk/Archives/Computing/2010_August_22#Matlab_-_simple_column_vector_formatting_question - about formatting my column vectors in order to get them to display in a row of different length columns without trailing zeros. The solution given has worked in that sense, but unfortunately Matlab is making my output misaligned when it finishes, because the top row contains some 1 and some 2-digit numbers, whereas the second, third etc. rows contain only single digits - for example, I end up with something like

1   0   3   5   7   11   13   17   19   23   29   31   37   41   43   47   53   59   61   67   71   73   79   83   89   97
2   0   0   2   0   0   2   2   0   0   2   0   2   2   0   0   2   0   2   0   0   2   0   0   2   2
        1       1   1           1   1       1           1   1       1       1   1       1   1

that, rather than having all 3-element columns aligned - the second and third rows are all correct, but because the top row has 2-digit numbers it goes wrong. Despite this, if I copy the string into any text-editor/web browser etc, or even into the Wiki editing box, it displays properly, in fixed columns - so it's clearly -storing- the string right, just not displaying it as I want. Is there any way to solve this problem at all? I've tried outputting the data as a matrix, adding column by column (with trailing zeros to make the vectors the right length), but this comes up with exactly the same problem.

Ordinarily I would just copy the output across to MS Word or something, but unfortunately this is for a project where we have to display the output of our programs by printscreening Matlab, so I have to get it displaying correctly in-program, if possible. Could anyone suggest anything? Perhaps it's not a matter of fixing the program but actually changing the Matlab settings themselves, but I have no idea after searching for a solution. Many thanks, Typeships17 (talk) 23:41, 28 September 2010 (UTC)[reply]

Disclaimer - I've never used Matlab. Having said that, if your maximum value is 99 can't you just zerofill your single digits to always be 01, 02, 03, etc? That way they'll line up in your viewer. Zerofill is a pretty basic function, and every programming language should have it. The Masked Booby (talk) 00:47, 29 September 2010 (UTC)[reply]

I too am not a Matlab user, but taking a quick look at the documentation, the fprintf function may serve your needs. For example, the statement fprintf('%5d', A); should (if I understand the documentation correctly) format all of the numbers in the array A as decimal integers right justified in a 5 character string. if A contains floating point values instead of integers, you might need to try fprintf('%5d', round(A)); to convert the matrix to integers. I hope this helps. -- Tom N (tcncv) talk/contrib 05:30, 29 September 2010 (UTC)[reply]

Follow up. Looking at the example referenced above, your solution may be to replace the num2str with sprintf. Try changing A = num2str(V); to A = sprintf('%5d', V); or A = sprintf('%5d', round(V));