Wikipedia:Reference desk/Archives/Computing/2009 July 16

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.

July 16

Sniffing BitTorrent Traffic

If I am using BitTorrent (Azuereus client) to download movies, will someone, say an ISP or a hacker, be able to sniff or find out the content of my traffic and hence, know what movies I am downloading? Acceptable (talk) 00:35, 16 July 2009 (UTC)

yes, this is very easy to do with each client. If you look at the piece graph/information of each torrent it will list your IP and anyone elses that is uploading/downloading. No need to packet sniff. it is public info Ivtv (talk) 01:43, 16 July 2009 (UTC)

This is true if the ISP/Hacker is taking part in the file upload/download. What about others who are not taking part in this way (such as the ISP providing internet access to the user)? --Jwoodger (talk) 03:14, 16 July 2009 (UTC)

Any user can see this data. It is clear on any client. I use BitComet and Under your tasks there are menu's you can click. One menu, I forget the name at the moment because I am at work you can see every IP attached to the torrent either uploading or downloading and how much of the torrent you have accumulated and your upload/download speed. It is out there in the open. Then, if you really wanted too you can take those IP's and do a whois to determine what ISP it is attached too. The short of it, when you download or upload via torrents, your IP is CLEARLY visible.

edit: Ok, I think I mis read your question. To re-answer it, Anyone that is using a TORRENT program that has the same torrent as you can see your IP on it. so if you are downloading show X, and someone else is downloading show X, they can see in the ip list your ip and you can see theirs. But if you were downloading show X and they were downloading show Y, they cannot. Your ISP, if they wanted too(even tho it is illigal) can record what you are doing while you are using their service. Also, I know that when you download a torrent your ip can be stored somewhere. So, if people wanted too they could gain access to the information you are worried about

Ivtv (talk) 04:37, 16 July 2009 (UTC)

Basically, if you don't want your ISP (or hackers, although that's unlikely anyway) to be able to view anything you're doing on the internet, you need to use some sort of encryption. Most secure websites (like banks and such) use SSL/TLS. BitTorrent clients, for the most part, don't have any kind of encryption. It would slow down your downloads, and all the other peers would have to support the same kinds of encryption. The pirate bay guys were trying to make an encrypted network with IPREDator, but it's still beta-invite-only. So, yeah, until something better comes along, anyone who can intercept your traffic can tell how much porn you download. Or whatever. :) Indeterminate (talk) 06:07, 16 July 2009 (UTC)

At least µTorrent supports protocol encryption and it's possible to set it enabled or forced. In the forced-mode it'll only connect to peers with encryption enabled. The torrent-files themselves aren't encrypted (unless you download it from a https server). Thykka (talk) 13:20, 16 July 2009 (UTC)

Note that as the internet transitions to network switches instead of network hubs over more of the hops, "fewer" people can intercept your traffic. This does not mean "zero" people can intercept your traffic; especially if you do not trust the ISPs to respect the privacy of your data. As a general rule, you can not be certain which network(s) your data will traverse when you connect to the internet, so you must assume that it is possible for some untrustworthy third-party to intercept your data. Nimur (talk) 09:16, 16 July 2009 (UTC)

How does protocol encryption help? The way your ISP will find out what you're downloading is when the copyright holder (or the FBI, depending) logs onto the torrent themselves and saves the list of peers. You can encrypt your packets all you like, but that won't stop them from getting a list of everyone who has downloaded a particular contraband file. APL (talk) 15:28, 16 July 2009 (UTC)

APL, that's a valid consideration. With modern encryption (properly implemented), we can "guarantee" that no third-party can eavesdrop. However, as APL points out, the fallacy with peer-to-peer networks is that the "second" party (whomever you are connecting to) is an unknown individual - so it is technically meaningless to verify that you have a secure connection to an anonymous recipient. Nonetheless, this is a totally different class of network eavesdropping (in both the technical and the legal sense). Encrypting the protocol categorically denies eavesdropping capability to anybody who is not an active participant in the transmission. (This may actually be beneficial, from a legal point of view, because it forces the eavesdropper to actively participate in a protocol connection. In some jurisdictions, this can implicitly bind all parties to certain terms of service, etc., etc., and certain obtained evidence may be inadmissible in court, etc, etc. Of course, your local jurisdiction may vary and Wikipedia can not give legal advice. Nimur (talk) 18:38, 16 July 2009 (UTC)

I'd just like to point out, to the 'at least uTorrent' comment, that the main protocol encryption used was developed between the uTorrent and Azureus Inc. people, way back when. Azureus/Vuze also has the very same 'force encryption' modes, with at least two different encryption methods.

On topic: a third party, not part of a particular torrent's swarm, would have a somewhat hard time figuring out what you were transferring with even plaintext encryption on. RC4 is a wee bit more effective, though (somewhat effective shaping tools don't really care though, BitTorrent traffic is somewhat easy to spot). Also, remember that the encryption is only on the protocol headers, not on any of the piece data! Washii (talk) 10:24, 18 July 2009 (UTC)

Playstation/PC

Kind of a dumb question... I recently bought a Playstation I game, and rather than drag out the Playstation my roommate has, I was hoping to play it on my PC. I run Ubuntu and have PCSX. However, my computer couldn't even mount the CD, quite possibly due to the black underside. Is there a way around this? I could just try to download the ROM from the internet I suppose, ha ha, suggestions? Magog the Ogre (talk) 03:07, 16 July 2009 (UTC)

You tried to mount the disk through the emulation software correct? I've never been able to mount the disk in the traditional windows way of clicking on it in "My Computer" (Or the Linux way, for that matter), but doing it through the emulator software worked just fine.

Failing that, it could be a problem with your CD drive. I know that Playstations with slightly "misaligned" laser assemblies will sometimes read silver disks but not black ones, or vice-versa.

Incidentally, In my opinion a lot of PlayStation 1 games look a LOT better on an emulator than they do on hardware. It's amazing the difference you get drawing the exact same geometry on a modern PC video card. I was disappointed when I discovered that the PS3's emulator didn't work like that. APL (talk) 05:47, 16 July 2009 (UTC)

Thanks all, you were right APL; I had to download a BIOS from the internet, and it opened, though it appears this game is not running correctly, and unfortunately I must drag out the old PS1 beast itself. Magog the Ogre (talk) 11:41, 16 July 2009 (UTC)

SATA hard disk for a 2006 MacBook

Hey guys, I have not been here for a while now ... I guess things have changed ...

I have a question, I habe a 60 GB hard disk on my 2006 MacBook and I am considering getting a bigger alternative. How does this http://www.amazon.com/Western-Digital-Scorpio-Drive-WD2500BEVS/dp/B000SIG5QW/ref=sr_1_3?ie=UTF8&s=electronics&qid=1247714298&sr=8-3 look? I don't think I need to get much bigger than this ... any suggestions? How about the price? Please respond ... :) Kushal (talk) 03:25, 16 July 2009 (UTC)

At 5400rpm, it might be a little slow, but it shouldn't be too bad. And that price is pretty good, those drives usually go for about $90. I did about the same kind of upgrade on my laptop (40gb -> 250gb) and it totally gave new life to my old laptop. Indeterminate (talk) 05:58, 16 July 2009 (UTC)

Thank you for your quick response, Indeterminate. I already have a 5400 rpm HDD (60GB 5400-rpm Serial ATA - http://support.apple.com/kb/SP31). It might be worth mentioning the model number is MA254LL/A. One more question, is there any reason I should get a SATA 3 Gbit/s instead? Will it even work? Thanks. Kushal (talk) 14:53, 16 July 2009 (UTC)

Well, it'll be faster. :) But yeah, if your laptop is 3 years old, it might actually not be faster - it might fall back to 1.5gbps. If there's a big price difference, I wouldn't bother, but it's your call. Indeterminate (talk) 23:57, 16 July 2009 (UTC)

The price difference is not that great. and perhaps I could reuse the newer drive (hopefully I don't have to) later on. Thanks a lot. Now the only question is the disk size ...I wonder if 320 GB will be big eno.. Kushal (talk) 14:28, 18 July 2009 (UTC)ugh. .

Haha I meant enough. :P BTW, I am also looking at

http://www.amazon.com/exec/obidos/tg/detail/-/B001JSSDGU/ and unable to decide which one. :( Kushal (talk) 21:36, 18 July 2009 (UTC)

(I'm in the business, I always get asked the "big enough" question! :) Big enough for what? Because this article is about to get archived, I'll try to answer blindly. If you are storing pictures and music, and compressed video, You'll take a long time to fill a 320GB. And if you fill it quickly... Well, suffice to say the 500GB won't be enough either. That's when you start considering a large external drive. Hope this helps, or try my user page! Mxvxnyxvxn (talk) 04:55, 22 July 2009 (UTC)

Thanks for your reply. To tell you the truth, I feel stupid to be asking this question ... I think I will be dealing with photos and some digital videos from my i85. I don't keep my video clips for long, most get deleted within a month and some end up on YouTube (usually as private videos). I am glad you answered my question. I am pretty sure I will be fine with a 320 GB internal hard disk drive. Thanks. Kushal (talk) 15:35, 22 July 2009 (UTC)

i think my memory card reader is disabled

i think my memory card reader is disabled (the thing that you put SD cards in). is there a way to re-enable it? it worked fine before I did a system restore, now it doesn't recognize any SD cards and doesn't show up in "My Computer".--Drawngray19 (talk) 12:48, 16 July 2009 (UTC)

Go to the Device manager it may be that that piece of hardware has been disabled. Here's a guide on how to get to it if you don't know how to (http://pcsupport.about.com/od/tipstricks/f/finddevmngr.htm). ny156uk (talk) 17:52, 16 July 2009 (UTC)

i was looking there already, but i can't seem to find anything that says "memory card reader" or "5-in-1 reader" so i dont know what to do.--Drawngray19 (talk) 15:50, 17 July 2009 (UTC)

network adapters??

how do i find which network adapter i have in my computer (ie. wireless)?? and which driver do i install to make "airopeek" work on my computer??

sushil —Preceding unsigned comment added by Sushil shenoy (talk • contribs) 13:38, 16 July 2009 (UTC)

I have never used Airopeek, but to find out about your network adapter, first open the Device Manager and then click "Network adapters". You didn't tell us what OS you are running on your computer, so for instructions on how to access the Device Manager, I'll refer you to our Device Manager article. Tempshill (talk) 16:08, 16 July 2009 (UTC)

i run on win xp !! simply the best !! —Preceding unsigned comment added by Sushil shenoy (talk • contribs) 15:51, 18 July 2009 (UTC)

Sending huge files as time intervals

Is it true that if we had virtually zero latency across networks, we could send arbitrarily large files almost instantaneously? If we converted the binary value of the file into a number, x, and then sent two pings to another computer, x miliseconds apart, the other computer should be able to create the original file, right? (Idea from Dinosaur comics).

Even if x were very, very large, couldn't we compress it even further, by taking factors of that number and sending those factors all at the same time? Say for instance the factors of x were y and z. We could send the transmission like so:

ping!----[y]----ping!
  pong!--------[z]-------pong!

So we should be able to send any number, no matter how large, in a very short time, with enough factors, right?

Obviously, we don't have zero latency. But we could have some known time slice that was large enough that both computers could be sure they would both measure accurately, say "number of half-seconds" between the two pings. Two computers on the same network should at least get that kind of accuracy. Then if you broke down the factors so they were small enough, each one would only be a few "half-seconds."

Are people already using this kind of compression?

— Sam 14:00, 16 July 2009 (UTC)

Comment : Why are you factoring the numbers? That's an expensive operation. Why not just break the file up into chunks of some number of bytes in the traditional way? APL (talk) 14:13, 16 July 2009 (UTC)

Or that. I just figured factoring would allow the number to be broken into even smaller chunks, but I don't know which would be more efficient. Point being, either way you could break a large number down into many smaller ones and just send those smaller ones as very short time intervals, all at the same time (or very slightly offset). Would that work, and does it already exist? — Sam 14:30, 16 July 2009 (UTC)

Your factoring is nothing more than a form of data compression. You are using a well-defined method to take a large data item and represent it as a small data item (two or more factors). So, you are asking: "Can we use data compression to transfer large files over a network faster than sending the large file itself?" Yes. It is used all the time. Most large files are compressed in some fashion before being transferred over a network. There are also programs for some slower networks that compress all traffic on one and and decompress it on the other. -- kainaw™ 14:34, 16 July 2009 (UTC)

Using time intervals between pings is not a practical method of compression. Here's why. Let's assume we chop a document into kilobit chunks and turn each chunk into a binary number between 0 and 2¹⁰⁰⁰. Half of these numbers will have a 1 in the most significant place, so average value is about 2⁹⁹⁹ (give or take 1). Some of the numbers will be prime numbers, for which factorisation will not help. So at least some of the time intervals involved will be of the order of 2⁹⁹⁹ time slices which is 2⁹⁹⁸ seconds. Turn this into years, then compare it with the age of the universe.

Now you could refine the method by using smaller chunks, which makes the average time intervals shorter. Ultimately, you can use a separate pair of pings for each bit - say half a second apart for 0, one second apart for 1. Then what you have done is reinvent Morse code. Gandalf61 (talk) 14:46, 16 July 2009 (UTC)

This is why I was suggesting taking factors, and not just chopping up the number. Say we were sending a kilobit. That would be a number in the order of 2¹⁰⁰⁰, right? So if we take the hundreth root of that, we get 1024, So if we sent out a hundred pings in a row, and sent a hundred pings again 1024 ms later, then that would be read as 1024¹⁰⁰ = 2¹⁰⁰⁰. Now obviously I'm seeing that this is silly, since it should take much less than 1024 ms to send one kilobit of data. Is there any case where this might be useful? If the clocks were so synchronized that they could count the nanoseconds accurately, say, then 1024 ns would actually be very fast, right? And it would always be able to send the kilobit in 1024 ns no matter what the speed of the network (or rather, 1024 ns + initial latency) if the computers could confidently measure the time interval. No? (Well, I guess there are a lot of hypotheticals in that last thought, so I guess we're in hypothetical land now...) — Sam

63.138.152.238 (talk) 15:15, 16 July 2009 (UTC)

If such reliable clocks could be synchronized, they could be used for transmitting data directly. The flaw here is that the bit-rate is actually very low on the average. Assuming a uniformly random distribution of numbers to be sent, ranging from 0 to N, you need an average of N/2 symbols. It "seems" like you are sending very little information, ("two bits"), but in fact you are simultaneously sending "no information" (... zero-value bits) for the entire intermediate time interval. This must be counted as part of your coding scheme (because if you sent anything in the intermediate time, you would change the transmitted message). Like all perpetual-compression-schemes, this method cannot and will not violate the Shannon information theorem - the data transmission rate can be no faster than half the Nyquist-rate. You might be interested in other methods, like CDMA, PSK, and the like - these use chipping or symbol coding by methods other than "binary bit" - but in the same sense, they always must be limited by Nyquist criteria. (The Computer Scientist can invent whatever coding method he wants, but when he passes it off to the Electrical Engineer, it must have a hardware manifestation which is invariably going to be bound by certain physical law)... Nimur (talk) 18:50, 16 July 2009 (UTC)

If you have nanosecond accuracy in the timing of when you can send signals out on the wire, you're likely to have at least a 1 bit/ns throughput. That is, as you count off the nanoseconds, you can monitor the communication channel. If there's signal, it's a 1, if not, it's a zero. So in 1024 ns, you're likely to be able to transmit at least 1024 bits of information in such a fashion ... which is exactly equal to the kilobit you're trying to encode. The only possible application for something along the lines of your idea that I can envision is if you have some sort of system where the recovery time for the signal is much longer than the timing accuracy. That is, if you can initiate a signal with nanosecond accuracy in any arbitrary nanosecond, but once you trigger the signal, it suppresses further signals for a significant period of time. An example is a theoretical system transmitting signals with high intensity laser pulses. The signals can be very brief and timed to the sub-millisecond level, but it might take the capacitor bank several milliseconds to recharge for the next pulse. -- 128.104.112.87 (talk) 18:48, 16 July 2009 (UTC)

You just can't win...the laws of information theory are every bit as rigid as the laws of thermodynamics (to which they are closely related). One law is that you can't compress arbitary data without loss. Your scheme MIGHT compress numbers that happen to not be prime - but prime numbers take MORE space to send because you have to send the two factors '1' and the full value of the number. When it all averages out - you lose. Perfect, lossless compression is like perpetual motion machines - it's DEFINITELY not possible. So, forget all of the compression...all you have left is "turn the file into a number - 'send' the number as a delay. Fine - OK - so let's send (say) the Declaration of Independance (of Lithuania as it happens!): it's 818 bytes. So if you send that as a number, it's 6544 bits long - making the length of your transmission significantly longer than the expected life of the universe! So you can't do that. How about sending each letter as a delay in the range 1-256 ms? Well that's only 209 seconds...but we can do better. If we chop each letter into two 4-bit numbers in the 1-16 ms range - then your system can sent the 1636 four bit numbers in a mere 26 seconds! If we go to 3272 two bit numbers in the 1-4ms range then it's only going to take 13 seconds...and (guess what?!) if you send each bit as a delay of 1 or 2 ms - then you can still do it in 13 seconds. But hold on - rather than sending "START....delay...STOP" why not just send "ONE" or "ZERO" - no delay is needed - and you only need to send one 'special code' instead of two. In fact, what we've just demonstrated is that the most efficient possible way to send the lithuanian declaration of independence is as a simple, uncompressed binary code. Boring, eh? Well - that's life. You can't do better - it's quite simply impossible - and that's as true as that perpetual motion machines are impossible. SteveBaker (talk) 01:54, 17 July 2009 (UTC)

regarding pointers

can pointers be declared using different datatypes? can a float value be stored using *ptr? —Preceding unsigned comment added by Srividhyaathreya (talk • contribs) 14:11, 16 July 2009 (UTC)

What language? Pointer handling is different in different languages. APL (talk) 14:13, 16 July 2009 (UTC)

I'm sure in most common languages the compiler will work out/require that the pointer is associated with a particular datatype, so the answer is yes in general, when you are using a high level programming language such as pascal, or if using C. In a lower level language such as assembler you have to do this yourself sometimes - ie the you need to remember what type of data a memory pointer is pointing to. Are there any exceptions?83.100.250.79 (talk) 15:18, 16 July 2009 (UTC)

The obvious issues are if you try accessing data that might not actually be a float (or might be in a different byte order). In that case, you can still read it or write it as a float, but the actual value will be garbled if it is re-accessed as something else. Another worry is word-alignment. In C, for example, a byte* can point to any address; while a float* should point to a 4-byte-aligned address (on many architectures). Some machines will give you an alignment exception if you use a 32-bit float on a non-word-aligned pointer. Others will re-align the word at runtime. Nimur (talk) 19:44, 16 July 2009 (UTC)

And if you are using C, you could of course destroy yourself if you had used malloc to reserve enough space to store a 16-bit word, but subsequently you end up storing a 64-bit float. Tempshill (talk) 21:10, 16 July 2009 (UTC)

Don't frighten the questionees - memory overflows do not cause one to physically disintegrate.83.100.250.79 (talk) 22:21, 16 July 2009 (UTC)

Also think of the questioner, we wouldn't want to frighten him or her either. --Anonymous —Preceding unsigned comment added by 70.254.170.8 (talk) 01:06, 17 July 2009 (UTC)

the damage has probably already been done ;(

Pointer (computing) might answer any specifics. in C I think it's

float *peter to declare a variable called peter used as a pointer for/to floats

in Pascal I think it's

var
 peter: ^float  different syntax, same effect

though you are obviously asking about something like C Pointers in this context are always pointing to memory locations so they are always integer types themselves - probably 32 or 64 bit. If you wanted the pointer itself to be a non-integer datatype that's a different question, and an odd thing to do.83.100.250.79 (talk) 01:03, 17 July 2009 (UTC)

HELP NEEDED

Hi All, the other day i inserted a flash drive in to my USB port and since i have open the drive, i can't see my "FOLDER OPTIONS" in "TOOLS" menu... and moreover i also can't use the "SEARCH" feature in "START MENU"...

Can anyone let me know how to overcome this problem?

Does this look like what is happening to you? Google is our friend. Kushal (talk) 15:14, 16 July 2009 (UTC)

autorun.exe

How do i see the autorun.exe of any flash disk/pen drive??? —Preceding unsigned comment added by Sushil shenoy (talk • contribs) 15:19, 16 July 2009 (UTC)

The autorun.exe file is usually marked as "hidden", so the easiest way to do this is to tell Windows to "show hidden files". In Windows XP, do Control Panel->Folder Options, and then I think it's on the "Advanced" screen. In Vista, open Control Panel and then type "folders" in the search field, then choose "Show hidden files and folders". After this is turned on, the autorun.exe on a flash drive will be visible (and easily deletable, which I assume is what you're aiming for). Tempshill (talk) 15:58, 16 July 2009 (UTC)

exactly ..... i hate autorun.exe! slows down everything !!! and at times can get infected also !!

You mean autorun.inf I think. 68.123.28.58 (talk) 05:50, 19 July 2009 (UTC)

COBOL Help

Hello. I installed OpenCobol, which uses Visual C++'s cl command to make binaries from COBOL source files. So, I made a COBOL hello-world program (TEST.COB), and compiled it, but it just creates (1) test.dll, (2) test.dll.manifest, (3) test.exp, and (4) test.lib. Anyone know how to make a .exe or at least use those files to output the message to standard out? I tried running the .dll using rundll32, but nothing happened. Here's the source for TEST.COB:

000100 IDENTIFICATION DIVISION.
000200 PROGRAM-ID. "TEST".
000300 ENVIRONMENT DIVISION.
000400 DATA DIVISION.
000500 PROCEDURE DIVISION.
000600 DISPLAY "HELLO".
000700 STOP RUN.

Thanks for any help.--H. Gutmans (talk) 18:53, 16 July 2009 (UTC)

Nevermind. Sorry. I needed to add a -x to the cobc command to get it to output a .exe.--H. Gutmans (talk) 19:08, 16 July 2009 (UTC)

Who am I connected to?

Is there any way of checking that I am connected (broadband) to the ISP I expect, and that my broadband connection has not been hijacked and put through some expensive thing? I know my anti-virus software will check for diallers, but is there any more direct way of checking please? 78.145.23.157 (talk) 19:40, 16 July 2009 (UTC)

If the hijacker is good enough, it is impossible to tell. The hijacker will inspect and fake all traffic so even something like traceroute will show you incorrect information to make you think you are on your network. It really isn't worth it do go to that much trouble for some random Internet user. So, just run traceroute on something like Google. You will see each IP address between you and Google. You can check to ensure that the ones closest to you are your ISP. -- kainaw™ 19:48, 16 July 2009 (UTC)

Dialers affect users using dial-up, not those on broadband.--Xp54321 ^{(Hello! • Contribs)} 20:54, 16 July 2009 (UTC)

Yes, I think the OP is asking: When he connects to the Internet via his laptop or iPhone or whatever, might he be connecting to the Internet via some expensive ISP that charges him a lot of money? Offhand I can't think of any way this could happen, because unlike with a telephone line provided by your local phone company, you have not authorized your ISP to charge you for services that are performed by other companies. At least, I haven't ever heard of such a thing. Tempshill (talk) 21:07, 16 July 2009 (UTC)

Mathematica 6.0

Does anybody know what algorithm Mathematica 6.0 (student edition) uses when the Sort[list] function is called? --72.197.202.36 (talk) 20:34, 16 July 2009 (UTC)

It seems to be Quicksort which we also have an article about.-KoolerStill (talk) 23:16, 16 July 2009 (UTC)

How to compare two folders?

What would be the easiest way of comparing two folders and checking that their subfolders and the names of the files within them are the same please? I just want to confirm they are the same, not move files around. The folders may be on different drives. I am using Windows XP. Thanks 78.145.23.157 (talk) 21:29, 16 July 2009 (UTC)

command prompt: tree C:\path\to\folder1

command prompt: tree C:\path\to\folder2

put the windows side by side

Ivtv (talk) 22:10, 16 July 2009 (UTC)

Even better!

CMD: tree c:\path\to\folder1 > c:\folder1tree.txt

CMD: tree C:\path\to\folder2 > c:\folder2tree.txt

CMD: fc c:\folder1tree.txt c:\folder2tree.txt

(If you don't know what '>' does, it redirects the output into a file of name some (path and) file)

This should work pretty handily without any extra software on I believe Win2k+ Washii (talk) 10:34, 18 July 2009 (UTC)

Get the free winmerge.exe (http://winmerge.org/). Wonderful tool! Saintrain (talk) 00:55, 17 July 2009 (UTC)

I agree! We do have an article on WinMerge though. Jay (talk) 09:37, 17 July 2009 (UTC)

Security

http://news.runescape.com/newsitem.ws?id=2003

How secure is this, really? Someone was claiming people would "work around it", but if you had such a key I can think of no way, beyond physical theft, to compromise the security of someone else's account. Or maybe I'm misunderstanding it entirely. Is it possible to hack someone's RuneScape account if they've got one of these keys? Vimescarrot (talk) 21:47, 16 July 2009 (UTC)

The security token article may be of interest to you. Tempshill (talk) 22:13, 16 July 2009 (UTC)

I cannot access this page, but I am assuming that you order a physical device and through a random number generator every 10 seconds it matches the same number generator to your account. World of Warcraft does the same thing. if you implement this to your account it is next-to-impossible to crack it. each number generator is unique Ivtv (talk) 22:21, 16 July 2009 (UTC)

How to identify a running application in Ubuntu?

Say I have a GUI application running on Ubuntu, for example the tracker applet or the sound recorder or the screenshot application and I'd like to report a bug against it. The command ubuntu-bug <packagename> would take care of that but how would I know the name of the package in the first place? (This is also very useful if you would want to use killall <NameOfApplication> or simply want to know how to start the application with parameters from the commandline.)

Does it boil down to using top or htop to look through possible application names or asking at the forums or is there a smarter, better way? Any help will be greatly appreciated. Thanks. Jarl Arntzen (talk) 23:15, 16 July 2009 (UTC)

Yeah, I usually use top or ps auxf. The tree organization of ps auxf sometimes help find the program, and if you just started the program, it's often the last item in the ps list. There must be some better solution, though. Indeterminate (talk) 23:22, 16 July 2009 (UTC)

Thanks! That's certainly one possible way. Have anyone else some other, good technique to suggest? Jarl Arntzen (talk) 23:34, 16 July 2009 (UTC)

Many window managers have an "identify window" widget. If yours doesn't, the following Perl script will change the cursor to a crosshair, you click on a window, and it will tell you the command line of the process that opened that window:

#!/usr/bin/perl -w

`xwininfo` =~ /Window id:\s*(\S+)/ or die "could not find window";
$id = $1;

`xlsclients -l` =~ /^Window\s*$id:.*?Command:\s*([^\n]+)/sm
    or die "could not find client";

print "$1\n";

--Sean 14:28, 17 July 2009 (UTC)

JSON texts versus values

RFC 4627 says that "A JSON text is a serialized object or array.", and that "A JSON value MUST be an object, array, number, or string, or [one of 'false'/'true'/'null'.]" What reason is there for requiring that the top level construct be an aggregate? My impulse would have been to make the text be a value so that there was only one concept to address. Is it perhaps related to XML's restriction that the top level be a (single) tag? --Tardis (talk) 23:15, 16 July 2009 (UTC)

We may have to wait for The Design and Evolution of JSON, but I've also found that weird. --Sean 22:58, 17 July 2009 (UTC)

What's wrong with this javascript widget?

I have User:Henrik/sandbox/google-search installed in my monobook.js. It's supposed to put a google search box in the left menu column, which indeed it has done for ages. Today I found no such search box & am v.depressed since I make major use of google searches of wikipedia. Does anyone have a clue what's going wrong? thanks --Tagishsimon (talk) 23:20, 16 July 2009 (UTC)

WP:VPT would be a better place for this question. Algebraist 23:22, 16 July 2009 (UTC)

I can't see anything wrong with it, and it works fine here. Have you tried bypassing your cache? Algebraist 23:25, 16 July 2009 (UTC)

Thanks. It was a cache problem, or, was solved by clearing the cache. --Tagishsimon (talk) 23:28, 16 July 2009 (UTC)

Windows needs disc to boot

My PC with a Gigabyte motherboard and Windows Vista has an odd problem. It will not boot unless the Vista DVD is in my DVD drive. It then does not boot *off* the DVD — it just seems to need the DVD to get started, and then it proceeds to boot from the hard disk. I may have faulty recall here, but I think a similar problem occurred before I had installed the current hard disk and installed Vista — the old hard disk, with Windows 2000 on it, required a boot disc in the DVD drive in order to proceed to boot from the hard disk. Changing the boot device order in the BIOS does not seem to affect this issue, other than when the DVD drive is set to be the boot device, it boots from the DVD. Any suggestions? Tempshill (talk) 23:52, 16 July 2009 (UTC)

this does not seem to make much sense at all. Try resetting your bios to default, then boot to windows and update your bios. did you do a clean install of vista? Ivtv (talk) 00:00, 17 July 2009 (UTC)

This sounds contradictory (though I'm guessing you're pretty sure you're accurately describing the symptoms). Are you very certain that the system is not booting off the DVD? Maybe there is a broken bootloader; can you try booting off the DVD and reinstalling the boot-loader (or using a Linux CD to install GRUB)? Nimur (talk) 00:23, 17 July 2009 (UTC)

Some more details: I'm very certain it's not booting off the DVD, which runs an installer. Yes, this was a clean install of Vista. Each time I boot the system, after the disks are listed and the memory check is done, it seems to hit the hard disk for a long time (during which the LED seems lit at half-intensity, telling me it's flickering fast), and then an error message appears saying, "DISK BOOT FAILURE, INSERT SYSTEM DISK AND PRESS ENTER". I have now clumsily tried installing GRUB from a Knoppix CD (I fumble with Linux) and after mounting the C: drive into /mnt/temp, and trying "grub-install --root-directory=/mnt/temp /dev/sda", it hit the hard disk for a while and then said "The file /mnt/temp/boot/grub/stage1 not read correctly." I also have booted the Vista install CD and tried installing the Windows Boot Manager, and got a file error that I didn't write down. I haven't tried re-flashing the BIOS; I guess I'll attempt that (and then reinstalling WBM) next. Tempshill (talk) 04:40, 17 July 2009 (UTC)

Try shuffling your boot order. Sometimes if you have the optical drive as the first boot device the BIOS can be stupid and stop at that when it detects no disc in the drive. Set it to hard drive and make sure the drive priority is correct. --antilived^{T | C | G} 04:59, 17 July 2009 (UTC)

I found that my BIOS is a couple of revisions old, but the only BIOS flash utility from Gigabyte I was able to find on their website is for 32-bit Vista only (yay Gigabyte), and from the descriptions, it appears that the changes are some CPU ID additions and one added feature. I checked my boot order; it's HD first (always was) and I verified the drive priority is correct (SATA HDD, then a PATA HDD master). I stared at the drive light while booting, this time, and logged the following bizarre behavior, which does occur with 100% reproducibility:

1. Memory test completes on the BIOS screen
2. Drive indicator light turns "half-intensity" (which I interpret to mean it's getting accessed repeatedly but not continuously) for 25 seconds
3. The BIOS lists all the connected drives
4. The drive indicator light turns "half intensity" for 35 seconds
5. The BIOS lists all the PCI devices then says "Verifying DMI pool data"
6. Drive light is off for a couple of seconds
7. Drive light turns on, half-intensity for an amazing 122 seconds
8. Text appears saying "Press any key to boot from CD/DVD..." At this point I leave it alone and it boots Vista off the hard disk.

So, yes, it takes me 3 minutes to boot my computer, and nearly all of that time is spent accessing the hard disks, if my drive indicator light isn't lying. Could this be the root of the trouble? Tempshill (talk) 07:49, 17 July 2009 (UTC)

Strange, indeed. (1) nothing to do with the BIOS version, if it ever used to work at all (2) looks like it is searching madly for a boot sector, as indicated by the "disk boot failure" error message. Having finally found the boot instructions on the DVD, it then proceeds to load these into memory and run the rest of the OS from the hard drive.

It sounds like the boot sector is corrupted. So stranger still if it did, indeed, do the same with another hard drive and OS. If it's done it with more than one disk, it is more likely a kinked or cracked ribbon cable to the drive. Triple check, flatten out and reinsert the cable, ideally replacing it with a new one. (I've had cable problems create Disk Boot failure errors with no other trouble once the HDD was running).

The DVD should have a Repair section on it. Fire that up and go to the Booting section, which should allow you to "fixmbr" (rewrite the master boot record).This will overwrite the Grub, which does not sound a very successful installation either. Or get a hand from this BCD repair programwhich works without the DVD. Leave the BIOS set up to start from the HDD. - KoolerStill (talk) 12:40, 17 July 2009 (UTC)

OK - I tried the Vista disc's "Repair -> Startup" process, and it went through a diagnostic and said that no problems were detected. Following a Googled tip, I did bootsect.exe /nt60 all from a command prompt after booting from the DVD; bootsect went through and stated for both my C: and D: hard disks that it had "Successfully updated NTFS filesystem bootcode." This didn't lead to any change in my 3 minute boot times or the need for the Vista disc in the drive. Thanks for the tip on the cables — I guess I'll take a look this weekend. I would have guessed that a cabling problem would cause ongoing HD access failures and wouldn't have expected problems only at boot. Tempshill (talk) 16:18, 17 July 2009 (UTC)

It's possibly a more insidious failure, e.g. failing hard-disk that is intermittently flaking out. By the time the OS boots, error-correction can make this seem transparent or nonexistent. It's also possible that there is a power brownout in effect - your hard-disk may be right on the edge of consuming more power than your PSU can supply (again, intermittently failing). I would treat this system very cautiously and back up all important data; hopefully a fix can be found, but these symptoms may indicate a more severe crash is imminent. Nimur (talk) 20:54, 17 July 2009 (UTC)

It's weeks later and this is in the archives now, but in case any person of the future finds this thread: I "fixed" the problem. My boot drive was a SATA drive and the second hard disk (used for backup) was an ATA drive. I disconnected the ATA drive from the system and now the computer marches quickly from the BIOS screen through the device ID screen and then boots normally off the SATA drive, and the Windows disc is no longer needed. I have no explanation for the malfunction and don't know whether it's the ATA drive itself or the ATA cable. Thanks Nimur for pointing to physical problems. Tempshill (talk) 00:34, 10 August 2009 (UTC)