Jump to content

Wikipedia:Reference desk/Archives/Computing/2010 April 8

From Wikipedia, the free encyclopedia
Computing desk
< April 7 << Mar | April | May >> April 9 >
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.


April 8

[edit]

USB 3.0

[edit]

Is USB 3.0 essentally a parallel version of the USB 2.0 with three sets of data cables to allow two way communcation between devices plus backwards compatiblity with 2.0 devices? —Preceding unsigned comment added by 74.67.89.61 (talk) 02:45, 8 April 2010 (UTC)[reply]

Does USB 3.0 answer any of this ? StuRat (talk) 02:49, 8 April 2010 (UTC)[reply]
It's not parallel communication as that term is normally used. -- Coneslayer (talk) 12:25, 8 April 2010 (UTC)[reply]

Signal a program when Internet connection made

[edit]

In Kubuntu, is there a way to send a signal to a program whenever an Internet connection is made? I'd like to be able to mirror a local folder onto a remote folder (one-way) via SSH, with updates taking effect as fast as possible; I figure the way to do this is to write a daemon that uses inotify to monitor the folder, waits for a connection if I don't already have one (or if the one I'm on didn't work the last time the daemon tried it), and then invokes rsync. I have both wired and wireless interfaces (eth0 and wlan0) and am usually using only one of them at a time. NeonMerlin 03:23, 8 April 2010 (UTC)[reply]

As a Debian-derived system, I'll guess that it has an /etc/network/if-up.d directory. Put a script in there and it'll run every time an interface is brought up. man 5 interfaces for details 98.226.122.10 (talk) 07:45, 8 April 2010 (UTC)[reply]
Yes, that's certainly the case for Ubuntu. As wireless connections are almost always managed by NetworkManager, it runs /etc/NetworkManager/dispatcher.d/01ifupdown when an interface goes up or down. That script in turn runs the contents of /etc/network/if-up.d and /etc/network/if-down.d respectively. -- Finlay McWalterTalk 09:23, 8 April 2010 (UTC)[reply]

do Huawei wireless broadband dongles support voice?

[edit]

The wireless broadband dongles work on the 3G mobile phone/data network, so I was wondering. If they do support voice do you know if they support voice on common linux distributions? --Polysylabic Pseudonym (talk) 08:37, 8 April 2010 (UTC)[reply]

The one I have in the UK (Vodaphone) is a purely data device with no automatic voice support. You may find it possible to use a VoIP provider, such as Skype, but the quality may not be high. --Phil Holmes (talk) 15:42, 8 April 2010 (UTC)[reply]

How to merge bookmarks files in Firefox?

[edit]

I have a short bookmark .json file that I would like to merge with my existing Firefox bookmarks. I arrange my bookmarks in folders, and the names of some of the folders could be the same. What would be the best way to merge them please? 92.29.42.231 (talk) 11:40, 8 April 2010 (UTC)[reply]

.json is the serialization for JavaScript objects, right? Where did you get it from? Given that Firefox stores in plain HTML last I checked, the merging is probably non-trivial. As a cheat, you might try importing them into another browser on your machine, then use Firefox's native File->Import option, but it probably wouldn't sync up exactly the way you want. —ShadowRanger (talk|stalk) 18:33, 8 April 2010 (UTC)[reply]
Firefox Bookmarks are actually stored in a SQLite DB. (since Firefox 3) It can be set to output a backup html file to the profile, but if I remember correctly this isn't the default behavior and a setting needs to be changed. In about:config add browser.bookmarks.autoExportHTML and set it to true, it will save the bookmarks on exit, this can dramatically slow shutdown times. The file should be in your profile folder. There are ways to merge them, but the simplest might be to use the Weave extension.
http://kb.mozillazine.org/Import_bookmarks
http://labs.mozilla.com/
http://kb.mozillazine.org/Profile_folder
209.226.104.91 (talk) 18:04, 9 April 2010 (UTC)[reply]

.json was what I got from Firefox regarding saving the bookmarks - I cannot remember if I asked it to export the bookmarks, or if I found the files it uses and copied them. According to the json article, it is meant to be human readable, so if I could find an editor that could display them in a clear layout rather than just as a continuous long string, then I could try merging them manually. 92.28.237.243 (talk) 19:15, 8 April 2010 (UTC)[reply]

Have you tried opening the file in Firefox itself? It can display stuff like XML in a well-structured form. —ShadowRanger (talk|stalk) 19:27, 8 April 2010 (UTC)[reply]

Firefox does not load websites after computer left unattended for a long time

[edit]

After leaving my WinXP computer on but unattended for say a hour or so, when I came back Firefox will not load any webpages. After a long pause it says something like "website unavailable". Even when I click the Try Again button, same result, although instantly and without any delay this time. This is even with dependable websites like Google or the BBC news. However Internet Explorer has no problems displaying there same websites in the same circumstances. After IE has made a connection, then Firefox makes a connection easily as normal. Firefox has no problems loading webpages when doing so soon after the computer has been switched on and started.

What could be the problem and how could I fix it please? 92.29.42.231 (talk) 11:55, 8 April 2010 (UTC)[reply]

Firefox is known for consuming large amounts of RAM, especially when its been running for a long time. In newer versions this is less of a problem but still exists somewhat. I suggest when you're going to leave your computer for a long time to exit firefox. 82.44.54.207 (talk) 15:14, 8 April 2010 (UTC)[reply]
Are you using an internet connection that is timing out for some reason (e.g. a dial up connection)? IE has a setting that allows it to recreate the connection - this may be what's happening. --Phil Holmes (talk) 15:40, 8 April 2010 (UTC)[reply]

I do exit Firefox when leaving the computer, at least I close the Firefox window which I assume is exiting Firefox. I have this problem with Firefox 3.6.3 and earlier versions. My connection is broadband. 92.29.42.231 (talk) 17:04, 8 April 2010 (UTC)[reply]

Assuming Firefox no longer has an entry in the taskbar, then yeah, you exited it. What extensions, plugins and themes do you have installed? —ShadowRanger (talk|stalk) 18:34, 8 April 2010 (UTC)[reply]
Also, are you on a corporate network? Behind a firewall, NAT, etc.? —ShadowRanger (talk|stalk) 18:45, 8 April 2010 (UTC)[reply]

Extensions: 1-Click YouTube Video Downloader, Adblock Plus, CacheViewer, deskcut, Flagfox, Image Zoom, TinEye. Themes: default theme. Plugins: Adobe Acrobat, Java Platform SE6 U13, Java Platform SE 6 U18, Microsoft DRM, Mozilla Default Plugin, Shockwave Flash, Windows Media Player Plugin Dynamic Link Library.

I see I have two Java things - do not know if that is an error or not. I am not on a corporate network. I use the default Windows XP firewall. More specifically, Firefox says something like 'cannot find server' when I have this problem. Thanks 92.28.237.243 (talk) 19:37, 8 April 2010 (UTC)[reply]

I doubt the multiple Java copies are the cause (though you may want to uninstall all your Java stuff, clean out the plugin(s) manually if necessary, and reinstall the latest version just to be sure). I'm not familiar with most of your extensions. Adblock Plus should be fine, but you might try disabling the rest of them (remember to restart the browser after doing so) and seeing if the problem recurs (this assumes the recurrence is somewhat predictable). If it doesn't, re-enable plugins one by one and test until you find the culprit. —ShadowRanger (talk|stalk) 19:50, 8 April 2010 (UTC)[reply]

Logical, but since I'd have to wait an hour between disabling, even doing half of them at a time, then I'm probably just going to live with it. 78.146.107.183 (talk) 20:05, 12 April 2010 (UTC)[reply]

atanfull

[edit]

This post waxes lyrical about atanfull, apparently a trigonometric function. From context it sounds like he's talking about atan2, but I can't find a reliable source to confirm this (if I could, I'd make a redirect from atanfull->atan2). Searching confirms that atanfull is implemented in at least PlayBASIC and DarkBASIC. Can anyone find a reliable source that shows atanfull is indeed just atan2? -- Finlay McWalterTalk 12:31, 8 April 2010 (UTC)[reply]

googling found this (p. 56). It looks like atanfull is atan2 with a result in degrees instead of radians. It appears peculiar to "darkbasic", a lame looking Basic dialect from some game company. I personally wouldn't bother with the redirect, but whatever. 66.127.52.47 (talk) 23:58, 8 April 2010 (UTC)[reply]

Quantum Computer Data Transfers

[edit]

Lets say in the near future quantum computers were common, laptops were quantum computers, even cellphones had quantum pocessors. How would the qubits be transfered in between them? Could a version of modern connecton cables be used like USB, or would it require a completely new way of connectng devices, a complete redisign of the cables used? —Preceding unsigned comment added by 74.67.89.61 (talk) 13:01, 8 April 2010 (UTC)[reply]

Quantum cryptography discusses this question. --Sean 15:17, 8 April 2010 (UTC)[reply]
Err, that's a very different thing. Quantum computers and quantum cryptography are not the same thing at all. --Mr.98 (talk) 15:27, 8 April 2010 (UTC)[reply]
The linked article is all about quantum networking, as is the OP's question. I don't get your point. --Sean 13:58, 9 April 2010 (UTC)[reply]
I'm a little confused about what you're asking. In modern computers the processor states are never transferred. All that is transferred are memory states (e.g. a file on your hard drive). That's not the same thing. Presumably to transfer quantum computer data it would be translated into standard electrical signals and then re-entered as quantum data, the same way you transfer, say, information stored on a flash memory drive. The actual medium of how the memory is stored does not really matter in such situations. You are not transferring the cores of one computer physically into the cores of another. (Addendum: quantum computers will probably not become common in the near future. Even if we did advance quantum computing by leaps and bounds in the short term—which would be somewhat miraculous by itself—it's not clear that they would have an advantage over silicon processors in most computing situations.) --Mr.98 (talk) 15:27, 8 April 2010 (UTC)[reply]
Yes but wouldn't a quantum computer use qubits to transfer information and process it just like a conventional computer uses bits to process and transfer information? My question is would current cables/wireless data transfer methods be cabable of transfering qubits (because they use 1s and 0s like conventional bits but also use a superpositon of inbetween) or would we need to go back to the drawing board and completely redesign our data transfer methods? —Preceding unsigned comment added by 74.67.89.61 (talk) 16:11, 8 April 2010 (UTC)[reply]
Bits aren't "things", they're just a unit of information (like saying an inch is just a unit of length). The core of a quantum computer uses qubit logic, but the stored memory is almost certainly going to be some conventional form of 1s and 0s (which can be used to handle all types of states). So when you are running a quantum computer, you would set up the qubit states, run the quantum algorithm (which lets you do things you can't do as easily with a standard computer), and then get the results and put them back into stored memory. You don't store the qubits states as qubits—you can't, I don't think, as the act of reading them destroys the quantum states. (That is, separately, why quantum cryptography should work—it guarantees that certain information physically has not been tampered with, because looking at the information modifies the states.) --Mr.98 (talk) 16:19, 8 April 2010 (UTC)[reply]
Why would you change them back to conventonal bits and not leave them as qubits if talking between quantum devices? Thats like when digital computers came out having the devices compute the data using digital bits, then converting them to analog signals to communicate between devices only to change them back to digital to continue computing. If we use bits to transfer and compute information in and between didgital devices, why wouldn't we compute and transfer data using qubits in and between quantum devices instead of wasting time converting them to bits. —Preceding unsigned comment added by 74.67.89.61 (talk) 18:03, 10 April 2010 (UTC)[reply]
Storing qubits in a conventional storage medium is equivalent to measuring them. If you're doing a large quantum computation, all involved storage devices have to be quantum. If it swaps to disk, the disk has to be quantum. If it's a distributed computing project, the network has to be quantum (unless you can divide the problem into small quantum pieces with a classical interface between them). -- BenRG (talk) 19:12, 8 April 2010 (UTC)[reply]
BenRG's statement is valid - and expresses the issue with quantum computing. An earlier question about quantum computing brought up some points with regard to designing algorithms. Quantum computers force us to redefine "algorithm" in such a way that we can express, represent, and compute a program using a quantum physical system. In the same way, they could conceivably require us to redefine "storage". The act of storing or transferring a known set of information is a classical representation. A file is the same set of bits every time you view it, and when you modify it, the results are deterministic. If we were to redefine a "quantum computer storage system", those fundamental paradigms would not apply anymore, and the concept of stored data or data transfer would also need to change. Nimur (talk) 20:07, 8 April 2010 (UTC)[reply]
It's easy to define how a quantum storage device should behave. They're just hard to construct, because of the no-cloning theorem. When you read a qubit from the storage device, it has to be perfectly erased from the device as though it had never been there (failure to do so is equivalent to measuring it). That means, in particular, that storage and retrieval have to be thermodynamically reversible processes. The same constraint applies to every piece of hardware that's involved in the computation.
In the earlier thread you say "The problem of quantum computing is that we do not know a way to map boolean algebra into input/output states for quantum mechanical entanglement." I don't understand what you mean by that. Certainly it's known how to run ordinary (classical) algorithms on a quantum computer. A big part of Shor's factorization algorithm involves applying a classical modular exponentiation algorithm to the qubits. Since the set of classical computable problems is the same as the set of quantum computable problems, we know how to solve every solvable problem on a quantum computer. The difficulty is just speed. Solving problems on quantum computers using the known classical algorithms wouldn't save any time. People are looking for quantum algorithms with better big-O complexity than the best known classical algorithms. That's hard, but so is finding fast classical algorithms. -- BenRG (talk) 03:30, 9 April 2010 (UTC)[reply]
Right - the speed boost comes from the use of quantum phenomena in order to generate the same calculation results that a classical algorithm would. To my knowledge, there is not a general mapping between boolean algebra and quantum qubit state changes. If you have a set of logical AND and OR gates, it's trivial to wire them up classically - but it's not generally true that we can come up with an equivalent quantum system that produces the same outputs. Nimur (talk) 12:06, 9 April 2010 (UTC)[reply]
The classical reversible gates are a subset of the quantum gates, namely those whose matrix representations are permutation matrices (a subset of the unitary matrices). The Toffoli gate, for example, is universal for classical reversible computation and is also a quantum gate. Any formula in boolean algebra can be computed by a circuit containing only Toffoli gates, which is a quantum circuit. -- BenRG (talk) 07:21, 10 April 2010 (UTC)[reply]
Qubits can't be sent over ordinary wires because they leak information into the environment. It might be possible over superconducting wires, I don't know, but that's wildly impractical without a room-temperature superconductor. (Of course, everything about quantum computers is wildly impractical.) More realistically, you could use fiber-optic cables. The cables could look like USB from the outside, but they couldn't actually be USB. -- BenRG (talk) 19:12, 8 April 2010 (UTC)[reply]
Well, superconducting supercomputers aren't completely out of the question. If you mean home computers, then you may have a point. StuRat (talk) 04:52, 9 April 2010 (UTC)[reply]

gmail vs yahoo mail

[edit]

How come to log in to gmail you just have to type your "Username" (without the "@gmail.com"), but to log in to yahoo mail you have to type in your "Yahoo ID" plus "@yahoo.com"?

It is, of course, easier to log in to and check your email at gmail simply because you don't have to type in "@gmail.com".

Did yahoo not think of this? Or is there some other reason? Bus stop (talk) 15:57, 8 April 2010 (UTC)[reply]

Probably just a matter "let's do it this way". If you want to rationalize it, the Yahoo solution may also strengthen brand loyalty. And Yahoo can more easily merge in other systems, as there is an implicit name space if you use the domain name. --Stephan Schulz (talk) 16:07, 8 April 2010 (UTC)[reply]
Google usernames are irrelevant of the domain name (e.g. gmail.com or googlemail.com) and this is good because you as you've said you can log in faster, but bad because it eats up the available names at a much faster rate (if someone registers myemail@gmail.com, it's not possible for someone else to register myemail@googlemail.com). Yahoo also has multiple domains (yahoo.com as well as yahoo.co.uk, yahoo.ca, ymail.com... etc), but they allow you to create a username per domain. So someone could create myemail@yahoo.com and someone else could create myemail@yahoo.co.uk. When you login they have to ask the full email address so they know which mailbox you're trying to log into. Hope this helps! ZX81 talk 16:14, 8 April 2010 (UTC)[reply]
(ec) Yahoo! Mail is more than the @yahoo.com domain, I think. See Yahoo!_Mail#Ymail_and_Rocketmail. Presumably you can use @ymail and @rocketmail logins for the same interface? --Mr.98 (talk) 16:15, 8 April 2010 (UTC)[reply]
It would be simple to program the site to tack on "@yahoo.com", only if the "@" sign is absent, and use the full address as is, otherwise. Alternatively, they could have a button after the base address that defaults to "@yahoo.com", but can be selected to produce a pull-down menu with all the supported domain names. StuRat (talk) 16:20, 8 April 2010 (UTC)[reply]
I somewhat agree. I think drop-down menus are a huge pain, though. But there is probably a better way for them to do this. --Mr.98 (talk) 16:25, 8 April 2010 (UTC)[reply]
While a pull-down menu is best for long lists of choices, for a small list, maybe 3 or 4, you could just list each, with a radio button in front of them. StuRat (talk) 16:31, 8 April 2010 (UTC)[reply]
Thanks everybody. Well, I have a newfound appreciation for Yahoo's way of doing things. Bus stop (talk) 16:56, 8 April 2010 (UTC)[reply]

I don't ever have to type in the "@yahoo" when I log into Yahoo mail. Perhaps I have bookmarked a secondary sign in screen? 10draftsdeep (talk) 18:20, 8 April 2010 (UTC)[reply]

If your Yahoo e-mail has @yahoo.com on the end then you can leave the @yahoo.com off and the log in will work. If your Yahoo e-mail has some other domain on the end, then you have to log in with the complete e-mail address. If you're signed out and go to mail.yahoo.com, the sign in box shows an example log in with @yahoo.com on the end, but this is likely to help remind users whose e-mail addresses have non-yahoo.com domains on the end that they need to log in with their complete e-mail address. Next time you log in, ignore the example and see if it works with just your username. It works for me and for 10draftsdeep. --Bavi H (talk) 01:11, 9 April 2010 (UTC)[reply]

Enforcing strong passwords in Active Directory

[edit]

In Active Directory I can require some complexity for passwords. If I feel that this is to lax what do I do? Ie. how do I prevent "H3lloWorld" from beeing a legal password? Some protection can be given by enforcing password length, but this I feel is somewhat annoyoing and it does not prevent "H3lloWorld1234567890", or similiar use of padding to get the required length. 213.161.190.227 (talk) 17:11, 8 April 2010 (UTC)[reply]

Not a direct answer, but I would warn against overly strict rules for passwords. When I worked at EDS, they had such strict rules that nobody could remember their password, leading everyone to write them down on Post-It Notes and stick them on the side of the computer. Obviously, the EDS security policy was self-defeating. StuRat (talk) 17:15, 8 April 2010 (UTC)[reply]
Well, writing passwords on post-it notes at least doesn't expose them to internet attacks. But yeah, going crazy with password policy isn't worthwhile. The 1980's-era Rainbow Series (I forget which volume) of DoD security standards said passwords should be assigned by the security officer (using a random generation scheme) rather being selected by the user, which would help with the complexity issue. These days though if you're doing something serious, the preference is for multi-factor authentication, so you'd provide each user with a hardware token like a SecureID (or one of its far cheaper competitors available these days) and not rely purely on passwords for authentication. 66.127.52.47 (talk) 17:33, 8 April 2010 (UTC)[reply]
That explains my first experience with the password on a DoD PC in 1986. ---— Gadget850 (Ed) talk 23:06, 8 April 2010 (UTC)[reply]

Consequences of misusing Mac OS X Terminal

[edit]

Hi, just wondering what would be the most major effect of misusing the Mac OS X Terminal program. Would my computer fail to start up completely? Would certain programs not work? How powerful is the program potentially? Chevymontecarlo. 20:41, 8 April 2010 (UTC)[reply]

I believe it has the same power as any other *NIX style CLI. Which means if you misuse it badly, you can delete the OS out from under yourself. —ShadowRanger (talk|stalk) 21:27, 8 April 2010 (UTC)[reply]
This statement is a little harmful:
sudo rm / -r -f
--Chmod 777 (talk) 21:43, 8 April 2010 (UTC)[reply]
Wikipedia:Don't stuff beans up your nose, please. Consider removing or explaining in more detail the danger of such a command... --Mr.98 (talk) 22:25, 8 April 2010 (UTC)[reply]
The above command will permanently remove all contents of a hard-disk all hard-disks and media mounted on a computer. It will also execute without any request for confirmation (especially if the user has already used sudo). Nimur (talk) 21:29, 9 April 2010 (UTC)[reply]
Yeah. That's the one I was talking about. Don't do that! —ShadowRanger (talk|stalk) 21:45, 8 April 2010 (UTC)[reply]
(edit conflict) Oh, yes. For example, the rm command can quite easily be instructed to erase your entire hard drive. The command line offers you basically complete control over your computer (like Finder, but more flexible). Basically, you should know the meaning of commands. rm deletes stuff -- it's dangerous. ls lists files -- all it does is give you information, so it's safe.
The other important thing that you should understand is the layout of your filesystem. Everything under /Users/yourusername is your files. You could delete it, and your computer will start up just fine (and generally speaking, programs will still work), but your files would be gone. If you delete something else (like, say, the system utilities in /usr/bin), your computer will be unbootable, but your files would be recoverable (given some technical know-how). Paul (Stansifer) 22:23, 8 April 2010 (UTC)[reply]
Also, there is nothing magical about the CLI. It's fairly hard to inadvertently damage your MacOS-X installation with the command line (it's easier to damage your personal files, but still not happening automatically), and you can just as well ruin your computer using another tool. --Stephan Schulz (talk) 23:07, 8 April 2010 (UTC)[reply]
There is nothing magical about prompts in general but it is easy to do something wrong with them. A number of years ago I totally fubared a Windows box I had by running a batch file that didn't quite change directories correctly and instead of erasing a bunch of temp files, erased my Windows directory. When you are doing things with powerful tools and no GUI, all sorts of mayhem is easily possible... it's easier to accidentally do something with dire consequences in the shell than it is in the GUI (especially since on Macs, a huge amount of the Darwin/*nix stuff is actually hidden from the GUI by default). --Mr.98 (talk) 01:54, 9 April 2010 (UTC)[reply]
How many computers have been ruined by a UNIX CLI prompt (and by ordinary users, at that), and how many by clicking on the icon of a downloaded virus? --Stephan Schulz (talk) 07:41, 9 April 2010 (UTC)[reply]
Just because viruses are more common doesn't mean that mucking around in a command prompt (UNIX, DOS, whichever) is safe. Malaria is more common than Ebola, but that doesn't make the latter anything to sniff at. --Mr.98 (talk) 13:49, 9 April 2010 (UTC)[reply]
Both those examples suffer the same symptom - they do not provide feedback to the user about what actions (harmful or otherwise) that the program is doing. This is a matter of user-interface design, not "command-line text-only" or "icon / window". The command rm -i forces an interactive mode - it's very irritating, but it asks for confirmation. Similarly, a proper GUI program will inform the user about any modifications it is making. This design guideline isn't a solution for malicious and misleading software, but it prevents unintentional modifications. Nimur (talk) 10:10, 9 April 2010 (UTC)[reply]

iPhone question...

[edit]

I just jailbroke my iPhone 3G. I also just found out about iPhone OS 4.0, which makes me wonder, "Am i eligible for an upgrade once the next iPhone comes out, since I jailbroke it? I know my warranty's voided, but am i still eligible?" Thanks! —Preceding unsigned comment added by 24.243.18.28 (talk) 21:21, 8 April 2010 (UTC)[reply]

Probably not. That said, OS 4.0 was only demo-ed, I don't think they've published anything definitive. Maybe this time Apple will say yes to jailbreakers... Okay, probably not. —ShadowRanger (talk|stalk) 21:25, 8 April 2010 (UTC)[reply]
The article Jailbreaking for iPhone OS may help. Apple consider jailbreaking to be a violation of copyright (and by implication prosecutable under the DMCA). Cuddlyable3 (talk) 23:01, 8 April 2010 (UTC)[reply]