Wikipedia:Reference desk/Computing

Welcome to the computing section
of the Wikipedia reference desk.

skip to bottom

Select a section:

• Computing
• Entertainment
• Humanities
• Language
• Mathematics
• Science
• Miscellaneous
• Archives

Want a faster answer?

Main page: Help searching Wikipedia

   

How can I get my question answered?

• Select the section of the desk that best fits the general topic of your question (see the navigation column to the right).
• Post your question to only one section, providing a short header that gives the topic of your question.
• Type '~~~~' (that is, four tilde characters) at the end – this signs and dates your contribution so we know who wrote what and when.
• Don't post personal contact information – it will be removed. Any answers will be provided here.
• Please be as specific as possible, and include all relevant context – the usefulness of answers may depend on the context.
• Note:
  • We don't answer (and may remove) questions that require medical diagnosis or legal advice.
  • We don't answer requests for opinions, predictions or debate.
  • We don't do your homework for you, though we'll help you past the stuck point.
  • We don't conduct original research or provide a free source of ideas, but we'll help you find information you need.

Ready? Ask a new question!

How do I answer a question?

Main page: Wikipedia:Reference desk/Guidelines

• The best answers address the question directly, and back up facts with wikilinks and links to sources. Do not edit others' comments and do not give any medical or legal advice.

See also:

• Teahouse
• Help desk
• Village pump
• Help manual

April 2

Reading/copying Unix floppy on Windows XP system

One of the managers here at work thought I might be able to help them out with an issue they're having. I'm trying to help but am a bit out of my depth. And it's Friday night, so any real help won't be in until Monday morning.

I have a 3.5" floppy that is used on tools running some form of Unix. They want several more copies of this disk made. All I have available to me is Windows XP, a USB floppy drive, the original disk (they have more, they only gave me one), and a blank disk.

When I put the source disk into the drive and try to open it on my laptop, I get a warning saying that the disk is not formatted and I'm asked if I want to format it. Any ideas? Thanks, Dismas|^(talk) 00:33, 2 April 2011 (UTC)

Probably ext3 format or ext4. There are some Windows drivers for ext3, but I don't know if they support format and copy.[1] I used to use CopyQM Plus to copy non-Windows floppies and to create image files back in the day (they were system disks for a Kentek printer that booted from a hard drive).[2] ---— Gadget850 (Ed) ^talk 01:18, 2 April 2011 (UTC)

If you simply wanted to copy the disk without actually reading its contents, you can use any "floppy image" programs, like RawWrite or WinImage, to copy the disk. 118.96.157.155 (talk) 02:08, 2 April 2011 (UTC)

No, I don't actually need to read the contents. I just need to make copies.

• What I got from fs-driver.org didn't allow me to read the disk
• CopyQM won't work as it's not made to work on anything later than Win98
• RawWrite doesn't work because I don't have a disk image on the disk. It still fails telling me that the disk isn't formatted.
• I suspect the results for WinImage will be the same as RawWrite.

Anyone else? Dismas|^(talk) 02:22, 2 April 2011 (UTC)

Really? Did you image ("read") the disk first before writing it in RawWrite? Also, even if Windows keeps saying that the disk is "not formatted", if you copied the disk using RawWrite, it is still copied no matter what. Copying a disk has nothing to do with Windows saying the disk "formatted" or "not formatted". 118.96.157.155 (talk) 02:29, 2 April 2011 (UTC)

Really. I opened the program, clicked on the "Read" tab, and it's asking me for a disk image file. If I go looking for one on the disk, the program tells me that the disk is not formatted and asks if I want to format the disk. Dismas|^(talk) 02:32, 2 April 2011 (UTC)

Don't do it. Instead save the image on your hard disk with the "Read" tab, then write the saved image to a blank disk with the "Write" tab. To use CD burning analogy, the "Read" tab is used to "rip" the disk to a disk image, and the "Write" tab is used to "burn" the image to a disk. 118.96.157.155 (talk) 02:37, 2 April 2011 (UTC)

I didn't. What I'm trying to get across is that there is no image. There is nothing, according to Windows or RawWrite, on the disk. There's nothing there to rip as far as the program is concerned. Dismas|^(talk) 02:39, 2 April 2011 (UTC)

There is indeed no image in the disk. You need to create the image first from the disk, then write the created image to a new, blank disk. 118.96.157.155 (talk) 02:41, 2 April 2011 (UTC)

When the "Save As" dialog comes up, you should choose a place to save the (newly created) image on your hard disk. The floppy drive is selected from a different drop-down box in the main application window. -- BenRG (talk) 03:04, 2 April 2011 (UTC)

(indentation reset) Here's a walkthrough:

1. Insert the source disk into drive A:.
2. Launch RawWrite, then do the following:
   1. select A: in the Floppy drive list box,
   2. select the Read tab,
   3. enter C:\asdfghjk.img (or any non-existing file on your hard drive) in the Image file text box,
   4. press the Read button, then
   5. follow further instructions onscreen.
3. Eject the source disk from drive A:, then insert a blank disk into drive A:.
4. In the RawWrite again, do the following:
   1. select A: in the Floppy drive list box.
   2. select the Write tab,
   3. enter C:\asdfghjk.img (the file name you enter above) in the Image file text box,
   4. enter 1 in the Number of copies text box,
   5. press the Write button, then
   6. follow further instructions onscreen.

Good luck. 118.96.157.155 (talk) 03:07, 2 April 2011 (UTC)

Well, I no longer have the disk. It had to be returned. And I don't see how "Read" means "Create an image file". Doesn't seem very intuitive to me at all. Dismas|^(talk) 06:25, 2 April 2011 (UTC)

I am sorry you couldn't copy the disk on time. Next time someone is asking me how to copy floppy disks, I probably should tell them to run the DISKCOPY A: A: command instead. It is simpler and more intuitive (yet far less powerful) than RawWrite or WinImage. Oh well. 118.96.154.96 (talk) 08:34, 3 April 2011 (UTC)

I wonder if you are able to download and burn a (small) linux live CD. Boot your PC from the CD and use linux to copy your floppy. Don't have a CD?... maybe some virtual machine software will work directly with the live CD's .ISO file and let you start linux without needing a CD. Astronaut (talk) 07:24, 2 April 2011 (UTC)

Microsoft Comfort Mouse 4500 in Ubuntu

Hi everyone, I recently bought a new Microsoft keyboard and mouse. The keyboard works perfectly except for a few of the random Messenger/Documents/Zoom buttons which I don't care about, but the mouse is driving me crazy - the scroll wheel seems to act as a Page Down/Page Up button! Scrolling down by one click is the same as using the page down button, and scrolling up acts like Page Up. I'm using Ubuntu 10.10 and I thank you in advance for your help :)

110.175.208.144 (talk) 06:13, 2 April 2011 (UTC)

A mouse will typically come with a driver CD which allows you to customize this type of behavior. Is the problem that this driver lacks Linux support (or specifically Ubuntu) ? (It wouldn't surprise me if MS does everything possible to sabotage Linux users.) StuRat (talk) 16:40, 2 April 2011 (UTC)

Duplicate file detection

Is there an easy way to detect and perhaps remove duplicate files from a directory? The files could be anything: documents, text, images, music, etc, but there could be hundreds of files to process. Any duplicates would have some things in common, but other things could be different: filenames, dates, other meta data. I thought about writing a program to do this, something using MD5 and moving duplicates to a different directory, but maybe there's already such a program available (preferably for free). Astronaut (talk) 10:01, 2 April 2011 (UTC)

There are lots, I like this one 82.43.90.38 (talk) 10:26, 2 April 2011 (UTC)

See fdupes. ¦ Reisio (talk) 20:40, 2 April 2011 (UTC)

I like Duplicate Cleaner http://www.digitalvolcano.co.uk/content/duplicate-cleaner , and Visipics for finding similar images http://www.visipics.info/index.php?title=Main_Page . Both freeware. Edit: I used the 1.6 version of Duplicate Cleaner. I havn't tried the latest version yet, which seems somewhat different. 92.15.2.39 (talk) 16:31, 3 April 2011 (UTC)

Similar file detection

Following on from my question immediately above, is there a program that can identify similar image files (again, preferably for free)? By similar, I mean perhaps an image has been converted to a different file format (eg. .jpeg -> .png), or has been resized, or has broadly the same colour palette (I'm thinking something along the lines of the colour selection in Google's image search). Astronaut (talk) 10:01, 2 April 2011 (UTC)

This program can 82.43.90.38 (talk) 11:30, 2 April 2011 (UTC)

Content-based image retrieval ¦ Reisio (talk) 20:41, 2 April 2011 (UTC)

Visipics, as mentioned above. Something similar that would be very useful would be a program that sorts similar images into the same sub-directories. 92.15.2.39 (talk) 16:33, 3 April 2011 (UTC)

Laptop

Which brand is most reliable among laptops. Is it OK to buy one without an OS ? Will it be as easy to install Windows (XP or Vista) as on a conventional desktop ? Is the laptops BIOS any different from desktops ?  Jon Ascton  (talk) 17:57, 2 April 2011 (UTC)

Of the major brands, I doubt there is much difference in quality. I have an 11 year old Dell that a relative still uses; I've most experience with Acers, and they've proven to be very reliable. But others will doubtless have good experiences with other brands, and maybe bad experiences with these. You'll find it very difficult to find a laptop without an OS; these days most have Windows 7 Home Premium or Home Basic - netbooks might still have XP. The BIOS is much the same as for a desktop and a laptop. In general installing XP on a laptop is straightforward, but you will increasingly run into problems where there is no XP driver available for some of the laptop's hardware. Laptop manufacturers figure they're shipping the machine with Windows 7, so they need only code a driver for Windows 7. -- Finlay McWalter ☻ Talk 18:20, 2 April 2011 (UTC)

http://tuxmobil.org/laptop_oem.html ¦ Reisio (talk) 23:34, 2 April 2011 (UTC)

Note that you'll probably wind up paying more for Windows this way. Microsoft sells OEMs Windows for an undisclosed but probably large markdown over buying it off-the-shelf. It makes sense to buy OS-less computers to install Linux on them, but not Windows. (Sadly, it often makes sense to buy a Windows computer, even if you're going to install Linux on them, because there aren't many options when purchasing OS-less or Linux laptops.) Paul (Stansifer) 08:10, 3 April 2011 (UTC)

One caution: It may be impossible to fully test a laptop without an O/S. So, if you do purchase one this way, make sure you install an O/S and fully test it out, well before the warranty expires. StuRat (talk) 23:25, 3 April 2011 (UTC)

RAR for mac

Anyone have a recommendation for a good free application for unpacking RAR files using a Mac?--141.155.147.132 (talk) 21:03, 2 April 2011 (UTC)

There's a command line version of WinRAR available here and a mac version of 7zip available here. I've never used a mac so I don't know how good they work 82.43.90.38 (talk) 21:40, 2 April 2011 (UTC)

StuffIt Expander (link) can extract .rar files. 118.96.156.239 (talk) 01:36, 3 April 2011 (UTC)

Thank you both for the effort. However, let me repeat my question with a bit of emphasis: "Anyone have a recommendation for a good free application for unpacking RAR files using a Mac?" I downloaded the first, installed it, then, when I tried to use it, I learned you must register for it to work, which requires getting a serial number, which in turn requires payment. Stuffit is "free" alright, you just have to "successfully sign up with or purchase from any one of our affiliate partners during the TrialPay checkout".--141.155.147.132 (talk) 04:10, 3 April 2011 (UTC)

Okay, I have downloaded this, blindly, though at least it's through a site that appears to be reliable.--141.155.147.132 (talk) 04:21, 3 April 2011 (UTC)

Oops. Sorry. I gave you the "wrong link" (StuffIt Expander is indeed free). Here is the corrected link, but I guess it's too late now. 118.96.154.96 (talk) 08:16, 3 April 2011 (UTC)

P.S. If you are still interested in StuffIt Expander and want to download it, ignore all the "Buy", "Special Offer", and "TrialPay" links, and all the banner ads on the page. As I say, StuffIt Expander is indeed free. 118.96.154.96 (talk) 08:43, 3 April 2011 (UTC)

UnRarX seems to be what you're looking for. link -- Bob drobbs (talk) 05:23, 5 April 2011 (UTC)

April 3

SeaMicro SM1000-64 System

Hello RefDeskers!

  The SeaMicro SM10000-64 Server is stated to comprise 256 dual-core processors, with 4 processors on each of 64 Compute Cards. Does this mean that the SM10000-64 is a single enormous 256-processor server, or is it a cluster of 64 quad-socket servers?

  Thanks as always. Rocketshiporion♫ 03:03, 3 April 2011 (UTC)

Cluster of 64 4-socket servers, and the cpu's are not very powerful. It's for cloud hosting, not a supercomputer. 75.57.242.120 (talk) 08:27, 3 April 2011 (UTC)

If I am not mistaken, the SM10000 is a cluster of 256 servers and associated infrastructure in custom packaging. A SeaMicro white paper (SeaMicro Technology Overview) seems to agree. It describes each server as a credit card-sized board that is installed in groups of four into a motherboard. It then describes each credit card-sized board as a node. I doubt the Atom processor which the SM10000 uses has support for inter-socket multiprocessing since it is a low-end processor. Rilak (talk) 08:56, 3 April 2011 (UTC)

Awww... No supercomputing on the SM1000-64. Darn! And I thought I'd found a supremely-powerful supercomputing platform. Rocketshiporion♫ 05:15, 4 April 2011 (UTC)

W3C standards as a source for browsers, or vice-versa?

I've heard it claimed that new feature additions to the W3C standards for HTML, XHTML, CSS, DOM and JavaScript are usually made to comply with, or arbitrate a compromise between, what the major browsers are already doing, and that it is the browser developers that invent new functionality. Is that right? I thought it was more common for features to be first codified in W3C working drafts and then implemented in browsers. NeonMerlin 03:22, 3 April 2011 (UTC)

I don't know much about the W3C specifically, but industry standards bodies are normally made up of representatives from organizations interested in implementing the standards, and implementations are written in parallel with the standards themselves, so that neither one comes first. It's also common for standards to simply document existing practice and introduce nothing new. An ivory-tower "standard" developed without the input of likely implementers would probably just be ignored. -- BenRG (talk) 22:15, 3 April 2011 (UTC)

All the principal browser vendors are part of the W3C. ¦ Reisio (talk) 01:11, 4 April 2011 (UTC)

Yes, but in the olden days they all gladly innovated independently, giving us things like the infamous <blink>blink tag</blink>. Much of the internet has been developed by "rough consensus and working code", and this carried over into browser land. Nowadays, however, the W3C spec is, more-or-less, leading implementations. --Stephan Schulz (talk) 02:17, 4 April 2011 (UTC)

Not really. ¦ Reisio (talk) 21:47, 4 April 2011 (UTC)

DVD world "Zones" - How to increase the allotment?

On my Dell Inspiron 1720, I have only about 4 more times to change the zone/regions. I plan to be some type of a jetsetter in the coming years, so it would be essential to keep changing DVD regions. What can I do to increase the allotment of region changes? Or how can I make it a universal-region DVD? --70.179.169.115 (talk) 04:12, 3 April 2011 (UTC)

Unfortunately, there is no 'legal' way of altering the number of times you can change the region code. If you want to try, please by all means try it yourself- I think under the no legal advice section is that we cannot tell people how to circumvent these security measures. General Rommel (talk) 10:48, 3 April 2011 (UTC)

General Rommel - perhaps you can quote the law or statute that you have in mind here - any jurisdiction will do. I find it difficult to believe that it is illegal to alter these settings on hardware that you own, with properly licensed software, and for a legitmiate reason e.g. you are frequently moving to different countries. It may be technically difficult, it may void certain warranties, it may not even be a sensible solution to the original problem, but how exactly is it illegal ? Gandalf61 (talk) 12:33, 3 April 2011 (UTC)

I agree with the sentiment, but see Modchip#Legality. To not buy multiple copies of a DVD is a circumvention of copy protection, and so is preparing to not buy them. 81.131.41.83 (talk) 13:33, 3 April 2011 (UTC)

I was not expressing a sentiment - I was asking General Rommel for a source for their assertions. Modchip#Legality says "The nonuniform interpretation of applicable law by the courts and constant profound changes and amendments to copyright law do not allow for a definitive statement on the legality of modchips", so it sits on the fence. Our article on DVD region code lists several ways of bypassing region codes through software or firmware; none of these methods are stated to be illegal. Indeed, the article says that the legal staus of region codes and mechanisms that enforce them is unclear in several jurisdictions, and in New Zealand, for example, they have no legal protectionat all. Gandalf61 (talk) 14:46, 3 April 2011 (UTC)

(EC) I don't know if 'not buying' is really the issue. I do agree there's a fair chance it may fall foul of the Digital Millennium Copyright Act restrictions on circumvent copy protection measures (although for other reasons it may depend on the reason you need to use a different region see [3] for example). As with many areas of copyright law in the US, I don't think this has been tested but see [4] and also consider the fact the US Copyright office has refused to grant an exemption for removing regional restrictions [5]. [6] (consider why the EFF felt it necessary to ask in the first place) and [7] (the US Copyright offices view appears to be that the current limited number of changes is fine). This is even more likely to be a problem if you use something, like VLC, which uses DeCSS to get around the restrictions, see DVD-Video#Content Scramble System as there is case law strongly suggesting it is not allowed (unless exempted) AFAIK (looking a bit more I think it's only distributed which is accepting as not allowed, possessing has not been tested and a is common I suspect it never will [8]). However as our article DVD region code notes, the legality of DVD regions themselves is questionable in some countries like Australia and possibility NZ. The OP however appears to come from the US at the current time. P.S. I seem to remember discussing this before so there may be more sources in the archives. Nil Einne (talk) 15:06, 3 April 2011 (UTC)

From VLC media player: VLC is one of the free software and open source DVD players that ignores DVD region coding on RPC-1 firmware drives, making it a region-free player. However, it does not do the same on RPC-2 firmware drives. - might be worth a try. -- 78.43.60.13 (talk) 11:40, 3 April 2011 (UTC)

All drives manufactured since 2000 are RPC-2 drives. But I think that VLC also bypasses region coding on RPC-2 drives (i.e., the article is wrong). There certainly are players that do. Also, the various DVD ripping programs work on modern drives. There are custom patched firmwares for some drives that disable the region lockout. There are also some software products (none of them free, as far as I know) that transparently remove the protection for all applications. The most straightforward solution is probably ripping to the hard drive and playing from there. As mentioned above, the legal situation is unclear. -- BenRG (talk) 01:44, 4 April 2011 (UTC)

I should have said that that was purely first reaction statement. But I believe it's just not gright to.. But if you do want, I can say that usually using a cheap $20 DVD player from an Asian electronic store will usually do the trick. General Rommel (talk) 13:04, 4 April 2011 (UTC)

One thing to note, and at the risk of stating the obvious, your PC's DVD drive doesn't know which country you are in. If your existing Region 1 disks play now, they will still play if you visit Europe, Asia, etc. You only need to change the drive's region if you want to play DVDs from other regions or you buy new DVDs locally. For the latter, it might be better to buy a local drive as Gen Rommel suggests, especially if you can get one that is already multi-region capable. Astronaut (talk) 13:35, 4 April 2011 (UTC)

DOS in laptops

Some models of laptops are coming with what they advertise as "Free DOS". What exactly is this ? DOS, as I remember, was an operating systems that is now history, it was used when there was no GUI and commands had to be written as text, what is meaning of using DOS in today's computers ?  Jon Ascton  (talk) 14:14, 3 April 2011 (UTC)

FreeDOS. -- Finlay McWalter ☻ Talk 14:16, 3 April 2011 (UTC)

I've seen a fair few laptops like that in Malaysia. I'm not surprised if the situation is similar in India. AFAIK it's far less common in NZ and I would guess other places like the US, UK etc. FreeDOS is basically selling the laptop without an OS (by actually including an OS they can show something when they turn it on and I guess they can claim they aren't encouraging copyright violations because they did include an OS it's not their fault if the customer doesn't like it). I doubt many people use the FreeDOS for long if at all because there's not that much you can do with it. They could use Linux, FreeBSD or something else but that creates additional support issues. The reason for selling without OS is I'm pretty sure to reduce cost (or at least create the illusion of reduced cost, as our article notes they aren't always cheaper), with the assumption the customer is going to install their own own OS perhaps one of the aforementioned options but far more likely some copyright violation version of Windows (probably Windows 7 Ultimate or perhaps Windows XP Pro) or occasionally Mac OS X and hence the reason this is common in places like Malaysia and India. Notably while they could probably add a cheap version of Windows 7 for not that much more they may figure many are just going to change it to Ultimate anyway. I've also seen this tends to be with the cheaper models. Of course not everyone will do this, some may get the shop to install a legitimate version of Windows for them. (In other words, even if the customer is going to pay for the OS, you again create the illusion of lower cost by excluding the OS cost from the upfront cost.) I don't know whether the cost will be the same since Windows OEM licensing confuses me but labour time wise it's probably insignificant compared to many developed countries where it would likely add a fair amount to the cost. Nil Einne (talk) 14:37, 3 April 2011 (UTC)

Nil Einne is spot on. A laptop is sold with freeDOS so that the seller doesn't need to buy a copy of Windows to put on it. Most people won't use the FreeDOS: they'll quickly upgrade, typically to something else free (pirated Windows, or some sort of Linux). If you do plan on installing Windows legitimately, it's almost certainly cheaper to buy a laptop with it installed (if you can), as Microsoft gives discounts on their OEM licenses over retail licenses. See this helpful guide. Buddy431 (talk) 16:04, 3 April 2011 (UTC)

Won't totally removing this DOS from laptop's harddisk will present any problems ?

Of course it is perfectly possible that the customer already has a license for a version of Windows they are comfortable with, and has no interest in an upgrade. 84.239.160.59 (talk) 17:03, 3 April 2011 (UTC)

Removing DOS shouldn't be a problem: when you install your new operating system, you'll reformat the disk, essentially erasing what was on it before. And you are correct, if a person has a valid license from an old computer, or a license that's good for multiple computers (I think Microsoft's retail licenses are typically good for up to three computers at once), they could install that. Note, however, that most licenses for OSs that come with the software are only valid for the machine they came with (so called OEM licenses). It is not legal (or at least against Microsoft's terms of use, legality depends on jurisdiction, how good your lawyer is, etc.) to install one of these on a new machine that it did not come with. Technically, you may be able to do it anyway, depending on the reactivation process, and how well you can lie to Microsoft if you have to reactivate by phone. Buddy431 (talk) 19:23, 3 April 2011 (UTC)

Home Wi-fi networking question

I have several devices in a part of the house where there's no wired Ethernet port. I have a Wi-fi AP in the house but the devices in question don't have built-in Wi-fi support. What would be the best or simplest way to connect these devices to the (W)LAN in the house?

I've seen "Wi-fi extender" products with Ethernet ports, which seem to be a possible solution. Are wireless routers these days also usable as Wi-fi extenders? (I'm asking because, given the choice, I'd rather get a device usable in several ways than a special-purpose one.)

I remember seeing mods for converting a wireless router into the kind of Wi-fi extender I just described, but that was from quite a few years ago. If it's no longer necessary to do any mods, using a device unmodified is preferred.

--108.36.90.190 (talk) 16:57, 3 April 2011 (UTC)

If you say that the devices are incapable of WIFi or Ethernet, or any other wireless way of communication, well that would mak it very tricky.... General Rommel (talk) 01:45, 4 April 2011 (UTC)

I believe the OP was saying it's the part of the house rather than the device that lacks ethernet. --Sean 16:32, 4 April 2011 (UTC)

Some wireless routers are. See Wireless Distribution System. This page touches on using Apple's AirPort Extreme or AirPort Express routers to form a WDS — the page is short on detail but has links and a very pretty picture. Of course, may I ask whether you have considered whether the simplest solution is to add wireless capability to these devices, via a PC Card or USB device? I'm pretty sure that in all cases, that'll be more straightforward than trying to configure multiple routers to talk nicely to each other. Comet Tuttle (talk) 18:08, 4 April 2011 (UTC)

I believe you would be looking for a WiFi Bridge, one that will take wireless signals and allow a regular ethernet hookup. These devices were popular a few years ago when wireless technologies were just getting a footing, but slowly fell out of favor. You will find them now labeled as "Gaming Adapters," mostly marketed to connect gaming consoles with only ethernet ports to Wifi. For example, there is this one however I did find one that looks like it will fit your needs exactly. We of course do not endorse Amazon or Netgear, but this looks like that second one will fit the bill perfectly. Let us know how it goes! --rocketrye12 ^{talk/contribs} 20:55, 6 April 2011 (UTC)

Limits of lossless data compression

How close is something like 7-Zip to the theorectical limit for lossless data compression? I recall reading something about this in the past - why cannot the best possible algorithmn for lossless compression be deduced? Thanks 92.15.2.39 (talk) 17:10, 3 April 2011 (UTC)

How do you define 'best' ? Fastest, highest compression? Specific to which domain of files to be compressed? All algorithms make trade-offs to better suit the type of file they intend to compress, there cannot be a 'best' algorithm for 'all' possible files. But the theoretical limit for an algorithm suited to one specific file, might be 1 bit. Unilynx (talk) 21:26, 3 April 2011 (UTC)

Why not 0 bits ? That could be used to mean that the file is in some default state, hopefully the most common state for that type of file. StuRat (talk) 23:17, 3 April 2011 (UTC)

A large file of apparently random bits might be a bunch of zeros encrypted by some algorithm with some key. In that case the entire file can be compressed down to almost nothing if you can find the key, but is probably incompressible if you can't. This shows that optimal compression is at least as hard as breaking current encryption algorithms. Compression is also related to science, because the shortest representation of data is more or less the same as the simplest theory describing the data in the sense of Occam's razor. So optimal compression is as hard as doing science. It's related to artificial intelligence too, for similar reasons.

The only estimate of an optimum compression ratio that comes to mind is Shannon's 1951 paper Prediction and entropy of printed English, where he used the prediction ability of human subjects to estimate an entropy of about 1 bit per letter for written English (about 8:1 compression of ASCII text, in modern terms). The modern state of the art is around 8:1 for 1,000,000,000 bytes of an XML dump of Wikipedia, but this is not directly comparable since it includes a lot of metadata (and punctuation, which I believe Shannon didn't consider). 7-Zip didn't do very well on this test, with a total size about 40% higher than the state of the art (using PPMd). -- BenRG (talk) 02:15, 4 April 2011 (UTC)

I thought it had been mathematically proved that the theorectical limit for (dictionary-less, lossless) compression was a lot lower than achieved by current algorithms? 92.29.115.116 (talk) 10:31, 4 April 2011 (UTC)

If you look at the average over all possible files, the limit of compression is 1 (as in "no compression"), as can be seen easily by a counting argument (to uniquely identify an n-bit file, you need n bits). If you look at a finite subset, then Shannon indeed describes that limit. English text has a lot less entropy than random noise, and most program texts have less, again. In principle, encrypted files should have about the same entropy as their un-encrypted counterparts. However, they are mapped into the space of all texts in a way that makes them very hard to differentiate, and hence in practice they compress badly. --Stephan Schulz (talk) 11:51, 4 April 2011 (UTC)

There's no such proof, but I'd be interested in seeing your source. Maybe you're thinking of the Shannon limit? All state-of-the-art compressors use entropy encoding that is very close to an optimum proved by Shannon, but that's the easy part. The hard part is coming up with the probability estimates to feed into the entropy coder. -- BenRG (talk) 19:07, 4 April 2011 (UTC)

The "theoretical limit" is the Kolmogorov complexity, which is not a computable function for the same reason that the halting problem is undecidable. The article about it explains more. 75.57.242.120 (talk) 20:31, 4 April 2011 (UTC)

random numbers and the system clock

Why don't computer random number generators just use the last few digits of the system clock, ie. measuring fractions of a second? I know they use the clock, but only as a seed. Why would the last few digits of something measuring fractions of a second not be random enough? Is it because a computer program runs so quickly, and with such regularity, that there would be inevitable patterns in the stream of numbers thus produced? Thanks, It's been emotional (talk) 17:14, 3 April 2011 (UTC)

You pretty much have it. If the program asks for random numbers quickly enough (that is, faster than the counter wraps around), they aren't going to look random at all - they'll be in increasing order. Secondly, under similar loads, computers tend to run at about the same speed, so that any two invocations of the call to get random numbers will be at about the same difference in time so the two numbers will have a non-random difference between them (this applies even if the counter wraps around several times) - this is a killer if you have a loop to get multiple random numbers, say to shuffle a deck of cards or simulate a population. In either case you would need to pass the resultant value through a function which would "deorder" the values, and amplify the small variations between close numbers: this is effectively what pseudorandom number generators do. They also have the benefit of being repeatable; that is, with the same seed, you will always get the same string of "random" numbers - handy for debugging. -- 174.21.244.142 (talk) 17:37, 3 April 2011 (UTC)

Additionally, from a practical perspective, getting the current time from the system clock will require a system call, which is a comparatively expensive operation. Grabbing the time once, and then using a quick pseudorandom number generator thereafter, means the program has to make only one syscall; grabbing the time whenever a random number is needed will be a lot slower. —Bkell (talk) 17:49, 3 April 2011 (UTC)

On x86 processors the Time Stamp Counter can be read without a system call, probably a lot more quickly than the execution time of a decent PRNG. But, as already mentioned, it's not random enough. -- BenRG (talk) 01:26, 4 April 2011 (UTC)

Thanks to all, very interesting and exact, It's been emotional (talk) 08:54, 7 April 2011 (UTC)

Sound Card and Microphone Questions

Ok I have two questions. Firstly how do I check my sound card on Windows Vista. I have read that looking under the "Sound Video and audio device", section of the device manager or that running "Dxdiag" and checking the sound tab, should have my sound card listed however the only thing listed in the device manager is"high definition audio device", while the sound tab of dxdiag says "Digital output device (SPDIF) high definition) both of which seems more like a description of a sound card, then a sound card itself.

My second question is how do I get my speakers to output what I say into my microphone directly on Windows Vista. I have read that right-clicking the volume control, in the bottom right of my screen, clicking on playback devices. Then right-clicking on speakers and selecting levels should let me uncheck a box to unmute my microphone, but the only thing there is a section called "speaker/headphones" which isn't muted. So could anyone help me with these problems. 86.162.151.128 (talk) 17:32, 3 April 2011 (UTC)

For what it's worth, on the Vista system I am currently sitting at, the relevant entry in Device Manager is "Sound, video and game controllers", and the two items are "ATI Function Driver for High Definition Audio" with a part number, and "IDT High Definition Audio CODEC". These are indeed just driver files. The dxdiag tool has three relevant tabs, "Sound 1", "Sound 2", and "Sound 3", for the three hardware audio outputs on this machine. Comet Tuttle (talk) 18:39, 3 April 2011 (UTC)

RAID 1 and Subversion questions

Hi,

I am new to setting up RAID systems. I'm currently setting up a Subversion server using Windows XP, having used this guide to getting Windows XP to format two drives (not including the Windows drive) as a single software RAID 1 volume. No errors have been reported.

Question 1: I now want to perform the practical test of disconnecting each of the drives in turn to see whether the other drive has in fact got all the data expected, but I realize that as soon as I boot up the system with one drive disconnected, something is going to get written to the active drive, so the two drives won't be mirrored anymore. How do other people perform practical tests of a RAID 1 volume's mirroring?

Question 2: When one of the two drives in the RAID 1 volume fails, and I install a replacement, what tool is used under Windows XP to mirror the good drive to the replacement drive?

Question 3: A Subversion question. All workstations will be running Windows. Is it stupid for me to set this server up as a Windows XP machine rather than an Ubuntu machine? I have the vague notion that the most recent Ubuntu may have better tools available for administration of the server; but I find it a little appealing to run a Windows server to service Windows clients — we've run into one problem in the past with a GNU/Linux Subversion server, when my users were renaming the case of files, and the Linux server allowed both the files "hello.c" and "hElLo.c" to exist in the same directory, which caused problems with all the Windows clients. In a vague way, I have the notion that a Windows server might prevent other, similar problems.

Thanks! Comet Tuttle (talk) 18:34, 3 April 2011 (UTC)

It's been ages since I worked with Windows XP, so I don't remember the answer to your second question.

Regarding #1, there is not really a non-destructive test for this. However, you could use dd on both drives after booting from a live CD, store these images somewhere else, and restore them after you performed your tests. Another option would be to use three drives, always rotating one out and checking it in another computer to see if it contains the needed data, then locking that drive away in your storage location/vault/whatever.

Regarding #3, please keep in mind that XP was never meant as a server operating system and thus places a limit on the number of simultaneous connections (it's a licensing issue, not really a technical limitation per se). So if you want to go the Windows route, prepare to have to shell out some $$$ for W2K8 server. On the other hand, if you wish to follow the F/LOSS route, take a look at mount options; for some file systems it is possible to force upper- or lowercase. Maybe there's an option in subversion for that, too - but I'll leave answering that to a subversion expert. -- 188.105.131.8 (talk) 09:04, 5 April 2011 (UTC)

Thanks for the partial answer. What would be useful to me in question #1 would be a jumper that would write-protect a hard disk. Comet Tuttle (talk) 17:56, 5 April 2011 (UTC)

A tablet computer question

Dear Wikipedians:

With tablet computers replacing netbooks as the latest "fad" in computing, I myself am thinking of getting a tablet. I will probably not get an iPad, for reasons that will become obvious after I have outlined my questions below. But I do have the following questions before I make my tablet computer purchase decision.

I remember that for my recent desktop PC, I went to my local computer store and bought all the parts, like motherboard, CPU, memory chip, a new SATA 250 GB hard drive, and then assembled them at home. Since the hard drive is new, there is obviously nothing on it. So I had to partition, format and then install my favorite operating systems. Now the PC works like a charm.

Now I am wondering, for tablet computers, do I have the option of going out there and grab different parts and assemble my own tablet? Barring that, is there a tablet computer that comes as a "clean slate" and allow me to install whatever operating systems I want on it, including options for multi-booting, just like what I do with my desktop PC? Barring that, if I get a tablet, say, with Windows 7 preinstalled, can I wipe it out and then setup a Windows XP/Linux dual boot system, just like what I do with my laptop computer? And by "tablet" I don't mean the clunky early tablet computers with the pivotable screen like the Lenovo Thinkpad X, but rather the "cool" one-piece tablet like iPad's form factor.

Now to push the envelope a bit, for smartphones, do I have the option of going out there and grab different parts and assemble my own smartphone? Barring that, is there a smartphone that comes as a "clean slate" and allows me to install whatever operating systems I want on it, including options for multi-booting, just like what I do with my desktop PC? Barring that, if I get a smartphone, say, with Android preinstalled, can I wipe it out and then setup a Windows XP/Linux dual boot system, just like what I do with my laptop computer?

And if the answers to my above questions are "no"s in their most restrictive sense (i.e. the only option I have is to put up with whatever crap the OEM throws at me). Then my question is: What is the reason for this erosion of my digital freedom, where I am able to do anything I like with my desktop PC, yet can do nothing with my tablet/smartphone? And should I be worried about this trend of OEMs gradually taking stronger controls of us consumers, and telling us what we can or cannot do with the gadgets that we bought with our own money? Am I witnessing the beginning of the end of our liberal, western democratic states?

Thanks,

L33th4x0r (talk) 23:15, 3 April 2011 (UTC)

Increasingly, many tablets and smartphones are designed so that it is difficult to access the actual layer (like behind the nice interface of a phone) and do things like delete folders. As well as that, Hardware for such things nowadays tend to be customized by the manufacturer. And as for the customization of tablets and phones, it will become increasingly difficult because of the fact that these devices are ever so becoming smaller, which means that it will become very hard to DIY. You could try building your own tablet, but you will probrably end up with a large sized one instead. General Rommel (talk) 01:42, 4 April 2011 (UTC)

I'd hardly call the trivially larger form of the Lenovo Thinkpad X tablet clunky, particularly when comparing it to something so much less versatile, less powerful, & incredibly less extensible such as either model iPad. You're not going to find a time where a Thinkpad X won't fit and an iPad will, it just isn't going to happen. ¦ Reisio (talk) 02:08, 4 April 2011 (UTC)

At the moment tablets are closer to an Embedded system then a fully functional "computer". That might change in the future, but at the moment you probably find that most of what you are "hoping" for is not possible on tablets or smart phones. There might be some hackers who get a different OS running on a tablet or a phone, but no doubt it will take a LOT of hacking and will also probably offer only limited functionality compared to the device it was designed to run on. Might be worth to note that the same thing could have been said of most laptops ten years ago, now it's not uncommon for people to dual boot macbooks and the like, it's possible when tablets become more common and ubiquitous they might become also more interoperable. Vespine (talk) 02:17, 4 April 2011 (UTC)

The issue with tablets is that they are generally lightweight and the size people want them in precludes any real DIY stuff with the hardware. When you shove so many components into such a small place, the only way you can make a tablet that will offer any real competition to the iPad, Xoom, and other such tablets is to solder chips to a circuit board yourself. Tablets (and smartphones) have only a couple circuit boards, and there is rarely space for connections that are easily used by human fingers, much less interchangeable components. Unlike desktop PCs where there is usually plenty of space to work and long running standards for component sizes, placement, and connections a smartphone or tablet has a small amount of space to work in, and components are shaped and put together much like a puzzle. Even in the best case scenario where components were completely standardized, the amount of DIY work you could do on a tablet would be minimal at best. Your best bet right now is to look into rooting an android tablet. Building the actual hardware will result in something cost prohibitive in small runs, or incredibly bulky. Caltsar (talk) 18:54, 4 April 2011 (UTC)

Thanks for all the responses. I now fully understand the reasons behind not being able to DIY tablets and smartphones. I understand that with more powerful tablets and smartphones in the future, and them becoming ubiquitous, open platform and general purpose variants of those machines may eventually emerge. I am putting my faith now in riding the Kurzweilian exponential technological progress curve. 174.88.242.201 (talk) 21:37, 4 April 2011 (UTC)

Resolved

April 4

PAM

can u explain me in detail the pulse width modulation using frequency division multiplexing? — Preceding unsigned comment added by Saiffuddin (talk • contribs) 02:24, 4 April 2011 (UTC)

Have you read our article Pulse-width modulation? Comet Tuttle (talk) 03:09, 4 April 2011 (UTC)

jdbc insert query problem

I am using JDBC with oracle data base, All the table are well formed normalization . After INSERT THE TABLE I NEER TO GET THE INSERTED PRIMARY KEY VALUE

PLEASE HELP ME

tHANKS IN ADVANCE

—Preceding unsigned comment added by 192.8.211.14 (talk) 08:05, 4 April 2011 (UTC)

Welcome to the Wikipedia RefDesk! I hope you find the RefDesk to be a useful resource. Please do not write in ALL CAPS, as it is considered the online equivalent of shouting, and is hence bad netiquette. Thank you.Rocketshiporion♫ 01:16, 6 April 2011 (UTC)

You had to create a sequence before you created the table to get auto-numbering, which I assume is what you are doing or you would know the primary key. Oracle doesn't have a "this was the number" column like most other databases. You have to fetch it beforehand. The sequence will have a field "nextval" that contains the value that will be used next in autonumbering. There is a problem. What if you fetch the next val. Then, I fetch the next val. Then, you insert. Then I insert. We both think we are using the same primary key. Another solution is to set up a trigger that saves the primary key and the username of the person performing the query to a separate table whenever an insert is performed. That is useful as long as you don't repeat the same issue above with two people (or two programs) using the same username. Yet another method is to do an insert and then fetch using all the info you just inserted sorting from most recently changed to most distant and grab the first row returned. Then, you get the ID off that. Finally, there is another rather foolproof way. Set up a server-side script that hands out unique IDs. You select an ID from it and it is guaranteed to be unique and not handed out before. Then, use that when you do your insert. You will know that you are always using a unique ID and you will know what the ID is. -- kainaw™ 12:41, 4 April 2011 (UTC)

What about sequence.currval? 98.103.60.35 (talk) 16:50, 4 April 2011 (UTC)

That has the same race condition as using nextval or any other variant. -- kainaw™ 17:31, 4 April 2011 (UTC)

Are you sure? I am under the impression that nextval is session specific, and, as such, is a simple and correct solution. Googling for oracle sequence currval session suggests this is so. 88.112.59.31 (talk) 17:40, 4 April 2011 (UTC)

My experience is that it failed miserably when I had to mass-import a lot of records. Even with each record within a single transaction to fetch the ID and insert the record, I got multiple records with the same ID - resulting in insert errors and loss of data. -- kainaw™ 19:49, 4 April 2011 (UTC)

I am fairly certain you did something else wrong. When used properly, .currval works fine, and is session specific. See Oracle documentation or google as suggested above. 88.112.59.31 (talk) 13:14, 5 April 2011 (UTC)

I should state that my solution in that instance was to alter the transaction to exclusive-lock the data table (not the sequence table - trying that didn't work). Then, get the next ID from the sequence table. Then, insert the record. Then, ensure it was inserted by fetching it again and comparing the values. Then, if all went well, release the lock. Otherwise, throw a tantrum, permanently locking the table, until a human was available to see what else went wrong and alter the program yet again. -- kainaw™ 19:52, 4 April 2011 (UTC)

Thunderbird and Archiving

Would anybody happen to know if Thunderbird archives mail, say, when it gets to a certain full level (rather like the message history of WLM)? I have Thunderbird as my default mail client, with four email accounts running on it - three of which are Gmail and one from my ISP. One of my Gmail accounts (the one I have used longest) now has no mail in the inbox prior to 29th March 2011. I know this 'lost mail' can still be accessed by going to the 'Google Mail' folder (which has all the mail in, whether sent, received, or flagged as spam), and I can also access it just by logging in from the browser, so it's not a problem. I am just wondering where all my mail 'disappeared' to. Would it have been archived for some reason? If so, where to? I'm using Thunderbird 3.1.9, if that's any help. TIA! --KägeTorä - (影虎) (TALK) 12:39, 4 April 2011 (UTC)

Searching Twitter feeds

I'm having a little difficulty searching on Twitter. I'd like to find all mentions of the terms "atemporal" and "atemporality" from the feed GreatDismal, but the search query https://search.twitter.com/search?q=atemporality+from:GreatDismal yields no results despite references existing. The feed is not marked private. Is the query malformed? Takk, Skomorokh 15:35, 4 April 2011 (UTC)

Not all tweets are indexed. Nanonic (talk) 15:50, 4 April 2011 (UTC)

Ah, that explains it, thanks. Any feasible workarounds? Skomorokh 16:03, 4 April 2011 (UTC)

Does Verisign have a domain registrar for verisign.com?

Whois doesn't say. I was just wondering if the registry operator itself for the .com top level domain got to manage its domain registry information directly or if the rules dictated that even they have to formally have a domain name registrar. 20.137.18.50 (talk) 18:46, 4 April 2011 (UTC)

They are registered with Network Solutions, which they own once owned (paid something like $20B during the dotcom bubble). At one point they claimed to own the .com TLD itself. I don't remember what happened to disabuse them of the idea, but there must have been something. Oh man (just looking now), the article about NS is interesting. 75.57.242.120 (talk) 04:36, 5 April 2011 (UTC)

April 5

GNU Octave - GUI with Autocompletion & Tooltips

Good Morning RefDesk!

  In the MATLAB GUI, when the first few letters of a command are typed in, a box pops up with suggestions of possible matches, and provides (upon mouseover) a description of the function of each possible match. Does an equally functional GUI exist for the GNU Octave platform?

  Thanks as always. Rocketshiporion♫ 00:32, 5 April 2011 (UTC)

If you haven't found QTOctave yet, here it is: QTOctave. My experience with it has been so-so; it's a Spanish program that the developer internationalized to an English version (and there's a few rough edges still). On the whole, if I must use Octave on Windows, I use QTOctave (because it's better than Cygwin); but when I'm on *nix, I've found the command-line version works smoothly and integrates with the plotting and other features really well. I use KATE (a KDE text editor) for editing my .m files on *nix; it's totally un-connected to Octave project, but it has syntax-highlighting support and can run a command-line and powerful macros. You can set up more custom features using plugins (though I don't know any for hot-linking function-names; in C, some people use CTAG to analyze source-code and generate hot-linkable meta-documents; CTAGS supports MATLAB and has a KATE plugin). Overall, KATE is a versatile lightweight IDE for MATLAB-style interactions with Octave and other programs; the command-line in the dock is very MATLAB-esque and handy for all kinds of programming jobs. Nimur (talk) 14:30, 5 April 2011 (UTC)

Stopping retweets flooding my feed

I've tried blocking retweets from some of the people I follow on Twitter, but my feed is still flooded with retweets by them. Anyone know a way of actually stopping them? DuncanHill (talk) 10:25, 5 April 2011 (UTC)

What do you use to access twitter (twitter.com, third-party app etc)? Darigan (talk) 11:52, 5 April 2011 (UTC)

I just go to the website from time to time. DuncanHill (talk) 12:44, 5 April 2011 (UTC)

I found this: http://tj.luo.ma/post/1215741762/turn-off-twitter-retweets - It offers instructions for turning off retweets, although, it appears that you have to turn off the retweets one user at a time. I'm using the new twitter interface at the moment, to turn off the retweets in that:

1. Click on the profile of a user you are following so that it shows on the right of your screen

2. Beside the green follow button, there should be a phone symbol, and a little green retweet symbol - Hit the green retweet symbol so that it turns a white/grey colour. Now, retweets from that user will not show up in your timeline.

Hope that makes sense, if not let me know, and I'll try to find a better way to describe it. Darigan (talk) 13:40, 5 April 2011 (UTC)

That's what I did but it didn't stop them :( DuncanHill (talk) 13:51, 5 April 2011 (UTC)

You aren't alone in despising "re-tweets." Twitter hates them too: accounts who re-tweet will be blocked from search and are subject to total account shut-down. The official policy from Twitter is that "re-tweeting" wastes computer and human time. Nimur (talk) 14:36, 5 April 2011 (UTC)

As we talking about the same thing? I doubt twitter has that policy otherwise they wouldn't provide a button [9]. The ref you link to appears to be about people who repost twitter stuff, I guess you could call it a re-tweet but it doesn't appear to be what twitter means. It's not clear if people who just re-tweet (with the button) and do it a lot will also be blocked, from the ref and common sense I presume it may happen but it seems unlikely just doing it occasionally is going to lead to a block since as I said they provided a button. I presume if people use the re-tweet button, then what DH is going will work. I suspect these people aren't actually re-tweeting but posting the same message themselves hence blocking re-tweets probably isn't going to work. Note I rarely use twitter. Nil Einne (talk) 15:53, 5 April 2011 (UTC)

No, they are retweets, as they are labelled as retweets in the feed with a link to the original person who tweeted them. Blocking retweets doesn't seem to actually work. DuncanHill (talk) 15:58, 5 April 2011 (UTC)

The first page I linked to said that following the instructions would not cause already re-tweeted tweets to disappear, but would stop new retweets from showing up - Are you still getting new retweets after following those instructions? Also, not a direct solution to your problem, but when I access twitter from a computer, I use tweetdeck - I find it far more useful when it comes to finding tweets I might be interested in. There are a few other popular apps as well... hootsuite springs to mind. Would it be worth your while trying one of these apps for twitter? Darigan (talk) 20:03, 5 April 2011 (UTC)

Yes, I'm still getting new ones. DuncanHill (talk) 20:36, 5 April 2011 (UTC)

....hmmmmmm, as the tree said to the lumberjack, 'I'm Stumped' (ba-dum cish). Unless somebody else responds with a better response, I would suggest giving one of the apps a go. Mashable has a comparison of 19 different apps here: http://mashable.com/2009/06/27/twitter-desktop-apps/ - Sorry my earlier suggestions turned out to be duds. Darigan (talk) 22:18, 5 April 2011 (UTC)

are these 2 queries equivalent?

select * from a inner join b on a.id=b.id

select * from a left join b on a.id=b.id where a.id is not null

t.i.a. --83.103.117.254 (talk) 13:36, 5 April 2011 (UTC)

No. The first query would not return rows where the a.id is invalid but not null (that is, the a.id is not null but there is no b.id matching that a.id), where the second query would. 118.96.164.80 (talk) 15:39, 5 April 2011 (UTC)

I suspect a typo: they would be the same if the second query was select * from a left join b on a.id=b.id where b.id is not null. AndrewWTaylor (talk) 17:33, 5 April 2011 (UTC)

"The same" as in "returning the same results", that is. They might differ in how the SQL server executes them. AndrewWTaylor (talk) 17:36, 5 April 2011 (UTC)

php

Resolved

In php, is there a way to make a script go back to the start and run again, like the "goto" command in .bat files? —Preceding unsigned comment added by 82.43.90.38 (talk) 15:48, 5 April 2011 (UTC)

Place the script in a loop:

$do_it_again=true;

while($do_it_again)

{

// do a bunch of stuff here.

// If you want it to stop, set $do_it_again=false;

}

There are many many many other ways to do this. IE: Place the whole script in a function and call the function over and over. Basically, the concept of a "GOTO" is not used in normal C-style programming, so it isn't reflected in all the code that derives from that style (almost all of professional programming). -- kainaw™ 15:58, 5 April 2011 (UTC)

I checked - because I know that PHP tries to cater to every bad programming style imaginable, and PHP does have labels. So, you can do this:

begin:

//Do some code here.

goto begin;

This has been implemented with limited functionality. So, using the goto may fail, depending on exactly how you use it. -- kainaw™ 16:02, 5 April 2011 (UTC)

As Kainaw notes, "Goto" is considered pretty bad form in modern programming — it's a recipe for spaghetti code. You're better off figuring out how to control your program flow so that you don't need it — either with functions, or do/while loops, or what have you. It should be a very, very rare instance where you should need to use goto. --Mr.98 (talk) 16:08, 5 April 2011 (UTC)

Though putting the whole thing inside one goto is hardly spaghetti, and Donald Knuth's "Structured Programming with go to Statements" defends their use in some cases: "... what I am really doing is striving for a reasonably well balanced viewpoint about the proper role of go to statements. I argue for the elimination of go to's in certain cases, and for their introduction in others." In this particular case, a goto has the advantage that at the start you can name the label "repeat_everything:" and at the end you can say "goto repeat_everything". If you use a while loop, all you see at the bottom of the code is an uninformative closing brace. Also it has to check every time round to see whether 1 still equals 1, which is silly. 81.131.0.73 (talk) 17:08, 5 April 2011 (UTC)

Obviously goto's are not a universal evil. But they are often misused and can go very awry. Even Knuth agrees that they ought to be relatively uncommon. PHP in particular is not really designed to use them very effectively. --Mr.98 (talk) 20:52, 5 April 2011 (UTC)

I agree with Knuth and you, but I want to point out that most language implementations will optimize away a test on a compile-time constant expression like "true" or "1==1". I don't know whether PHP does, but CPython does, for example, despite doing very little other optimization. -- BenRG (talk) 05:38, 6 April 2011 (UTC)

Even Knuth's example from back in 1973 (if I remember the year correctly) of a problem that required a goto is no longer applicable in most languages because of optimization. The non-goto solution has one more comparison check than the goto solution, but in optimization the extra comparison is omitted. It isn't a knock against Knuth. Computer languages have evolved a great deal since the early 70s. Everything we claim right now will likely be defunct in another 20 years. -- kainaw™ 12:14, 6 April 2011 (UTC)

Knuth's example is as relevant as it ever was because he was talking about source code clarity, not speed. -- BenRG (talk) 22:16, 6 April 2011 (UTC)

Knuth's argument in 1974 was "[The non-goto version] is slightly less readable, in my option, as well as slightly slower; so it isn't clear what we have gained." By 1992, when he wrote "Literate Programming," he dropped his argument of runtime speed and focused solely on the argument that use of goto is "...easier to read and to write." I feel that he was making an argument for clarity and speed in 1974, but has dropped the speed argument since then. -- kainaw™ 17:57, 7 April 2011 (UTC)

Thanks! 82.43.90.38 (talk) 16:30, 5 April 2011 (UTC)

Firefox 4 - how to prevent the white dot appearing

In Firefox 4, when I scroll a page a white button often appears. This button has a dot in the middle and an arrow above and below. It makes my scrolling unpredictable and often scrolls the page at top speed to one end or the other.

Is there any way to prevent this thing appearing please? Thanks 92.24.184.244 (talk) 17:44, 5 April 2011 (UTC)

I haven't used FF4 yet, but in previous versions of firefox, what you describe appears when you press your middle mouse button. It's for people who don't have or don't like scroll wheels, You press the middle mouse button and then 'drag' the page at an accelerated rate. (Even if you have a scroll wheel, it's still useful because few mice also have a horizontal scroll wheel.)

If you're doing this by accident when you're scrolling, you may be pressing too hard on your scroll wheel.

In any case, here's how you turn it off in older version of firefox, I imagine FF4 has a similar option if you just look around for it. APL (talk) 18:00, 5 April 2011 (UTC)

Go to Tools -> Options -> Advanced -> General, and under the Browsing section, uncheck "Use Autoscrolling" You may have to turn on the menubar (right click near the top, you should get a box that has checkmarks with Menu Bar, Navigation Toolbar, etc) 206.131.39.6 (talk) 18:04, 5 April 2011 (UTC)

Comment: this scrolling might not be the fault of Firefox. Some mouse drivers override basic scrolling with a customized version. Check your Mouse settings in Control Panel, too - you might find some advanced options there, especially if a vendor driver has been installed. Nimur (talk) 22:32, 5 April 2011 (UTC)

April 6

please help

YouTube download

I want to download this movie from YouTube. Among the best qualities, there are two options available: MP4 360p and FLV 480p. Which will be better in quality? Are MP4 videos better in quality? Please help, i have a slow internet connection and it takes a lot of time to download videos in my computer. I want the better quality video. --Houlok (talk) 05:54, 6 April 2011 (UTC)

In general I'm not sure there's an easy way to determine whether the MP4 or the FLV will be better quality — what you'd want, ideally, is whichever one is the least transcoded. So if the MP4 is just a conversion to MP4 from FLV, you'd want the FLV; if the MP4 is the raw file from which the FLV was made, you'd want the MP4. Sorry that isn't super helpful. In theory the FLV would be higher resolution at 480p, but my experience is that often these numbers are quite misleading (just yesterday I downloaded one from YouTube which KeepVid.com had said was higher resolution than another format, but they were, once downloaded, identical in terms of their actual pixel ratios). The biggest argument towards downloading MP4 is that you probably already have video software to view it with. FLV usually requires plug-ins or extra codecs of some sort. --Mr.98 (talk) 18:28, 6 April 2011 (UTC)

I added a more useful (sub)title. StuRat (talk) 23:58, 6 April 2011 (UTC)

Disable interpretation of tipping on the pad as click

Hi,

how can I disable the behaviour of my laptop touchpad to inerpret the tipping on the pad as a click? I am using a Dell Inspiron with With Windows 7 Home Premium. Toshio Yamaguchi (talk) 12:19, 6 April 2011 (UTC)

This is known as 'Tap to click,' and you will find the option to disable in the Control Panel under 'Mouse'. Depending on your drivers, it may be in one of a number of tabs. Let us know if you have any trouble.--rocketrye12 ^{talk/contribs} 20:46, 6 April 2011 (UTC)

Problem with SQL outer joins

I ran into a rather complicated problem with SQL today at work. We are using the following kind of SQL query:

SELECT a.name, b.name, c.name, d.name FROM
tablea a, tableb b, tablec c, tabled d
WHERE b.aid(+) = a.id
AND c.bid(+) = b.id
AND d.id(+) = c.did

The idea is that this query must return rows even if no matching row is found from tableb, tablec or tabled, but if no matching rows are found from tablec or tabled, then b.name must be null even if there's actually a value there. I tried various ways of configuring the outer joins but couldn't find a way to do it the right way. Sorry for the nondescriptive names used in this example, but the actual code is under our company's copyright. Can anyone help me to achieve this kind of thing in SQL? JIP | Talk 18:44, 6 April 2011 (UTC)

Just to make sure we understand what you want, could you give us sample data and the desired result ? StuRat (talk) 23:57, 6 April 2011 (UTC)

Suppose tablea contains the following data:

name	id
Tarja Halonen	1
Carl XVI Gustaf	2
Angela Merkel	3
Palpatine	4

tableb contains the following data:

name	id	aid
Finland	10	1
Sweden	20	2
Germany	30	3

tablec contains the following data:

bid	did
10	100
20	200

tabled contains the following data:

name	id
Helsinki	100
Stockholm	200

I want to be able to return every row in tablea, even though Palpatine can't be linked to any of the other tables. However, I don't want Germany to show up along with Angela Merkel, because it can't be linked to tablec or tabled, instead it should show up as null. JIP | Talk 17:44, 7 April 2011 (UTC)

igoogle side tabs

Hi. As of this morning, my fixes to remove the igoogle side tabs on firefox don't work (the tabs appear on the left side of my igoogle page, with idiotic things like "gadget shares" and "friends"--things I have no interest in, and waste my screen real-estate). Bad google. My greasemonkey script, having given good service for a year or so, now doesn't get rid of them, and the firefox plug-in doesn't work either. I'm going to have to switch to netvibes or similar if I can't sort this out. Any genius out there with an idea how to remove those stupid side tabs? Robinh (talk) 20:56, 6 April 2011 (UTC)

When you say "my fixes," are you referring to said greasemonkey script? Greasemonkey scripts are of course at the mercy of any code changes google may make to that page, in addition to any browser compatibility issues that may be introduced in major or minor releases. I did find this discussion here which you could try, though I haven't tested it myself. But...at the end of the day, as we say in web development 'if you didn't build it, don't bank on it!' --rocketrye12 ^{talk/contribs} 21:01, 6 April 2011 (UTC)

Hello Rocketry12. Thanks for this. Yup, my fixes are (or, were)-: the greasemonkey script, and a firefox add-on. Neither seem to work now. Why are google being such jerks? Robinh (talk) 23:16, 6 April 2011 (UTC)

(OP). Well, I spoke too soon. There is a "Super iGoogle" greasemonkey script, which has a patch, as of last week, that seems to work for me. Thanks, everyone! I can use igoogle again! Robinh (talk) 23:31, 6 April 2011 (UTC)

April 7

Intel Xeon E7-x870 Processors = Three Identical Models?

Good Morning, fellow RefDeskers!

  I'm looking at three Intel Xeon E7 Series processors; E7-2870, E7-4870 & E7-8870; and can't find any differences between them. All three of them have the following identical specifications. What is/are the difference(s) between these three processor models?

• 2.4GHz Base Clock Speed
• 2.8GHz Maximum Frequency
• 10 Cores (20 Threads)
• 30MB L3 Cache
• 6.4GT/s System Bus Speed
• 130W Maximum Thermal Design Power
• Advanced Encryption Standard - New Instructions
• Trusted eXecution Technology

  Thanks as always. Rocketshiporion♫ 00:00, 7 April 2011 (UTC)

A good question. According to our List of Intel Xeon microprocessors, the only differences I can see are the spec number, part number, and price. Presumably we've either got it wrong, or there is something we've missed.. Oops! see below...AndyTheGrump (talk) 00:28, 7 April 2011 (UTC)

On looking again: "28xx models support single- and dual-processor configurations, 48xx models support up to four-processor configurations, 88xx models support up to eight-processor configurations". AndyTheGrump (talk) 00:30, 7 April 2011 (UTC)

(EC) In other words, the various processors in this family differ in their hardware support for cache coherency in a symmetric multiprocessor configuration; probably related to the bus-snooping hardware. Higher-end SMP systems require more sophisticated (and higher-performance) bus-snooping logic to deal with more CPUs. Specific details from Intel: Intel Xeon Processor E7-... Families, and comparative Performance Factsheet, plus ISV quotes and anecdotes for various configurations. Nimur (talk) 00:38, 7 April 2011 (UTC)

BTW, are these Xeons using QPI yet? If not, I'd skip the Xeons and go for the "lower-performance" i7 series. All in all, I'd rather have fast memory and peripherals access, than fast CPU. Ah, yes, all are Westmere, and all new Xeons are finally using QPI. Nimur (talk) 00:43, 7 April 2011 (UTC)

Samsung Champ GT-C3303 Mobile-Phone

i have a Samsung champ camera GT-C3303 mobile. and a brand new computer with GB ram and 2.93 Ghz processor (core2due. with Samsung kies installed in it and i want to connect to internet using my mobile as modem connected with usb cable. i can connect to internet in my mobile but after connecting to PC i cant connect to PC. can u solve my problem thanks.. —Preceding unsigned comment added by 180.215.121.29 (talk) 06:39, 7 April 2011 (UTC)

I've seperated your question from the previous one, and added a title. Rocketshiporion♫ 10:03, 7 April 2011 (UTC)

Wanted: Content management system

Hello. I'm looking for a content management system for my organization's project management. The CMS must allow hosting multiple projects, each of them restricted to users who have been allowed to use the project. The CMS should support sharing of large files (his-res images and videos). Freedom, both in terms of use and cost are appreciated. Any suggestions are welcome. Thank you! 212.68.15.66 (talk) 10:03, 7 April 2011 (UTC)

How's WordPress? It's not strictly a blogging platform any more, and if installed/setup properly it should support all the things you listed above. 168.9.120.8 (talk) 11:55, 7 April 2011 (UTC)

With a few more requirements I could make a recommendation, maybe some more use cases? Are you looking for a traditional CMS with the capability to run a website, or more along the lines of a Digital Asset Management System? Or maybe the best qualities of both? As you might expect, Drupal is highly flexible, powerful and cost effective. --rocketrye12 ^{talk/contribs} 14:32, 7 April 2011 (UTC)

Advice on buying new PC

I'm buying a new PC. I've been advised that, given my requirements (word processing large documents with numerous jpegs), I don't need masses of RAM or hard drive space, but that I should focus on getting one with a fast hard drive and a good graphics card. I've now realized I don't know what this means, and anyway these criteria are often missing from on-line catalogues. What sort of h/d speed should I be looking out for and what sort of graphics card?--Shantavira|^{feed me} 13:39, 7 April 2011 (UTC)

Well personally I would think you do need a fair bit of ram for opening and working on large documents with lots images in them. Not excessive amounts, but at least 1GB if you want to have a number of large documents open at the same time, more if you want to multitask with other things. Hard drive speed shouldn't be an issue, any normal hard drive will be fast enough for just opening large documents (I'm guessing ~100MB each, since you didn't specify) and neither should the graphics card. You only need a really good graphics card if you're gaming or working with 3D stuff 82.43.90.38 (talk) 13:57, 7 April 2011 (UTC)

Also, the graphics card should be irrelevant for those requirements. The display and processing of large images may take advantage of a graphics card, but I have a hard time imagining the dinkiest integrated graphics card not being good enough. Paul (Stansifer) 14:14, 7 April 2011 (UTC)

The fastest hard drives in a desktop will be 7200 RPM and have a large cache (32 or 64MB most likely). You could also go with an SSD, but that could double the cost of your computer. With your requirements, you'll find that 4GB of RAM, any multicore processor, and a 7200 RPM hard drive will be more than fast enough for what you need with some room to grow. The graphics card isn't important in your requirements, but dedicated graphics won't pull from your CPU and RAM as much as Intel graphics. Look for Radeon, nVidia, and GeForce to signify a dedicated graphics card. If it says "integrated" or "Intel" graphics, and you can get a computer with similar specs for a similar price, go for a dedicated graphics card. It shouldn't matter which one if you don't do any gaming or major 3D or video work. 206.131.39.6 (talk) 16:27, 7 April 2011 (UTC)

Most modern operating systems (Ubuntu / GTK; Mac OS X ; Windows, since Vista) - all use the graphics card and GPU extensively during day-to-day operations. Even if you aren't a 3D-gamer or a CAD/engineering user, your GPU and graphics RAM will make a significant contribution to your user experience. Things like switching windows, maximizing, and even just running at high resolutions, all take advantage of a powerful GPU accelerator.

Truthfully, the PC market is segmented along price-points. If you buy a $300 system, it will underperform compared to a $700 system, or a $1500 system. It is almost impossible to put a high-end CPU in a low-end motherboard; so the "mix-and-match" approach is probably a bit misleading. Select your price-point, and accept the level of performance that the major system-vendors can provide you at that price. Then, if you decide you need a little more oomph in your RAM or hard-drive, upgrade (or downgrade) from the base-systems. You won't turn a $300 system into a competitive match for a $1500 system by swapping out two components, though.

A few rules of thumb: more RAM is almost always a good thing. A faster CPU is usually a good thing, but doesn't always make the most noticeable difference. For hard-drives, all modern drives have huge capacities: so if you want to "upgrade," select a faster hard-drive rather than a larger hard-drive. (Many portable computers trade down to a 5400rpm drive to save power, but consider bumping back to 7200rpm). For GPUs, more video RAM is probably a bigger improvement than any other GPU spec (unless you're a 3D gamer). However, the number of GPGPU-accelerated desktop programs in increasing, so you may find that your paint program or even your word-processor runs faster if you have a CUDA- or OpenCL capable card. Nimur (talk) 16:46, 7 April 2011 (UTC)

They use the graphics card, but they don't necessarily need a fancy one. I have Compiz's wobbly windows, and various other effects enabled on my netbook (the integrated graphics card is an "Intel Mobile 945GME Express", which I doubt is even fancy as integrated graphics cards go). Antialiasing is turned off, but the framerate is good. Paul (Stansifer) 17:08, 7 April 2011 (UTC)

The cheapest computer you can find at Best Buy will probably do the trick; will probably have a 2GHz+ dual core processor, 2-3GB+ of RAM, may or may not have a proper graphics card (nvidia or ati [not on the motherboard / intel] — worth the small extra cost). The future is now. ¦ Reisio (talk) 16:40, 7 April 2011 (UTC)

Is there any reason you believe a GPU on the motherboard is necessarily lower-performance than one on a breakout card? That's a very old-fashioned mindset from an era when circuit-packaging was still a very difficult task. Nowadays, even the highest-end GPUs and the fastest video-RAM can be integrated on the mainboard. And in some system architectures, shared main- and video-RAM yields higher performance than the old-fashioned, dedicated graphics RAM. (Essentially, a daughter-card forces a NUMA machine architecture, which is inherently lower-performance... again, an old-fashioned engineering tradeoff from the days when packing in 4GB of RAM was impossible). Nimur (talk) 16:49, 7 April 2011 (UTC)

He's not going to open it up (& a clerk at Best Buy wouldn't know :p), and most boxes with nvidia or ati stickers will have separate cards. You're talking about possibilites and I'm talking about cheap computers with three choices: intel, nvidia, ati. Onboard nvidia beats onboard intel, too. ¦ Reisio (talk) 17:00, 7 April 2011 (UTC)

I have to agree with Reisio here. IGPs are crap. Whether this will actually make a difference to the OP is another question but talking about possibilities is a bit pointless when a look in the real world will easily dispel any illusions. In fact while this isn't something I've looked in to a while, I'm pretty sure it remains the case that the crappiest same gen standalone card with limited memory primarily relying on system memory will still outperform the best IGP. It may be things are different with professional cards and systems, this isn't something I've looked in to, but these are expensive and not something the OP is likely to encounter. Note that in laptops the discrete GPU smay or may not be on the same board, however these aren't IGPs and will still communicate over the PCI-express bus and can be basically be thought of as standalone AFAIK. And I've never seen anyone use them in desktop systems anyway because they cost more but perform worse and create additional driver issues. P.S. I'm sure some small form factor PCs do use mobile GPUs, however these are likely to cost more then an equivalent more normal sized system so unless it's actually something the OP wants it's not worth considering. P.P.S. The fusion/APU concepts was supposed to shake things up a bit. Last I heard, this hadn't really happened but it looks like it's now starting to [10] [11] with the low end discretes starting to lose to high end Intel APU graphics. Probably things will get even more complicated when AMD higher end lines (like Llano) become available since Intel have struggled with GPU performance for so long (a quick search confirms I'm not the only one to think this) and Larrabee (microarchitecture) failed as a GPU but we're not quite there yet. Nil Einne (talk) 17:47, 7 April 2011 (UTC)

You don't say where you are, and you don't mention a budget, but if you're in the US, one approach is simply to go to Costco and buy whatever desktop or laptop PC that is within your budget. They have an excellent 90-day return policy if you find the PC is insufficient for your needs. Currently I think they only stock HP machines, but this may change over time. I've bought several desktop and laptop PCs there and returned a couple, and will continue to do so because of that return policy. Comet Tuttle (talk) 18:25, 7 April 2011 (UTC)

Changing window manager on Fedora 12?

Although I've used Linux for more than a decade now, both in home, university and work settings, I am still not exactly a high Linux guru. So I am writing here on Wikipedia to ask this question. I just found out that as I am running Fedora version 12, I have Compiz installed as default. I'd like to try it instead of the usual MetaCity, but I'm not sure how I would go about it. I first tried to run compiz directly, and this resulted in a message saying that this screen already has a window manager, and suggested I run compiz with the --replace option. However, I didn't want to do that, because I didn't know what would happen if Compiz somehow failed to start up, would I be left without a window manager at all. So I switched to a different screen with Ctrl-Alt-F2, which didn't even have X Windows running. Trying to run Compiz caused an error message about not having an X server, but trying to start an X server caused an error message about already having one (albeit on a different screen). So can anyone give me instructions on how to run Compiz on X Windows without necessarily having it become the default window manager, so I could always have MetaCity to fall back upon? And if Compiz does work OK, how to make it the default window manager? JIP | Talk 18:58, 7 April 2011 (UTC)

When you login, there is a session option that you can use to select what window manager you want to use. My experience with Fedora is that Compiz is tied heavily to KDE. The latest version of Fedora with all updates even replaced "KDE" with "KDE-Compiz". -- kainaw™ 19:07, 7 April 2011 (UTC)

I was not able to find such an option on my Gnome login screen. In fact, the login screen presents no information about sessions or window managers whatsoever. There is only an option to switch to a different user, or to change the session language or keyboard layout. JIP | Talk 19:33, 7 April 2011 (UTC)

Please ensure you installed the compiz manager (compiz-manager package) and not just compiz. As I noted, compiz is an add-on for both Gnome and KDE. So, the compiz package is what is needed for all types of compiz, not just the compiz manager. -- kainaw™ 19:47, 7 April 2011 (UTC)

I didn't have compiz-manager installed so I installed it. However, that didn't have any effect on the login screen, it is still presented without any information about sessions or window managers, as if the whole concepts didn't exist. JIP | Talk 19:59, 7 April 2011 (UTC)

In Gnome, Compiz doesn't have to do with sessions. Is there a "Visual Effects" tab in Preferences > Appearance? Does setting it to "Normal" or "Extra" make things fancier? Paul (Stansifer) 20:58, 7 April 2011 (UTC)

Gmail attachment

Resolved

Have written an outgoing message. I click on "attach file" and nothing happens. What can I do? I am running Vista & Firefox. Kittybrewster ☎ 21:06, 7 April 2011 (UTC)

Got it. Closed firefox & restarted it. Kittybrewster ☎ 21:40, 7 April 2011 (UTC)

audio

From the history section of the television article it says "by modulating the output signal of his TV camera down to the audio range, he was able to capture the signal on a 10-inch wax audio disc using conventional audio recording technology". What programs could I use on a Windows 7 computer to convert video data into audio, and back? 82.43.90.38 (talk) 22:10, 7 April 2011 (UTC)

Horizontal scrolling with scroll wheel on USB mouse in Windows XP

I would like to continue using the scroll wheel for vertical scrolling and for zooming (CTRL + <scrollwheel>), but in addition I would like to scroll horizontally. (I want to use it in Adobe Reader, Firefox, InternetExplorer and in open file system folders).
Is this possible in WindowsXP ?46.15.58.17 (talk) 22:47, 7 April 2011 (UTC)

Have you tried it? Do you have the latest driver from the vendor? ---— Gadget850 (Ed) ^talk 23:19, 7 April 2011 (UTC)

Have I tried what ? (I am relatively "computer illiterate" so please explain!) I bought an optical mouse, yesterday, and plugged it in. It works for vertical scrolling, but I have no idea what keys to press (or what ever to do) to make it scroll horizontally ?:-)
On the mouse it says: Model no.:HM5211
How do I find out who made it (and find their website address? to check whether or not I have the latest driver from the vendor? 46.15.58.17 (talk) 23:51, 7 April 2011 (UTC)