Wikipedia:Reference desk/Archives/Computing/2011 February 24

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Computing desk
< February 23 << Jan | February | Mar >> February 25 >
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.

February 24[edit]

Three-Disk RAID 1 Array[edit]

I want to make a RAID 1 array using three disks, and want to know if it's possible. Let me make clear that I am not referring to RAID 1E, where the capacity of the array is exactly half the capacity of all its disks combined. I mean an array of three or more disks, where each disk is a mirror, i.e. the capacity of the entire array is equivalent to that of just a single disk. Is there any hardware implementation of this anywhere? And I need compatbility with both Windows 7 and Ubuntu. Thanks, folks. Rocketshiporion 08:59, 24 February 2011 (UTC)

No, RAID 1 requires an even number of disks (usually two, but you can chain it together with RAID 0 to make RAID 10 or 0+1 volumes which are still an even number of disks). If you want to use 3 disks, then RAID 5 is probably your best bet.  ZX81  talk 10:29, 24 February 2011 (UTC)
Adding to my own answer, did you mean literally having the same disk mirrored THREE times? If so then yes that's technically possible, although I've personally never heard of a controller card that that'll actually support that.  ZX81  talk 10:39, 24 February 2011 (UTC)
Yes, I meant a array of three mirrors. If it hasn't been implemented anywhere in hardware, can it be done via software, and if so what are the software option(s)? Rocketshiporion 12:45, 24 February 2011 (UTC)
I have not seen anything like that. The primary reason is that RAID 5 is not really much slower than RAID 1 or RAID 0 when implemented through hardware. Examples of RAID 5 latency are engineered by writing tons of very tiny files - the weakness of RAID 5. Normal computer operations write a few large files. Further, RAID 5 will give you S*(N-1) capacity where S is the size of the drives and N is the number of drives. Assume you have 3 1TB drives. Mirroring all of them will give you 1TB capacity. Using RAID 5 will give you 1*(3-1)=2TB capacity. So, RAID 5 isn't worse than RAID 1 or 0 in practice and it is better when it comes to capacity. The demand to support more than one mirror just isn't popular. -- kainaw 13:37, 24 February 2011 (UTC)
It's not true RAID 5 isn't worse (in any way). If you use RAID 1, you will get double redundancy so can survive the failure of 2 discs whereas RAID 5 will only give effectively single redundancy and can only survive the failure of one disc. That's where the additional capacity goes. I would agree it's not done much, I suspect it isn't normally considered worth it for such a simple array, however there are more sophisticated schemes with double or even triple parity usually used with a much larger number of disks, see Non-standard RAID levels and RAID. I presume the OP is aware that RAID isn't a sustitute for good backup practices and the primary reason for the extra redundancy is so you can keep the array up with limited risk if one drive fails in cases where downtime is not acceptable (and with large arrays rebuilding may take quite some time). Nil Einne (talk) 17:55, 24 February 2011 (UTC)
Getting back to the question ZFS appears to be able to do triple mirroring in their fashion [1] [2] [3] [4]. (ZFS also supports triple parity.) Oracle/Sun does have storage appliances that support ZFS although I don't know if that's what you want. Bear in mind ZFS is still fairly new and fairly complicated (or at least I think fairly different from a lot of other file systems) and some of the claims made tend to be controversial. Nil Einne (talk) 18:38, 24 February 2011 (UTC)
While I never tried it out myself, I think I saw a reference in the mdtool man page that a software RAID1 on Linux can be used with three (or more) active disks. The usual three-disk configuration with RAID1 would be one hot-standby drive, though.
Also, I agree with Nil Einne that a RAID1 alone is not a replacement for a backup.
If you can live with the downtime, though, you can use a RAID1, especially with more than two disks, as a basis for your backup strategy: Power down the machine, remove one disk from the machine, label it with the current date/time and store it somewhere safe, insert a blank hard disk, power up the computer and have it re-build the RAID.
Assuming "somewhere safe" is that far away that a fire or other catastrophic loss of your main computer won't affect your backup, you can insert the backup hard drive in a new computer, boot from it, and add blank drives to restore the full RAID functionality.
My former employer backed up Windows Terminal Servers that way - they had no user data on them, and the backup software in use had some... well... issues with doing a full restore on bare metal. It was actually cheaper and less of a hassle to send a "disk monkey" into the server room each Friday evening. Of course, with Windows, you occasionally run into the issue that restoring a backup causes the machine to drop out of the Windows domain (because the machine security token had changed between the backup and the fallback), but that could be easily fixed by leaving and re-joining it. -- (talk) 22:02, 24 February 2011 (UTC)

Four-Disk RAID 1+1 Array[edit]

  I did not intend that a RAID array should be used as a subsitute for backup - in the event of data corruption, the RAID controller will happily replicate the corrupted data across the entire array! It is possible, using four disks, to make a RAID 1+1 array? I'm hoping this might be more feasible. The primary operating-system of concern is Windows 7 Professional, so that's the OS for which compatibility is a must. Rocketshiporion 06:53, 25 February 2011 (UTC)

Quake 3: Team Arena won't install on Win7 64bit, alternative games offering experiences?[edit]

I just tried to install my old copy of Q3:TA on my new Win7 64 bit machine and was mercilessly denied by Windows itself. Wouldn't even let me try! Anyways, what else out there can offer Quake-like gameplay on a 64 bit box? The Masked Booby (talk) 01:34, 24 February 2011 (UTC)

Are you sure you don't want to try to get Q3:TA to work? It's almost certainly possible. You might try the procedure described here. -- BenRG (talk) 01:50, 24 February 2011 (UTC)

opinion wiki[edit]

I am searching for a site where people may post there own opinion pieces, and articles. Is there such a site? -Shahab (talk) 05:30, 24 February 2011 (UTC)

Normally you would use a blog for that rather than a wiki. Why would you want other people to edit your opinion? Anyway,,,, etc. are popular blogging sites. There are some wikis like c2 which contain opinion pieces, but only on a certain range of subjects (like programming). And I guess there are some general wikis from a particular POV, like (talk) 06:33, 24 February 2011 (UTC)
Sites like Yelp contain people's opinions in the form of reviews of places. Chevymontecarlo 11:08, 24 February 2011 (UTC)

Blocking Websites in OS X at specific times?[edit]

Is there any way to block websites in OS Xu that meets the follow two criteria:

1) The website (such as is blocked in all browsers

2) The website is only blocked at specific times

Thank you for any help. --CGPGrey (talk) 10:54, 24 February 2011 (UTC)

Have you looked at the Parental Controls in System Preferences? There may be a setting somewhere in there. Chevymontecarlo 11:07, 24 February 2011 (UTC)
I have, but they only match need #1, not need #2 --CGPGrey (talk) 11:29, 24 February 2011 (UTC)
The easiest way to do this is probably through your router settings — most have some way of restricting sites, and many have ways of restricting times of the day. You might take a look there. --Mr.98 (talk) 12:37, 24 February 2011 (UTC)
It looks like SelfControl might do what you want. It seems a bit drastic, though; I prefer browser-based solutions like LeechBlock and Chrome Nanny. Paul (Stansifer) 16:12, 24 February 2011 (UTC)

How does a browser set up an HTTPS connection through a proxy?[edit]

How does a browser set up an HTTPS connection through an HTTP proxy? Does it use the CONNECT method to set up a tunnel to the remote endpoint (usu. port 443) and allow SSL handshakes to happen over the tunnel? Does it do it some other way? TIA. —Preceding unsigned comment added by (talk) 13:32, 24 February 2011 (UTC)

As far as I know, it's always done with CONNECT. That is certainly the normal way to do it. -- BenRG (talk) 02:38, 25 February 2011 (UTC)

Python problem[edit]

Hi, this may sound quite elementary but I'm having problems with it. For an exercise, I need to make this doctest pass:


 >>> nlist[2][1]
 >>> nlist[0][2]
 >>> nlist[1][1]


What does nlist[1][2] mean? When I create a list called nlist and just ask it to print nlist[1][2] it gives me an error saying that the string index is out of range. I think I'm having a blonde moment over something really basic and I'm gonna kick myself when it's explained. Thanks for any help (talk) 15:42, 24 February 2011 (UTC)

This would be valid syntax only if nlist was a list of lists. nlist[1][2] says "take the list of lists called nlist, retrieve the [1] element, and of that list retrieve its [2] in turn. In this case, it's either than nlist doesn't have 2 members (so nlist[1] is failing) or that the list nlist[1] has 0 or 1 members. What happens when you print nlist[2]? -- Finlay McWalterTalk 16:01, 24 February 2011 (UTC)
Since it says "string index", you should also note, if you don't know already, that strings are lists of characters. Paul (Stansifer) 16:06, 24 February 2011 (UTC)
Actually, they are not. Strings are an immutable sequence type (with the individual elements being characters). Lists are a mutable sequence type (with arbitrary elements). Things common to all sequence types (like indexing) work the same in strings and lists. --Stephan Schulz (talk) 16:10, 24 February 2011 (UTC)
That's true, but Paul is right to notice the error message; if one misreferences a regular list it says "list index out of range", but if one misreferences a string then indeed it says "string index out of range". So it looks like nlist is a string, but the code is trying to treat it as a list of lists (or a list of something else that's indexable). -- Finlay McWalterTalk 16:15, 24 February 2011 (UTC)
That syntax in Python is valid not just for a list of lists, but also a list of strings, a list of dictionaries, a dictionary of dictionaries, a tuple of lists, etc - basically any integer-indexable container class which holds an integer-indexable class (including objects of user-defined classes which have implemented the indexing operator). The syntax is saying "take the object nlist, and give me the contained item indexed by 1. Take this new object, and give me the contained object indexed by 2." If all you want to do is make the docstring pass, then something like:
nlist = {}
nlist[0] = {}
nlist[0][2] = 17
nlist[1] = {}
nlist[1][1] = 5
nlist[2] = {}
nlist[2][1] = 0
using dictionaries, will work. But that would be horribly confusing, as "nlist" would then not involve any lists at all, and likely won't help for the other use cases. -- (talk) 17:44, 24 February 2011 (UTC)
P.S. The "nlist" probably is supposed to be short for "nested list" - if you do a Google search for "nested list python" (without quotes) you should get a number of pages which discuss them. -- (talk) 17:49, 24 February 2011 (UTC)

OK, thanks for the help. I now have the following code:

#  Add your doctests here:
>>> nlist[2][1]
  >>> nlist[0][2]
  >>> nlist[1][1]

# Write your Python code here:

nlist=[[0, 1, 17], [0, 5], [1, 0]]

if __name__ == '__main__':
    import doctest

Still getting this error though:

File "__main__", line 3, in __main__ Failed example:






It expected the same as it got!! Surely this is not an error? (talk) 18:08, 24 February 2011 (UTC)

Your indenting in the test script is wrong; it should all be lined up vertically, but the first line isn't. -- Finlay McWalterTalk 18:19, 24 February 2011 (UTC)

SMS spam...but not quite[edit]

Today I received a "service message" on my phone containing nothing but the message "1 new message!" and the following link: centre-aligned on screen. Obviously my spam sense was immediately triggered. Visiting the link on a PC only displays an error "Sorry, your mobile device does not support this service". I still haven't clicked the link on my phone. My main question however is: how did this arrive as a "service message" rather than a normal SMS? And how did they centre-align the URL? It does not seem to be an MMS so it shouldn't be capable of any formatting. Is it even spam? And if so how is it accomplished? Zunaid 15:50, 24 February 2011 (UTC)

FYI: I had wget retrieve it with a user-agent faked for a Nokia phone, and it returns a page saying "Sorry, this link is not valid anymore. In case of any problem call 0861000983." is owned by some guy in Germany (who may have nothing whatever to do with this). Depending on the national dialling plan of whatever country you're in, that might be a perfectly innocent number, or an expensive number operated by scammers. -- Finlay McWalterTalk 16:12, 24 February 2011 (UTC)
Further: a Google search for that number quickly finds several posts by people claiming they've been billed on their cell plan for stuff they didn't order, including this. The name of the company matches the email address of the whois record for You may wish to call your own mobile company and tell them what happened. -- Finlay McWalterTalk 16:28, 24 February 2011 (UTC)
HOLY CRAP!!! That is scary indeed. I'd report it to MTN but I don't think they'd take it very seriously judging by people's experience in the link you provided.
So now the technical question: how did they make it appear as a service message and not a regular SMS? Zunaid 10:13, 25 February 2011 (UTC)
From reading the link provided it sounds like they are a Wireless Application Service Providers so probably have an agreement with your provider enabling them to send service messages. I don't know about the situation in South Africa but I know generally there's a lot of semi-dodgy stuff that goes on with those sort of things not to the extent of people being signed up for without actually doing anything but I think it's fairly common for people to sign up for a service without realising the actual cost (which tends to be quite high) or that it will be recurring until and unless they unsubscribe. I see a lot of ads on more questionable sites while browsing from NZ IPs with stuff like 'are you compatible' or 'read your fortune or 'know where you will die' where it says somewhere in the small print (can't remember in the actual ad or only when you click) you'll be charged $3 or $5 or something like that for recurring messages. Similarly with Malaysian IPs I see a lot of even dodgies ads telling me I can see what that girl looks like nude (well from the pictures, in lingerie) and clicking on them I find in the small print it'll be RM4 for recurring messages (I think they send some sort of wall paper). I think SMS reply confirmation from your phone is the norm but don't know how well it's implemented. Nil Einne (talk) 11:09, 25 February 2011 (UTC)

.net equivalent to system function of stdlib.h[edit]

does it exist? --Il signore degli dèi (talk) 16:41, 24 February 2011 (UTC)

stdlib.h is a pretty mixed bag of stuff; .NET should have equivalents for everything, but not in one place. For example, stlib's exit() call is Environment.exit() in .NET, but conversions functions like atof are in .NET's Integer and Double classes and the like. Which call in particular are you asking about? -- Finlay McWalterTalk 16:55, 24 February 2011 (UTC)
Surely the question is about system(3). --Tardis (talk) 22:25, 24 February 2011 (UTC)
.NET has a System.Diagnostics.Process class. Its Start method and its various overloads may be what you are looking for. Example:
. Regards, decltype (talk) 09:59, 25 February 2011 (UTC)

Wiki fundraising account[edit]

I have limited computer knowledge and want to set up a Wiki account to fundraise for slaughter bound horses. I already have a Wiki account. What's next? — Preceding unsigned comment added by Logan52286 (talkcontribs) 16:57, 24 February 2011 (UTC)

By "wiki account" do you mean "an account here, on Wikipedia"? If so, that's an inappropriate use of Wikipedia; this is an encyclopedia and contributing to that encyclopedia is the only purpose for which accounts here should be used. If you mean an account on some other wiki, you should discuss your goals with the other users of that website. -- Finlay McWalterTalk 17:01, 24 February 2011 (UTC)
In the UK, an account with Justgiving can be used for that sort of purpose. (You will need to demonstrate that you are genuine by setting up a charity or registering with HMRC first.) No doubt there are similar schemes in other countries.--Shantavira|feed me 17:54, 24 February 2011 (UTC)

where did the people who used to be on slashdot go?[edit]

I know there are places like reddit, but it's not as targeted. where did all the people who used to be on slashdot go, where do they leave the same level of insightful commentary on the same 'nerd' issues/news as a few years ago on slashdot? (talk) 17:40, 24 February 2011 (UTC)

Must be one of the following sites. The visitors to those sites are highly likely to also be the visitors of Slashdot. (talk) 10:39, 25 February 2011 (UTC)

How to reset USB's memory of ports in WinXP sp3?[edit]

When attaching an external HD to my computer via USB, I mistakenly turned it on and off too quickly, while the computer was recognising the new hardware. Now the thing is always recognised as malfuntioning hardware by USB. How can I reset USB, ie get USB to forget its memory of ports and what has been plugged in them? Thanks (talk) 21:27, 24 February 2011 (UTC)

Try turning off (unplugging) the external drive for a while and restarting Windows. (talk) 09:49, 25 February 2011 (UTC)
You can also go to the device manager and delete the USB device. Graeme Bartlett (talk) 10:26, 25 February 2011 (UTC)
... and if the device doesn't show up in the Device Manager, you might need to do these steps. (talk) 10:39, 25 February 2011 (UTC)

Thanks, I had a look at Device Manager. At the moment all I have attached to my computer via USB are an optical mouse, a printer, and a USB2 four port hub. Under Universal Serial Bus Controllers, there is listed: Unknown Device, USB Printing Support, USB Root Hub (repeated five times), VIA Rev 5 or later USB Universal Host Controller (repeated four times), and VIA USB Enhanced Host Controller. Is this unusual? The malfunctioning entry has disapeared. What do the different lines mean? Thanks (talk) 15:38, 25 February 2011 (UTC)

Update: the problem may have been a cheap external hub, probably the "unknown device" listed above, which was not working properly so that nothing happened when I plugged things into it. Thanks (talk) 14:51, 26 February 2011 (UTC)


How do you turn off "tap to click" on a laptop touchpad, running Windows 7? I've been to the control panel, mouse settings, hardware settings etc and I cannot find any options to disable it. (talk) 22:02, 24 February 2011 (UTC)

If it's well designed, the touch-pad should have its own settings for what actions are enabled (or at least mine does). You may be able to find the settings right in the control panel (that's where mine was). Otherwise, you can probably go to mouse properties, and disable the entire device (I know, not what you want to do), or, if worse comes to worse, uninstall the driver so that it can't work (again, not an ideal "solution"). Buddy431 (talk) 22:14, 24 February 2011 (UTC)
I always turn that off, too. What is the manufacturer and model number of the laptop? The most popular touchpad manufacturer is Synaptics — at least, based on how frequently I've personally seen their touchpads — and on at least one laptop of mine, I have had to download a Synaptics driver from the Synaptics website in order to let me turn off "tap to click". Comet Tuttle (talk) 23:10, 24 February 2011 (UTC)