Wikipedia:Reference desk/Archives/Computing/2010 February 23

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 23

Tabs help with latest version of firefox?

Hi there! I have a quick question regarding tabs in firefox. I'm using the latest version (3.6). It's pretty much the same as 3.5, except for one noticeable difference. When I right click on a link and select new tab, it opens a new tab immediately after the tab that is currently selected. In 3.5, it would open the new tab to the right of all currently open tabs. How do I change this to the 3.5 version functionality? I've checked Tools>Options>Tabs, and it doesn't have an option there. Thanks!--70.122.117.52 (talk) 01:11, 23 February 2010 (UTC)

Sorry for the very general advice, but if you type "about:config" in the URL bar and hit Enter, there will be a cute warning that you are voiding your "warranty", and then a vast number of little config items you can tweak. Type "tab" for the filter and see if anything looks useful. (I don't have 3.6, sorry.) Comet Tuttle (talk) 01:20, 23 February 2010 (UTC)

Yes, this new feature can be reverted to the 3.5-style "open new tabs at far right" by following Comet Tuttle's suggestion in the about:config manager. The key is browser.tabs.insertRelatedAfterCurrent. You can see some discussion on the topic at MozillaZine. Nimur (talk) 04:08, 23 February 2010 (UTC)

The Troublesome Ubuntu Netbook: Two Cases of a Linux Newbie Flailing Around

Situation:

I have a somewhat new Asus Eee PC 900. I got it off of Woot.com, and it has only a 4GB hard drive. Normally, this would REALLY suck, but the price made me just shrug it off. I removed the manufacturer's OS and replaced it with the Ubuntu 9.04 Netbook Remix.

Everything went really well, until about a month ago, when I totally, totally ran out of space. Every personal, nonsystem file was deleted. Nothing is on the computer anymore, except for what came with it, and Google Chrome.

But still, I only have about 200MB free at the most, after running "sudo apt-get clean", clearing the cache, and removing some Braille-teletype program thingy, which I don't need.

This is a huge problem now, because I keep getting nagged to install the new update, Karmic Koala. And I don't even have enough space for that. Or the updates to the other programs I have.

I need to:

1. Remove Firefox entirely. Every package that's there. I need all the space I can get.
2. Remove anything else unnecessary that you could possibly think of.
3. Fully uninstall OpenOffice.
4. It'd be nice if it could support Flash...well. Well enough to play videos on Youtube.
5. Find a way to clear some other stuff up. Just...pretty much anything that isn't essential to web browsing that got installed by default

My ultimate goals are to:

1. Have room to install update 9.10
2. Get it working to the point where I can play Youtube videos.

Thank you all so much for the help!

Saeb(talkjorn) 02:51, 23 February 2010 (UTC)

To remove all traces of something you can purge it. From a terminal run "sudo apt-get purge [packagename]", or if you use Synaptic I think it's a right-click and "Mark for Complete Removal". You can go through a list of installed packages in Synaptic and remove everything you don't need. In addition you can head to System -> Administration -> Computer Janitor which will suggest unused packages to be removed and other changes to save space. Be very careful, though, as this Janitor has a reputation for being a bit...overenthusiastic. Once you're done removing you can "sudo apt-get install flashplugin-nonfree". However this may not be required as YouTube is betatesting an HTML5 player which would be supported by Chrome without any additional plugins. It's worth a look if you don't mind losing fullscreen. Xenon54 / talk / 03:01, 23 February 2010 (UTC)

Why choose something as big as Ubuntu - even Ubuntu Netbook Remix is quite hefty for a small disk and small memory system. There are plenty of lightweight linux distros available which would still play YouTube videos, if that's all you want to do with it. Astronaut (talk) 09:28, 23 February 2010 (UTC)

Breaking an Easy Cipher

In an effort to teach myself more about cryptography, I came across the following cipher in a paper. Let $k_{1},...,k_{100}$ be one hundred randomly and independently chosen round keys of 128 bits length each. They have nothing to do with each other. So one round of the cipher is that I take the input (128 bits in binary) xor it with the round key and then find its multiplicative inverse in the Galois Field GF(2^128) with the modulus known (zero maps to zero). I do this 100 times and I get the ciphertext. The paper says that it is easy to break it. I searched around and I find the same answer but no one actually says how do you break this or why is this easy. Can someone please point me in the right direction? Any hints? Is it really obvious how to break this cipher? How would you do this? Thanks! 174.29.98.151 (talk) 07:24, 23 February 2010 (UTC)

Hint: think about interesting chosen plaintexts that you can send through the cipher. 75.62.109.146 (talk) 07:39, 23 February 2010 (UTC)

Depending on what you mean by "I do this 100 times", it can be real easy or mostly easy. If you mean you perform 100 iterations on each character in the plaintext, you have created a 1-to-1 cypher. A simple character replacement algorithm will easily work out what character maps to each element of the cypher text. If you mean that you perform an encryption using k1 on the first character, then use k2 on the second character, and so on... it is the same thing. Step 1, guess how many keys there are. Assume that I guess there are 20. I break the code into 20 chunks. The first contains characters 1, 21, 41, 61, etc... Since they all used the same key, I can do a simple frequency analysis to guess at the 1-to-1 map. If I don't find anything, I try a key length of 21, then 22, then 23. It will take a matter of seconds to analyze each key length. So, it will be trivial to step up to a key length of 100 and then get good values for a frequency analysis of every 100th cypher-text character. -- kainaw™ 07:49, 23 February 2010 (UTC)

PS3 component cable with second stereo audio outpht

My Playstation 3 will shortly be moving to a room where the TV ([1]) doesn't have an HDMI input, the TV does have a component video input though.

There's loads of PS3 component cables out there - however, I'll sometimes want my PS3's audio to come out of the TV's speakers, and sometimes from a set of external speakers, so I'm looking for a component cable with a second stereo audio out. Does anyone know if such a thing exists?

I can find loads of PS3 SCART cables with extra audio outs (e.g. [2], [3]) but none for component, and from what I've read online, with SCART I won't be able to get 720p video out?

Cheers, davidprior ^t/_c 10:57, 23 February 2010 (UTC)

SCART doesn't do 720p - standard (old) definition only.

I can't find a dual ouput component cable either - cable splitters exist and are cheap - eg try search "stereo phono splitter" - eg two of these would do [4].

Using these actually halves the output impedence - but that really shouldn't be an issue (the input resistances are usually very high) - it's unlikely that there will even be any audible difference.87.102.67.84 (talk) 16:12, 23 February 2010 (UTC)

If you really need one you can use a switch box [5] or maybe you have an old amplifier you can use to switch the audio (an old one won't even need to be turned on to operate)87.102.67.84 (talk) 16:24, 23 February 2010 (UTC)

It may not be necessary to split them. Many televisions and receivers have both audio-in and audio-out. You can run audio into one and then run from the out to the in of the other. It will allow you to have one, both, or none of the audio devices active. -- kainaw™ 16:23, 23 February 2010 (UTC)

Accorfing to the manual linked above there is a RCA stereo out which can be either fixed or variable output (p36) - I guess fixed output (not related to TV volume) is the one needed - set in the TV's menu (p14) Shown on p8 (called AV OUPUT).

This would mean that the TV would have to be on though...87.102.67.84 (talk) 16:33, 23 February 2010 (UTC)

Thanks everyone for their input. At least now I know that I ain't going mad when I couldn't find this. I think I'm going to go the splitter route - I was unsure if it would degrade sound, but hopefully it'll not be noticeable.

Should've said that I don't want to use the TV's audio out; one of the reasons for wanting to play sound though the speakers instead is that I use my PS3 to listen for the radio (via PlayTV) and want to be able to do this with the TV off - partly due to power use, but also 'cos when tuned to a radio station, PlayTV displays a fixed image on the screen... I like BBC 6 Music but not enough to burn their logo onto my telly ;-)

Thanks again, davidprior ^t/_c 20:54, 23 February 2010 (UTC)

Good hosting for static files?

I have about 10 gigs of personal files I want to be able to access from any computer with internet. Low bandwidth costs-- just me. No dynamic content, just a simple password-protection http .

Let's assume I don't want to host the files out of a home server. Who should I go to for cheap static file hosting. --CreedShandor (talk) 11:05, 23 February 2010 (UTC)

Does you ISP have an inclusive hosting package? Astronaut (talk) 11:20, 23 February 2010 (UTC)

Dropbox or something similar might do what you want. AndrewWTaylor (talk) 11:44, 23 February 2010 (UTC)

If it's 10GB that you only need a little bit at a time from (access more important than total redistribution), checkout http://nearlyfreespeech.net/ ¦ Reisio (talk) 11:59, 23 February 2010 (UTC)

From that site: "Storage charges apply at $0.01 per megabyte-month" -- so hosting 10GB would be $100/month, I reckon. But bandwidth is pretty cheap. Compared to Amazon S3, another scalable host, 10GB would be $1.50/month. 198.161.238.18 (talk) 19:29, 23 February 2010 (UTC)

See File hosting service and Comparison of online backup services for some options. 198.161.238.18 (talk) 19:49, 23 February 2010 (UTC)

Skydrive is free and holds up to 25 GB.--Drknkn (talk) 06:17, 25 February 2010 (UTC)

Mediafire —Preceding unsigned comment added by 82.44.55.204 (talk) 13:52, 25 February 2010 (UTC)

Mp3 to MIDI converter

Is there any free full version mp3 to MIDI converter? Thank you! —Preceding unsigned comment added by 83.183.172.203 (talk) 12:56, 23 February 2010 (UTC)

Midi doesn't work in the same way as MP3. MP3 is an audio recording that has had unneeded data removed to allow it to be compressed. Think of it as a recording of a person reading something. Midi on the other hand is a language to tell a soundcard to play specific sounds at specific pitches, think of it more like a script which is interprited by a person who then reads it. It would be possible to convert MIDI to MP3 quite easily although it would sound different depending on the sound hardware that played it, but it is probably impossible to automatically convert music to midi, especially if it contains vocals. Gunrun (talk) 13:20, 23 February 2010 (UTC)

The first result you get when you search on google for "convert mp3 to MIDI" explains why it isn't easy, but gives an example of of to use a program that tries to do this. Cheers, davidprior ^t/_c 13:54, 23 February 2010 (UTC)

MP3 relates to MIDI as BMP to SVG. You can always convert from SVG to BMP, and from MIDI to MP3, but the other direction is (in almost all cases) impossible. --Andreas Rejbrand (talk) 13:57, 23 February 2010 (UTC)

This has been discussed here before: see [6][7][8][9]. The consensus was as above: it's very difficult; there is some software that claims to do it but the software is error-prone and not very effective. --Normansmithy (talk) 16:49, 23 February 2010 (UTC)

Playing a sound in C/C++

Is it possible to play a sound that is stored in the computer's RAM instead of hard disk? Is there a function like this?

void playSound ( int * sound, int length, int sampleRate ) ;

I'd need suitable functions for both Windows and Linux. 88.148.207.106 (talk) 13:57, 23 February 2010 (UTC)

Simple DirectMedia Layer can do this on both platforms. Here is an example, and in general, see the "audio" section of this list of SDL APIs -- Finlay McWalter • Talk 14:16, 23 February 2010 (UTC)

If you're looking for the native sound API, for Linux you do (or did) just open /dev/dsp for write, configure it with the SOUND_PCM_WRITE_BITS, SOUND_PCM_WRITE_CHANNELS, and SOUND_PCM_WRITE_RATE ioctls, and write raw PCM data at it. That breaks on platforms which use PulseAudio (such as Ubuntu Karmic), as PulseAudio owns /dev/dsp exclusively; in this case you either use the PulseAudio API or use the padsp wrapper around your /dev/dsp software. -- Finlay McWalter • Talk 15:10, 23 February 2010 (UTC)

On Windows, it can be done with PlaySound[10], or waveOutOpen/waveOutPrepareHeader/waveOutWrite[11]. Alternatively it can be done with DirectX. --Normansmithy (talk) 17:01, 23 February 2010 (UTC)

To make explicit something that was implied by Finlay McWalter and Normansmithy, the C and C++ languages do not have any sound playing functions as part of the languages themselves, or as part of the C Standard Library, or the C++ Standard Library. Sound playing is platform-specific and you have to use a sound library (plenty having been mentioned above), and every API will differ a little. Comet Tuttle (talk) 17:25, 23 February 2010 (UTC)

I know it doesn't answer OP's question, but strictly speaking it is possible to "play a sound" with ANSI C (in a console application) using printf("\a");. (Disclaimer: I don't have a C compiler available to test this specific code, but I have used the concept before.) According to K&R2, '\a' is an "audible alert". Mitch Ames (talk) 01:17, 24 February 2010 (UTC)

Faster than Gigabit, cheaper than Fibre Channel?

Are there any technologies that are faster than Gigabit, but cheaper than Fibre Channel? My Gigabit network isn't fast enough for multiple video streams, and Fibre Channel is crazy expensive. 10ge looks like a contender, but still seems rather pricey. Is there anything else? While not yet a shipping product, could Light Peak be used to networking? It seems to promise low cost/high throughput --70.167.58.6 (talk) 16:52, 23 February 2010 (UTC)

List of device bit rates has a pretty comprehensive list of networks and communication channels of various kinds. Light Peak does seem like a candidate, but it's not available right now; Light Peak's inventors clearly intend it for a consumer price range (not the eyewatering costs of Infiniband or Fibre Channel). In the short term, if I was setting up something like a high-quality video-server I'd consider multiple point-to-point ethernet connections (that is, that the server has several Gigabit ethernet cards each with a single cable direct-connected to a single client without an intermediating switch) - that way you get a dedicated 1Gbit connection to each client. -- Finlay McWalter • Talk 17:15, 23 February 2010 (UTC)

Or link aggregation. -- Coneslayer (talk) 17:40, 23 February 2010 (UTC)

I can't find enough (heck, any) hard technical information about how Light Peak actually works, in order to answer the "could Light Peak be used to networking" part of your question. It seems like it's a single-initiator multi-drop wire protocol, suggesting that only the PC can be master. If that's true, I guess you could have an intermediary box (like you can in USB) to build a naff point-to-point connection; for a genuine network, then surely 10 Gigabit Ethernet will make much more sense. -- Finlay McWalter • Talk 18:27, 23 February 2010 (UTC)

(off) As I understand it light peak is just a wire (or pipe) for connecting devices - and will allow multiple communication protocols to be sent down it (eg ethernet, hdcp video etc) - it's definately physically two way, daisy chaining of devices was also mentioned. Beyond that, as you say, information is scarce - I assume the different protocols would be packetised and sent then un-packetised. I've no idea if there will be a new 'light peak' protocol as well. It's a complete mystery how the wildly different signal protocols will be collected together on a device. (I'm guessing something like optical ethernet)87.102.67.84 (talk) 21:53, 23 February 2010 (UTC)

Are you sure that its a case of Gigabit Ethernet not being fast enough per ce, rather thanyour current network card/switch/etc having a GbE interface but not being able to actually handle anything like a gigabit of data (from every port) each second, or some other link in the chain such as the video streaming app not being able to feed the data to the network card fast enough (bearing in mind that it'll have to do this while your CPU also has the overheads described here

In other words, would a better Gigabit network be what you're looking for, rather than a different standard entirely?

The reason I ask is that even for good quality video, bandwith requirements are a lot less than 1Gbit/sec - DVD is 10.5MBit/sec, a Blu-Ray's is 36 (this is a 1x player's read speed - so an upper limit), DVB-T uses about 5 for an SD television channel. So a network capable of 0.75 Gbit/sec after overheads could carry many simultaneous streams (71/20/150 at the bitrates above). If that still ain't enough, you could use link aggregation as suggested above - or if the multiple streams are actually multiple instances of the same stream, is some form of Multicasting an option?

Hope this helps, davidprior ^t/_c 21:45, 23 February 2010 (UTC)

He may not be dealing with compressed video, as you referenced; uncompressed DVD quality video weighs in at about 180 Mbit/sec and high definition video would of course be much higher than that. Just look at FireWire standards (mostly driven by digital video needs) and you find a bandwidth of 800-3200 Mbit/sec in use on that format. More to the point (you made a good one), a lot of gigabit equipment does peak out at 200-300mbit/sec without proper optimization. Using jumbo frames on all devices (a property of the network interface), a good quality Gbit NIC and switch, and an application with low overhead (so that the CPU does not limit it) is the only way to get close to the theoretical limit of 1000Mbit/sec. It would behoove the question asker to take a close look at exactly how much bandwidth and CPU time is in use (through the Windows task manager or similar tool for your OS) and establish exactly what is limiting you, then simply add more. The cheapest solution, if you are only a few multiples away from your goal (say 2 or 3 Gbit/s) is to simply add more Gbit NICs and use the OS tools at your disposal to 'bond' them so they can aggregate the bandwidth. A higher quality switch may also be needed, but these are had for far far less than the price of 10Gbit or other more obscure hardware. --144.191.148.3 (talk) 15:56, 24 February 2010 (UTC)

It may be helpful to specify what sort of circuit length you require Nil Einne (talk) 01:52, 25 February 2010 (UTC)

Gigabit is fast enough for dozens of video streams, even at 10 megabits per second each. Are you sure it is your network that is limiting you and not your storage? You may have to change your style of presentation to multicast on a schedule, rather than video on demand. Graeme Bartlett (talk) 20:53, 26 February 2010 (UTC)

What is the metric measurements for a 1/2" CMOS HD sensor?

What would be the length and width (in mm) of a 1/2" CMOS chip with a resolution of 1920x1080 pixels? --70.167.58.6 (talk) 17:46, 23 February 2010 (UTC)

Well, if it's half an inch, that would be 12.7 mm, and if that's the height, the width would be 16.93 mm, assuming it's 4:3. Belisarius (talk) 19:03, 23 February 2010 (UTC)

I think we can assume from the pixel dimensions that it's a 16:9 sensor. But generally speaking, "inch-sized" sensors don't have the direct relationship between their nominal size and their physical dimensions like you would expect them to. See Image sensor format#Table_of_sensor_sizes. The typical size listed there for 1/2" is also for a 4x3 sensor, so it's not going to be quite right for 16x9, either. Usually the physical size of a sensor is listed in the device's specifications. -- Coneslayer (talk) 19:15, 23 February 2010 (UTC)

Both kodak and aptina have detailed specifications of their sensors on their sites - but don't appear to make the type you mention. Kodak seems to make that type - but I can't find any tech documents - anyone else better at finding stuff on canon's website?87.102.67.84 (talk) 21:17, 23 February 2010 (UTC)

Is it a safe assumption that all sensor chips have the same aspect ratio (4x3)? Or because the one I'm asking about is for an HD video camera, it would the sensor be 16x9? —Preceding unsigned comment added by 70.167.58.6 (talk) 22:12, 23 February 2010 (UTC)

16:9 sensors definately exist - ie 1920x1080 pixels - I found them for 2/3" and 1/3" size formats eg see [12] 87.102.67.84 (talk) 22:40, 23 February 2010 (UTC)

I've written the company (Sony) and I hope they reply. It's funny that so much importance is placed on sensor size, yet I can't find out the actual measurements! --70.167.58.6 (talk) 18:17, 25 February 2010 (UTC)

Another very common DSLR sensor ratio is 2:1. --Phil Holmes (talk) 14:48, 24 February 2010 (UTC)

Most DSLRs are 3:2, after the 24x36mm image area of 135 film. -- Coneslayer (talk) 14:53, 24 February 2010 (UTC)

how do I get to a different iPhone app store? (nationalized)

I got my phone in France, and so the store, notably the top apps in each catagory, on the phone is set for France... is it possible for me instead of seeing all French billboards and discriptions, to somehow set it to the US so that I can see the things that are relevant for me, e.g. radio apps for the US market rather htan the French market, etc. THanks! 84.153.231.223 (talk) 21:43, 23 February 2010 (UTC)

[13] [14] Sounds like you need to connect to iTunes on a PC (or mac!), make sure that's for the right country, and then sync the jesus iPhone to the PC - which then makes PC iTunes country the default on your phone. Did that make any sense, or even work??87.102.67.84 (talk) 22:02, 23 February 2010 (UTC)

cameras that take 720p or 1080p photographs

hey guys are there any cameras that take 1280x720 or 1920x1080 photographs? thanks--Copr89 (talk) 23:15, 23 February 2010 (UTC)

Most modern digital cameras have resolutions higher than that, and it is trivial to crop or downscale a photograph, so if you simply need to produce images of those sizes, almost any camera will do. --Andreas Rejbrand (talk) 23:48, 23 February 2010 (UTC)

Also, it's incorrect to attribute 720p and 1080p to still pictures. Our articles state that they are for progressive scan lines and frame rates. With cameras, it's correct to talk about resolution and the number of (mega)pixels needed to store uncompressed images at those resolutions. Sandman30s (talk) 06:53, 24 February 2010 (UTC)

Monitor resolution stuck at 640x480

A few years ago I bought a PC that came with Windows XP. About 6 months ago I did a major hardware upgrade where I changed almost every component. I guess Windows XP didn't like that and wouldn't let me use it until I reactivated it. Instead of going through all that, I've been using the Windows 7 Release Candidate up until now. But that's expiring soon, so I wanted to go back to using XP. I managed to reactivate it, but now it seems the screen resolution is stuck at 640x480. I go into display settings and drag the slider to 1280x1024 and hit apply, but nothing happens. I've tried installing the latest video card drivers (for an Nvidia geforce 8800 GTS, if it matters) and updating the monitor drivers (for an HP vs17e). Is there anything else I can try, or should I give up and buy Windows 7? (Which in this economy, I'd really rather not do.) Digger3000 (talk) 23:29, 23 February 2010 (UTC)

Do you have a KVM switch between the PC and the monitor (or something else like that)? KVM switches sometimes eat the VESA signalling where a monitor identifies itself (and its capabilities) to the graphics card, leaving you with a "Generic Monitor" reported rather than what you actually have. -- Finlay McWalter • Talk 23:37, 23 February 2010 (UTC)

I don't have the thing you mentioned, but there is an adapter type thing between the monitor and the video card. I guess I can't plug my monitor directly into the card. Digger3000 (talk) 23:45, 23 February 2010 (UTC)

What kind of "adapter"? Perhaps a white DVI adapter? Take a look at [List of video connectors]] and see if anything catches your eye. Astronaut (talk) 04:06, 24 February 2010 (UTC)

Yes, a DVI adapter. Digger3000 (talk) 04:24, 24 February 2010 (UTC)

If you can drag the resolution to 1280x1024 in XP then this means that the computer has recognised the available resolutions. So the monitor probably has identified itself. Is there a nvidia "control panel" where you can set resolutions? This might be overriding the windows method.?87.102.67.84 (talk) 11:17, 24 February 2010 (UTC)

There's an example of using nvidia's control panel instead of windows here [15] 87.102.67.84 (talk) 13:09, 24 February 2010 (UTC)

There is an Nvidia Control Panel, but I can't change the resolution from there either. Digger3000 (talk) 20:07, 24 February 2010 (UTC)