Wikipedia:Reference desk/Archives/Computing/2007 June 23

From Wikipedia, the free encyclopedia

Computing desk
< June 22 << May | June | Jul >> June 24 >
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.


Contents


[edit] June 23

[edit] SD card packs?

Anyone know where I could buy packs of SD cards at a reasonable price? I'd seen a two pack of 2GB cards, for 50$... compare to the 2GB card itself only costing 13$. A pack of five or ten for around 50 or 100$, respectively, would be ideal, but anything would be nice to see. -- Phoeba WrightOBJECTION! 00:00, 23 June 2007 (UTC)

$10 for a 2GB SD card? I really wouldn't use it on any critical work as it may well have been fake ones using chips that are marked as "bad" (ie. can't stand for many rewrites). --antilivedT | C | G 01:11, 23 June 2007 (UTC)
13-16$. A microcenter near where I live sells cheap cards, I've never had a problem with them though. They also sell USB sticks, they just have them in baskets near the counter. I think they sell them at cost or with very little profit, since they put their name on them, I'd assume that they're advertising. Anyway, I'm talking about buying in packs, which almost always should make things cheaper, not more expensive -- Phoeba WrightOBJECTION! 03:47, 23 June 2007 (UTC)

[edit] Fastest Processor

What is the fastest processor made or that is going to be made? 68.193.147.179 00:18, 23 June 2007 (UTC)

Nobody can predict the future; but the industry is shying away from faster clock speed in favor of parallel architecture. At present, you can probably commercially obtain processors at a maximum speed of maybe 3.6 or 3.8 GHz. You should also note that in general, faster clock speed does not even mean faster computation; it most certainly does not directly equate to better overall performance. Nimur 00:46, 23 June 2007 (UTC)

Yea, the 3.8 GHz don't have that much advantage over 2.8 GHz. Nowhere near a 25% increase in speed. The limit of the current technology is almost at its limits in terms of speed and heat, which is why they're exploring the multi-cores. Of course, the theoretical limits would probably be somewhere around the speed of light, I would imagine, but who knows if we'll ever even get there, much less when. --Wirbelwindヴィルヴェルヴィント (talk) 02:48, 23 June 2007 (UTC)
At the moment: []http://www.newegg.com/Product/Product.aspx?Item=N82E16819115021 probably this]. I expect the market to shift to quad-cores within 5 years, and perhaps increase to even more after that. -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 03:06, 23 June 2007 (UTC)
Last I heard, Intel is getting ready to introduce 8-core processors sometime Soon™, and they've also been working on a quad core laptop that I think should be out this year. I expect now that core speed in GHz is going to become increasingly irrelevant, and the main figures to worry about are going to be number of cores. I read something about intel already testing out a 64-core processor setup; although it had very few performance gains over an 8 core, the fact that they're working that far ahead is amazing. I will eat my hat if, within five years, non-multi core processors are still in popular use. Of course, I don't actually own or wear hats, but it's the thought that counts -- Phoeba WrightOBJECTION! 03:39, 23 June 2007 (UTC)
Well yeah, I'm sure they can pack a lot of cores into one processor (the PS3's Cell processor has 8 "Synergistic Processing Elements") and I'm sure gamers will keep buying them up (look at how expensive the processor I linked to is). I think at this point the slowdown will be with software manufacturers forced to keep up with the ever-expanding amount of cores. Widespread adoption of any technology will only happen when a good chunk of new software supports it, and computer manufacturers, often catering to the "i don't want to know how it works just make it work" crowd, don't want to adopt a technology that might not be fully stable and compatible with various applications. -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 07:26, 23 June 2007 (UTC)
I agree. Even though I know how to program to take advantage of multi-processor systems, I don't think I would buy something above a dual core at this point. Only best case scenarios can increase performance much, and it's not actually like dual core effectively double the processing power. --Wirbelwindヴィルヴェルヴィント (talk) 09:14, 23 June 2007 (UTC)

The current best performers in the industry standard SPEC single-core benchmarks are (higher is better, picking the best performers from different processor lines):

IBM POWER6 (4.7 GHz) at 21.6 int, 22.3 fp
Intel Core 2 Extreme X6800 (2.93 GHz) at 18.5 int, 16.2 fp
Intel Xeon 5160 (3 GHz) at 18.5 int, 15.6 fp
Intel Itanium 2 (1.6 GHz/24 MB) at 15.7 int, 18.1 fp
AMD Opteron 2222SE (3 GHz) at 14.9 int, 15.2 fp
Fujitsu SPARC64 VI (3 GHz) at 11.3 int, 12.4 fp

The integer (int) score is the most relevant for most tasks, but the floating point (fp) score is important in many kinds of scientific applications, and I suppose current games need some fp performance as well. These are SPEC peak rates, reflecting the performance reachable on a single core after performance tuning. Intel actually has a published score of SPECint peak 20.8 for Xeon X5365 (3 GHz), but for whatever reason they don't show a FP score for that one. 84.239.133.38 10:37, 24 June 2007 (UTC)

[edit] How do you create new (winnable) sudoku levels?

Or levels in any puzzle, for that matter. The methods of solving sudoku are complex, so I'm curious how one would go about generating a playable level? Much help appreciated! Xhin Give Back Our Membership! 04:26, 23 June 2007 (UTC)

This is a better question for the mathematics desk, i'm sure they have a simple enough method for it, but I use GNUdoku to generate sudoku puzzles when I'm in the mood for it.

I can think of one approach:

1) Write a program to solve a sudoku puzzle (fairly easy).

2) Start with a finished sudoku grid, then remove numbers and run it through the solver until multiple solutions are found. Then back up one step and remove a different number instead. Repeating this process, you will eventually end up with a puzzle which has only one answer and is solvable, but where removing even one more number would make it have multiple solutions. StuRat 04:51, 23 June 2007 (UTC)

I think you overestimate how "complex" sudoku is. See computational complexity theory; you will find that Sudoku is really just a set of linear equations (off-hand, I would imagine it's probably solvable in time proportional to the grid size. Just because it is "difficult" (for a human) does not mean it is "complex." Nimur 06:35, 24 June 2007 (UTC)
Actually, the article on Sudoku states that it is NP-Complete, and thus it is not likely to be solvable in time proportional to the grid size. Note that Sudoku does not have linear equations, but requires solutions to be integers. 213.237.10.244 13:03, 27 June 2007 (UTC)

Thanks, everyone. Very helpful. Xhin Give Back Our Membership! 07:02, 25 June 2007 (UTC)

[edit] programming

i wanted to design an application that istructs the user to enter inputs for instance marks or income for a period of time then later computes the total.which programming langage does it best suit? and what are the links for the tutorials? unsigned by 80.255.43.45.

Almost any computer language could do this job. You need to be able to display and receive input from a user, and store data in a file or database and retrieve it. Most computer languages have these facilities. Do you know what platfrom you want to run it on, could it be just one PC, or a network of computers around the world? If you have multiple people entering data it is much more tricky! Is this your first program? One simple way is to use a spread sheet such as Lotus 123 or Microsoft Excel, and store the marks for one person in a row, each person can be a separate row. GB 07:25, 23 June 2007 (UTC)
Why does it have to have a database? The user's question is very similar to the tasks I've send that students do when learning Delphi (Object Pascal) for example - and they don't use a database. They simply use variables. Yup, that means when you terminate the application all your data is lost - of course you can then store the variables in the registry or an .ini file or something. But my point is, databases are definitely one of the more advanced things to learn for novice programmers - I'd say let them learn variables first. Rfwoolf 10:20, 23 June 2007 (UTC)
Spreadsheets!

[edit] A Windows analouge to wget?

Is there a something included in windows that does something similar to the unix tool wget? That is, downloads a file from somewhere on the internet without using a browser (either command-line or GUI)? I realise that there are versions of wget for windows, but I'm looking for something that is included with the installation. 217.174.67.69 13:01, 23 June 2007 (UTC)

  • wget will work for you on windows. Get cygwin and install wget package. --BACbKA 13:06, 23 June 2007 (UTC)
    • Yes, I realise that, but I meant something that came with windows, as I wrote in my question. If I just wanted wget, I'd just use one of the native windows versions 217.174.67.69 13:13, 23 June 2007 (UTC)
      I'm sorry, but you are going to have to either use the command-line ftp client, the command-line tftp client, or perhaps you could try to do something with the Windows Script Host. All three are often used by malware to download things; I've personally seen one use the ftp client, and I heard of malware using the other two. Unlike most Linux distributions, a default Windows install doesn't come with a lot of useful utilities. --cesarb 13:26, 23 June 2007 (UTC)
Ok, thanks 217.174.67.69 13:41, 23 June 2007 (UTC)

[edit] Upgrade RAM or Video Card?

I run Windows Vista on a Gateway machine with 1GB DDR2 RAM and an Intel Graphics Accelerator 950. I want better performance in general to photoshop, edit video, watch movies and even just run Aero smoothly. Which should I upgrade first, the RAM or video card? Any suggestions? Thanks. 68.39.175.57 15:23, 23 June 2007 (UTC)

I suggest you upgrade to at least 2GB, re-evaluate the applications you're running, and then probably upgrade your video card as well. -wizzard2k (C-T-D) 15:13, 23 June 2007 (UTC)
Nothing runs Vista smoothly, it's way too needy. You might as well go with a video card, since even 4GB of RAM (the max a lot of motherboards will support right now) won't help Vista run quickly. -- Phoeba WrightOBJECTION! 15:15, 23 June 2007 (UTC)
My point was mainly for his applications, not really Aero (I think thats a lost cause for now). I wouldn't even do those things on XP with only a gig of ram. -wizzard2k (C-T-D) 15:23, 23 June 2007 (UTC)

I know that the Intel Graphics Accelerator actually saps physical RAM from your system since it doesn't have its own dedicated RAM. For this reason I thought maybe upgrade video first. 68.39.175.57 15:23, 23 June 2007 (UTC)

Upgrade the RAM first. You really want at least 2GB of RAM for Vista to start with. Photoshop obviously likes lots of RAM, as does video editing. A video card actually makes little or no difference to programs like Photoshop anyway. --jjron 17:11, 23 June 2007 (UTC)

[edit] Political Blogs on the Internet.

Which political blogs (both liberal and conservative) are the most trafficked on the Internet.

It depends on how you define those two terms, which are not mutually exclusive. AllanHainey 11:32, 25 June 2007 (UTC)
It also depends on how you define internet traffic... Nimur 06:30, 26 June 2007 (UTC)

[edit] Mapquest prints

When I try to print a regular Mapquest map, it doesn't even attempt to print what is actually on the screen, but instead prints "Sorry! When printing directly from your browser your map may be incorrectly cropped. To print the entire map, try clicking the 'Printer-Friendly' link at the top of your results page." This is aggravating in many ways:

1) They wasted a piece of paper and ink, instead of popping up an error message with the same text.

2) They won't even let me try to print what was actually there. The cropping, if any, may have been acceptable to me.

3) It seems they are being dishonest, in that the real reason they want me to pick the "Printer-Friendly" link is to display (and print) more ads.

4) They simply crop the map severely before the print, when you hit that button, there isn't any magic that allows the uncropped map to be printed.

5) Them "taking control" of the print function like that seems like malware, to me. I never gave them permission to change the way my computer works to refuse to print certain pages.

I can work around this annoyance by using Print Screen, pasting into Microsoft Paint, removing the ads, and printing from there. But, I'd like to know if there is a way to stop Mapquest from taking control of the print function. Any ideas ? StuRat 17:37, 23 June 2007 (UTC)

Might not be mal-ware; CSS can be used to print something entirely different from what appears on-screen. Other than that - I always use the PrintScreen method myself Rawling4851 22:55, 23 June 2007 (UTC)
Yup, looking at the HTML source, Mapquest uses different CSS for when the page is printed. IE7 picks this up in the print preview. Looked to me as if the whole map was on there, but again, that's IE7's improved (compared to IE6, at least) printing abilities. MapQuest isn't evil, it's the CSS standard being correctly interpreted by your browser that lets something be printed that is different from what you see on screen. At least you get something similar; your computer could be instructed to print something entirely different. Rawling4851 23:28, 23 June 2007 (UTC)
Thanks, is there any way to ignore the CSS and print what's on the screen without going through the whole print screen method ? It seems quite reasonable to me to expect to be able to print what I see, as opposed to whatever the authors of the web page feel like sending to my printer. StuRat 02:24, 25 June 2007 (UTC)
Google Maps is the best.. tru dat.. DOUBLE TRUE..... --frotht 13:42, 25 June 2007 (UTC)

[edit] Screeching computer with flashing power light

My computer makes a screeching noise, the power light flashes, and the computer won't turn off.

zerula

That sounds really bad. I'd pull the plug out to stop it immediately, to prevent any further damage, then worry about fixing it later. StuRat 18:08, 23 June 2007 (UTC)
Hold down the power button for 4 seconds to turn off. When you switch on your PC, your system speaker will bleep once. Anything else - such as a constant note - indicates hardware failure. Consult your motherboard's manual to see what error is reported. --h2g2bob (talk) 20:24, 23 June 2007 (UTC)
Sounds like a PSU is going out to me. -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 20:25, 23 June 2007 (UTC)
I was asked to look at a computer that was "screeching" once. It was the "overheat" warning going "beeeeep beeeeep beeeeep". So, it is important to define what "screech" means. Youth in Asia 13:10, 24 June 2007 (UTC)
Agreed- I'd say from the description (at least the first two parts) that it sounds like a hard drive problem, but it could be any number of things if screech is undefined. A screech could be anywhere from a beep to nails-on-chalkboard -- Phoeba WrightOBJECTION! 14:12, 24 June 2007 (UTC)

[edit] How to Merge Partitions

Can someone tell me how to merge to HD partitions from windows XP? I have attempted this via a bootable disc, but I am not having an easy time as nothing seems to boot properly, even though it is set to boot from CD/DVD first in the bios.82.153.69.54 18:01, 23 June 2007 (UTC) PS: This is a laptop.82.153.69.54 18:13, 23 June 2007 (UTC)

Absolutely impossible without complete data loss on both partitions. And there are no native tools that come with xp to do this, you'll have to upgrade to vista or use the gparted livecd or something --frotht 13:43, 25 June 2007 (UTC)

Pooh. I tried gparted, but it wouldn't boot (like everything else!).82.152.168.118 16:07, 25 June 2007 (UTC)

hi , you can use Partition magic for this its a very good and simple to use.

[edit] gop.com

GOP.com doesn't respond when I click on it, does it not work in general or is it because I'm not in America (but in Austria). Thanks, Jeffrey.Kleykamp 19:51, 23 June 2007 (UTC)

Works for me (from the UK) Tomgreeny 21:09, 23 June 2007 (UTC)
Is it, maybe, because I'm in a German speaking country? Jeffrey.Kleykamp 21:12, 23 June 2007 (UTC)
Works from Germany. --Dapeteばか 21:13, 23 June 2007 (UTC)
Why doesn't it work here? Jeffrey.Kleykamp 21:44, 23 June 2007 (UTC)
Times out here (UK). Maybe the server's experiencing downtime? JoshHolloway 22:34, 23 June 2007 (UTC)
My theory is that the GOP people only want to allow Americans on by looking at the IP addresses but if it works 2/4 times in other countries then... Jeffrey.Kleykamp 22:37, 23 June 2007 (UTC)
(From Australia). GOP.com doesn't work for me, nor does RNC.org, but GOP.gov does, if that's any clue as to what's going on. - Akamad 01:15, 24 June 2007 (UTC)
This article from November, 2004 states that they are blocking international access. - Akamad 01:18, 24 June 2007 (UTC)
GOP.gov probably works in that it is the House Republican Leadership website. An nslookup points to 143.231.169.175, and 143.231.0.0/16 is assigned to the House of Representatives, so it stands to reason the GOP couldn't restrict access on that basis. GOP.com and RNC.org appear to be owned directly by the party, and their IP's belong to what it appears to be a provider based in Chattanooga, Tennessee.
I'm also a little dubious of the justification quoted in the article for disallowing access from abroad. It's not as if the security concerns aren't the same that any other high visibility website wouldn't have to face. Maybe they don't think it's worth the additional bandwidth and utilization costs to provide that kind of access, but it seems a little short-sighted to me. –Pakman044 03:38, 24 June 2007 (UTC)
Works for me (UK). Use a coralized version by adding .nyud.net:8080 after the .com, like this: http://www.gop.com.nyud.net:8080/ Coralizing should pick it up (hopefully in the US) and feed it to you through several proxies. It takes a lot longer than normal to load, and may get strange formatting marks, but probably won't timeout. --h2g2bob (talk) 16:18, 24 June 2007 (UTC)

[edit] LaTeX to Powerpoint

I have a large number of equations in Latex that I want to put into a powerpoint presentation. What is the neatest way to do this? deeptrivia (talk) 22:45, 23 June 2007 (UTC)

You can use TexPoint, it's free for 30 days. There's also TeX4PPT (which I believe is free). I've never used either myself but it seems to do the trick. I guess you can also use print screen, but that might not look nice if you have coloured backgrounds on the slides. Hope that helps. - Akamad 01:04, 24 June 2007 (UTC)
All these create non scalable images. Is there a way to get Latex equation to Microsoft Equation Editor format, or atleast scalable image format. deeptrivia (talk) 02:19, 24 June 2007 (UTC)
It's quite easy to get PostScript / PDF output from TeX, either with pdfTeX or by producing a standard DVI file and running dvips / dvipdf on it. However, to make the output truly scalable, the important thing is to install and use vector versions of the TeX fonts instead of the pre-rasterized fonts generated by METAFONT. —Ilmari Karonen (talk) 20:35, 24 June 2007 (UTC)
May I suggest Beamer instead of powerpoint? I used it for my Master's thesis, and I was very happy with it. You also might want to reconsider presenting a large number of equations in a presentation, as the audience will have a hard time quickly processing them. Which of the following do you think your audience would prefer? "Since |\{x\in\{0,1\}^*|x\sqsubseteq y\}| < \infty we have..." or "Since y has finitely many prefixes, we have...". 213.237.10.244 13:18, 27 June 2007 (UTC)