Wikipedia:Reference desk/Archives/Computing/2008 September 16

= September 16 =

Processes on start up
How do i stop annoying programs from automatically loading when I start up my computer? for example 'itunes helper' or 'adobe speed launcher' hiding in my processes list. I just don’t want to open at all is there a way to stop them loading? (Other then uninstalling the programs altogether). I am running Vista. —Preceding unsigned comment added by 203.202.144.223 (talk) 02:44, 16 September 2008 (UTC)


 * See http://www.howtogeek.com/howto/windows-vista/stop-an-application-from-running-at-startup-in-windows-vista/. --194.197.235.221 (talk) 05:18, 16 September 2008 (UTC)


 * As a simpler answer: Start -> Run -> msconfig, choose Startup tab when it opens. Basically, if it says 'Updater' in the path, you can generally disable it. jusched is a big one too (the Java auto-updater). Generally you can't mess up too much, though try to stay away from anything that you know is associated with your anti-virus (Symantec, AVG, Norton, long list there). Also, iTunes helper and Adobe speed launcher? Totally unneeded. Washii (talk) 05:51, 16 September 2008 (UTC)
 * Fixed to say "can't mess too much up."

Sweet thanks a lot. —Preceding unsigned comment added by 203.202.144.223 (talk) 05:54, 16 September 2008 (UTC)


 * There is actually a better way: msconfig isn't able to stop everything. A MUCH better program is AutoRuns from Sysinternals. It'll show you everything that goes on a start-up and gives you the option to stop it. 90.235.12.16 (talk) 11:13, 16 September 2008 (UTC)
 * I second that, "autoruns" is good, make sure you understand what you are doing however, or you may soon find you have disabled processes you should not such as useful drivers. Equendil Talk 17:34, 16 September 2008 (UTC)

Can't install an important Vista update
After installing Vista on a "Vista Ready" HP Pavilion desktop, I downloaded the available updates but got a 643 error on one of them. It turned out to be for Microsoft .NET Framework 1.1 Service Pack 1. But the Microsoft download website says this service pack was published in August 2004. If I download this version, wouldn't it clobber any later version of .NET Framework I might have? This is supposed to be an important security update. What do I do? --Halcatalyst (talk) 02:48, 16 September 2008 (UTC)
 * Don't worry about it. It's probably installed in some form already. The Windows Update site is relatively dumb. However, if you'd like to check it anyway, grab the .NET 3.5 Network Installer (or Redistributable) and install that. It should install .NET 1.1 SP1, .NET 2.0 SP1 and .NET 3.0/3.5. You can also grab the .NET 3.5 SP1 installer and run that. Then you can make triply sure they're all up-to-date. Washii (talk) 05:56, 16 September 2008 (UTC)
 * .NET 3.5 Redistributable, does not install .NET 1.1. .NET 1.1 is also not included in Windows Vista, it is a separate system from .NET 2.0 onwards and will not interfere with it if installing afterwards. - Jimmi Hugh (talk) 18:20, 16 September 2008 (UTC)


 * Can I request a tie breaker? ;) --Halcatalyst (talk) 01:26, 18 September 2008 (UTC)
 * Jimmi Hugh was right. I thought I saw 1.1 included in the list for 3.5, but I stand corrected. Also, installing these after having installed 2.0 or above won't affect anything but 1.1
 * Here's the link for .NET Framework 1.1 Redistributable: http://www.microsoft.com/downloads/details.aspx?familyid=262D25E3-F589-4842-8157-034D1E7CF3A3&displaylang=en
 * Here's the link for .NET Framework 1.1 SP1: http://www.microsoft.com/downloads/details.aspx?FamilyId=A8F5654F-088E-40B2-BBDB-A83353618B38&displaylang=en Washii (talk) 05:14, 18 September 2008 (UTC)

where do i buy kudzu?

 * (Moved to science desk -- OP meant plants, not software) --- OtherDave (talk) 10:58, 16 September 2008 (UTC) 

internet history
Hi There,

Is there any programme to clean internet history for good.

P.P.80.254.93.170 (talk) 08:46, 16 September 2008 (UTC)


 * You can clear your internet history for Internet Explorer by clicking tools > internet options, then clicking 'clear history'. You can also clear temporary-history-files within the same area (Tools > internet options > Delete files > tick 'delete all offline content' and click ok), in the same area you can also delete Internet cookies (tools > internet options > delete cookies... > 'ok'). The information should be then pretty much 'deleted' from the view of most everyday people. Clever people could still perhaps access information - and certainly recovery-specialists could potentially get information from your hard-drive, but realistically unless you're doing something illegal/highly dubious the above should be more than enough to remove your history from everyday viewing (i'll assume you are buying a present for someone and don't want them to look at page-history and find a link to the gift selling website...) 194.221.133.226 (talk) 09:30, 16 September 2008 (UTC)


 * Every browser I know of lets you clean out the cache. For something really drastic: (1) save the bookmark file, (2) uninstall the browser (a bit difficult if it's MSIE), (3) install a newer version (there always is one available) of the browser, (4) copy back the bookmark file, (5) tell your parents (if they notice) that you upgraded as a "security measure." That's because "security" always sounds good. (But don't over-egg it with talk about "Islamic extremists" or other piffle too obviously borrowed from US electioneering.) Tama1988 (talk) 09:37, 16 September 2008 (UTC)


 * I use,, just because it's so easy to use... but there's probably some free tools available too for those who know what they're doing with them. It also allows you to see deleted files, which i suppose some people might find useful, i've never used it. - Jimmi Hugh (talk) 09:43, 16 September 2008 (UTC)

Thanks for your help, the main reason is that I am selling my pc, and I would like to clean it from all internet history,web sites, coockies etc. I don't want anybody sneaking on me, such as, on which web sites I hit.

p.p.80.254.93.170 (talk) 10:36, 16 September 2008 (UTC)


 * If you're selling your PC, you should wipe the entire hard-drive clean (not just the browser history) using something like Darik's Boot and Nuke. That will remove any trace of what was on there before (and it's free!). Then, if you want, you can just reinstall the OS clean or just sell it without an OS on it. 90.235.12.16 (talk) 11:06, 16 September 2008 (UTC)


 * I should make something a little bit clearer: Darik's Boot and Nuke will wipe your ENTIRE hard-drive away. I mean, nothing on there will ever be recoverable, and you will not be able to boot it up because the OS has been wiped. So fair warning: if there's anything on there you want to save, be sure to save it before nuking it. And if you want your computer to start again, you'll need a Windows XP Install CD so you can reinstall the operating system (which will also be gone). 90.235.12.16 (talk) 11:10, 16 September 2008 (UTC)


 * That's a little excessive... unless you think there is a serious shortage of computing power at a Major goverment or scientific institute, the chances of anyone with the ability to recover past a single layer of format and caring about the data they find is unlikely. Windows 98 Boot floppy, "format C:". Sorted, sell it and mention it doesn't have an OS unless you reinstall it. Also, some computers (I love Dell, so they're probably the only ones) come with a reinstall system mentioned in the manual that sets the computer back to the state it was in when you first got it. If you run this, it won't format your drive, so you should run a program like the one I mentioned above or cyberscrub to clean the unused clusters before selling it. - Jimmi Hugh (talk) 12:18, 16 September 2008 (UTC)


 * I'm under the (possibly mistaken) impression that DOS format doesn't overwrite the contents of the drive even once. If you do a quick format (/Q) it never even looks at most of the disk surface. If you do a full format it tests the whole disk for bad sectors, but I think this is done by reading, not writing. It's easy to recover files from a formatted disk using software like PhotoRec. Probably the most convenient way to securely wipe your C: drive is to install TrueCrypt 5.0 or later and enable boot-drive encryption, choosing a nice long passphrase. You can keep using your computer while it scrambles the disk contents, and in fact up to the moment you give it away. Then you simply power it down and wipe the passphrase from your mind. -- BenRG (talk) 13:16, 16 September 2008 (UTC)


 * I can't speak for all versions, but the Format included in the standard "Windows 98 Bootdisk" created by XP/Vista, fills the entire partition with 0's, meaning pyhsical drive analysis would be needed to recover any data. - Jimmi Hugh (talk) 15:17, 16 September 2008 (UTC)
 * However I am under the impression it doesn't absolutely get everywhere. My suggestion is stick with what works, which is Dariks Boot and Nuke. --mboverload @  05:15, 20 September 2008 (UTC)

BIOS with embedded OS
Why don't BIOS have an OS embedded? Flash memory is cheap enough and OS like Puppy Linux only have 128 MB....Mr.K. (talk) 10:36, 16 September 2008 (UTC)


 * Some do - SplashTop, HyperSpace (software), and I think Dell are making a few machines with this (whether its their own tech or one of these I don't know). -- Finlay McWalter | Talk 10:42, 16 September 2008 (UTC)


 * Dell's thing is called "Dell Latitude ON" - http://www.engadget.com/2008/08/15/dells-latitude-on-instant-os-detailed-screenshooted/ I don't think they're actually selling it yet. -- Finlay McWalter | Talk 10:45, 16 September 2008 (UTC)
 * The disadvantages out weigh the advantages. While there are some executive system programs being embedded in some BIOS to cover those advantages (resident recovery system etc.), given the longevity of some system's, the constant upgrading of software features and the size of hard drives, there are few good reason's to stick it all in ROM. While you can of course Flash ROM, it still limits the Operating System. Also note, that any speed advantages ever seen from ROM have long gone, and nearly all ROM is now shadowed into RAM before executing. - Jimmi Hugh (talk) 12:13, 16 September 2008 (UTC)


 * There are versions of Linux that can be burned into a BIOS ROM too. Coreboot (which used to be called "LinuxBIOS") does exactly that. SteveBaker (talk) 13:59, 16 September 2008 (UTC)
 * Plus every embedded device you can find... I think his point was about featured desktop systems as opposed to utilities and specialist devices. - Jimmi Hugh (talk) 15:14, 16 September 2008 (UTC)

Camera / memory card
Will this type of memory card work with my camera - a Polaroid I633. If not... what type do I need? Thank you in advance! --Endlessdan and his problem 14:26, 16 September 2008 (UTC)
 * No. According to that Walmart page (assuming it's correct), the camera takes SD and MMC cards only. That is a MicroSD (actually MicroSDHC) card, which is a smaller version of SD. You can buy adapters, but there's no mention of the camera supporting SDHC; in that case, even an adapter would not allow the card to work. If the camera doesn't support SDHC then what you're stuck with is plain SD/MMC cards. Keep in mind that anything over 2GB is an SDHC card. 24.76.161.28 (talk) 14:46, 16 September 2008 (UTC)

I can't get to the ebay link but based on the title it suggests that the type is 'micro SD'. Based on the information on the walmart link the camera takes an SD Card. These are cheap and available pretty much everywhere. There's no need for you to get micro-SD - although the SD Card article suggests it is possible for micro-sd cards to be used in SD card slots...but there's no point. You can find SD cards everywhere. The listing doesn't say whether or not High-capacity SD cards (SDHC) work so I would get a normal SD Card (say 1gb or 2gb) - that should provide plenty of storage at a reasoanble cost. 194.221.133.226 (talk) 14:48, 16 September 2008 (UTC)


 * Two concerns, one minor


 * 1. Minor issue: Some[who?] have suggested that using an SD card might give you faster performance. I am not sure about it.


 * 2. Major issue: Most[citation needed] older consumer electronics do not support SD cards with a capacity greater than 2 GB. It might be that your camera falls under that category. As the eBay seller cautions, "Please make sure your devices can support microSDHC, we do not responsible for any compatibility problem." I am sure someone more knowledgeable about digital photography and flash storage will come by to answer your question shortly. However, in the meantime, y You might want to read Secure Digital card. (If your camera does not support SDHC, your only solution would be to get a card of 2 GB or smaller capacity.) Have a great day!  Kushal (talk) 14:54, 16 September 2008 (UTC)

Thank you everyone for your help. I am a complete clod when it comes to computer junk. So, would this kind of card be okay? I'd prefer a 2GB, but every time I do a search... it comes up with a micro card. Can someone kindly just point me in the direction of a cheap, non-micro 2GB card? --Endlessdan and his problem 15:00, 16 September 2008 (UTC)
 * I think I found one. I just used -micro in my search. Would this card be suitable for my camera? Again, many thanks. --Endlessdan and his problem 15:01, 16 September 2008 (UTC)
 * Too many edits. Yeah, that would work. 24.76.161.28 (talk) 15:04, 16 September 2008 (UTC)
 * (edit conflict) $7.25 to your door, assuming you're in the USA. -- Coneslayer (talk) 15:02, 16 September 2008 (UTC)
 * Unknown asian brand? 24.76.161.28 (talk) 15:04, 16 September 2008 (UTC)
 * A 2GB MicroSD card with an SD adapter, kind of like the first listing, will work fine. The only issue is that it doesn't appear your camera supports SDHC (though that's not always well documented, you may want to check the manual), so you have to go for 2GB or lower. But yes, that 1GB card will work too. 24.76.161.28 (talk) 15:03, 16 September 2008 (UTC)

Cool. I sorta assumed the cards I linked to on Ebay, the "scandisk" memory cards were bootleg. Should I avoid the Asian card for $7.25? --Endlessdan and his problem 15:08, 16 September 2008 (UTC)


 * PQI is a pretty well-known memory manufacturer. You can see many more reviews at Newegg.  I've used PQI cards (other sizes) with good success.  I've also used the vendor I linked to once or twice with no problems.  I think you'll have trouble finding non-Asian memory products.  -- Coneslayer (talk) 15:26, 16 September 2008 (UTC)


 * Cool. I will get that card then. Thanks! --Endlessdan and his problem 15:35, 16 September 2008 (UTC)


 * It is safe to assume that every single SD card on ebay is bootleg. --Carnildo (talk) 20:25, 16 September 2008 (UTC)

html help for tables
border=" 3px solid #DFFF00" creates a border around all the cracks and crevasses of my table. What can I do make it go only round the very outside. Thanks, html noob --217.227.85.122 (talk) 14:57, 16 September 2008 (UTC)

Could this be what you're looking for? You seem to be mixing HTML4 attributes and CSS. --Kjoonlee 15:31, 16 September 2008 (UTC)

--This what you want? --98.217.8.46 (talk) 15:44, 16 September 2008 (UTC)
 * Yep that's it. Why isn't it rendering the same on my screen? --217.227.89.144 (talk) 16:02, 16 September 2008 (UTC)


 * Look at the style guidelines I put on it. You have to set border-collapse: collapse; to make it stop automatically adding border lines to the TD elements as well. --98.217.8.46 (talk) 17:25, 16 September 2008 (UTC)

JPG Fullscreen, quick!
I need a quick way to display a jpg image fullscreen of my computer. Sorta like a boss key. Any suggestions? 15:04, 16 September 2008 (UTC)


 * I assume you are using Windows. Make your desktop image whatever JPG you want to show.  Remove the desktop icons (they shouldn't be there anyway).  There is a "show me my desktop" icon, button, and keyboard command for Windows.  I don't use Windows, so I don't know what the keyboard command is.  Press it and all open windows will be replaced by your desktop - which is the jpg image you want to show. --  k a i n a w &trade; 17:36, 16 September 2008 (UTC)


 * If the keyboard has the "Windows" key, then the "Show Desktop" keystroke is Win-D. The Win key has to be held down while pressing the D. -- LarryMac  | Talk  17:43, 16 September 2008 (UTC)


 * Kainaw - what do you have against icons on the desktop (in an operating system you don't even use)? AndrewWTaylor (talk) 13:02, 17 September 2008 (UTC)


 * Many people feel that the windows desktop analogy has been corrupted by software manufacturers who believe that it is the place for short-cuts to their applications, where it should only be for the documents and files you are currently working on, just like a real desktop in an office. - Phydaux (talk) 14:49, 17 September 2008 (UTC)


 * The average user does not use the desktop as a desktop. He or she uses it as a repository for absolutely everything.  Every application has a link on the desktop.  Every picture they downloaded is on the desktop.  Every attachment from their email is on the desktop.  Every install program they downloaded is on the desktop.  When I'm asked to fix a problem on their computer, I have to stand there as they spend 10-20 minutes looking around on their desktop for the icon they clicked that caused the problem.  By having to support users who put hundreds of icons on the desktop - to the point that they are all overlapping, I am at the opposite extreme of not wanting to see a single icon on the desktop at all. --  k a i n a w &trade; 01:32, 18 September 2008 (UTC)
 * And then they solve that problem by using Google Desktop. They then get in the habit of sticking stuff anywhere they want.  Then Google Desktop breaks. And they're screwed. I've seen it.  --mboverload @  05:12, 20 September 2008 (UTC)

Syncing
I am looking for a software solution which would allow me to efficiently sync the contents of two PCs (both running WinXP). The first PC has all of the data (in several thousand directories and files), a subset of which is updated daily (new files are added, some old files are deleted, some files moved around, some renamed, and some updated). The second PC is supposed to mirror the first one, which is done daily by copying the affected directories from the first PC to a USB key and then replacing the matching subset on the second PC with the USB key data. I am looking for a solution that would allow to automate the process&mdash;i.e., at the end of the day it should determine which files/directories on the first PC have changed, copy only the changed data to the USB key, and then update the second PC with that information. The size of the daily updates can be anywhere from a few megabytes to almost two gigabytes and over time is expected to grow on average. The PCs are not networked (and can't be, as they are located in separate physical locations and one is not connected to the Internet for security reasons).

Preferrably, the software should be free (or reasonably priced), run without installation (being able to run it from the USB key itself would be ideal), and be able to handle missed cycles (i.e., if the updated data from the first PC were copied to the key but never moved to the second PC, it should not be a problem on the next day when more updated data is added to the key), but I realize that meeting all of the criteria is probably asking too much. Nevertheless, I would most appreciate any suggestions, as no software I've tried so far was able to supply this supposedly trivial functionality with satisfactory results. Thanks!—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 16:08, 16 September 2008 (UTC)


 * Would Robocopy work? I believe it lets you set job times. Alternatively, you may be able to script the SyncToy Windows XP PowerToy. You'll need to install one or the other, which may somewhat preclude the "run off USB." I'm not sure how well robocopy especially would hold up for that. Washii (talk) 16:36, 16 September 2008 (UTC)
 * Or, you could schedule a batch script with xcopy: xcopy src dest /D /E /V /C /I /H /Y /Z
 * This will copy all newer files, copy all subdirectories, verify copies, continue on error, default to copying files to a directory, copy hidden+system files, suppress overwrite prompts, copy in restartable (could probably remove that).
 * However, it may be better to use robocopy. Xcopy has some evil niggling details (if a path is >253 characters, it fails with an 'insufficient memory' error). I believe the third-party XXcopy doesn't have this limitation and supports mostly the same commands. Washii (talk) 16:36, 16 September 2008 (UTC)


 * I haven't used it, but the PortableApps version of Toucan may satisfy your requirements. At the very least, it will run directly from the USB drive&mdash;that's the point of PortableApps.  -- Coneslayer (talk) 16:48, 16 September 2008 (UTC)
 * Xcopy was my firsh thought, too, but it's unsuitable because it ignores the fact that some of the files are being deleted and some are moved around. Robocopy and Toucan, however, look promising, at least at the first glance; I'm going to give them a test run.  Meanwhile, if anyone has other suggestions, I'm ready to try them out as well.  Thanks much!—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 17:08, 16 September 2008 (UTC)
 * If you use robocopy /MIR, be very careful that you know what it does. My reasoning glitched the first and second times I used it. Give it a couple of shots on a fake folder-set so you know what it will do in varying situations. However, that should handle the 'remove files' cases. Washii (talk) 17:20, 16 September 2008 (UTC)
 * Thanks for the word of warning :) I'm going to test robocopy next, as Toucan did not work out for me&mdash;it's much like many other tools I tried, as it does a fairly decent job of syncing a source and a destination, but not through an intermediary means (USB key, in this case).—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 17:33, 16 September 2008 (UTC)
 * I don't know how far you've gotten, but I did find out that just the robocopy executable should work all on its lonesome as long as you're using a Unicode-based Windows (at least Win2k and up). It should be fairly easy to write batch scripts for re-using the same parameters, but different paths, too. —Preceding unsigned comment added by Washii (talk • contribs) 17:53, 17 September 2008 (UTC)
 * I've so far been testing robocopy (and yes, it works on its own just fine) and found it to be an extremely useful and flexible tool (so thanks for the tip) which, unfortunately, does not work for my purposes. If I were able to copy the whole data set to the USB key, robocopy would do a marvelous job of syncing the first PC with the USB key and then the USB key with the second PC.  However, since the whole thing is currently at 250 gigs and is only expected to grow, using a USB key for that is out of the question (and hauling a USB hard drive back and forth every single day is something I am desperately trying to avoid).  The software I need should be able to work with subsets of the data (which probably means it needs to create a catalog of the first and second PC contents, be able to create USB key subsets/copy them to the second PC using those, and then update those catalogs accordingly at each run).  Any other suggestions would be greatly appreciated.  I have not tried the SyncToy yet, but by the looks of it it is going to have the same problem as robocopy (i.e., it would sync a source and a destination, but not via subsets recorded onto an intermediary device).—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 14:23, 18 September 2008 (UTC)

Sorting a text file by line length, linux
I need to sort a huge text file, ordering it by increasing line length. The shortest lines in the input are to be the first in the output, then the second-shortest lines, and so on. I would imagine this to be a one-liner, but am not all that familiar with bash scripting, and would appreciate help on how to achieve this. Thanks. --NorwegianBluetalk 18:15, 16 September 2008 (UTC)


 * I don't know of a simple way to do that, as sort doesn't have that feature. The following ugly little script does it by prefixing each line with its line length, calling sort, and then removing that prefix.

 python -c "import sys for line in sys.stdin: sys.stdout.write('%08d%s' % (len(line),line)) " | sort | python -c "import sys for line in sys.stdin: sys.stdout.write(line[8:])"
 * 1) !/bin/bash


 * If you had that saved in lsort.sh, your command line would be     87.114.18.90 (talk) 20:35, 16 September 2008 (UTC)
 * Thanks, I'll try it out! --NorwegianBluetalk 20:48, 16 September 2008 (UTC)


 * If you have Perl installed, you can do it with this one-liner:

 perl -e 'print sort { length $a <=> length $b } <>'
 * The only One caveat is that this won't work correctly if the last line isn't terminated with a newline. The best I can come up with for that case is

 perl -e 'print sort { length $a <=> length $b } map { chomp; "$_\n" } <>'
 * (edit: As mentioned below, it also won't work with files too large to fit in memory.) -- BenRG (talk) 21:01, 16 September 2008 (UTC)


 * Beautiful! The last line was newline-terminated, so no problems with that. Worked like a charm, and fast, too. Thank you! --NorwegianBluetalk 21:33, 16 September 2008 (UTC)


 * Ben, does Perl's sort work properly on a dataset that's bigger than the process's available mem/swap? I copped out of the sort and used the coreutils sort because I was under the impression that it did, but on reviewing its man page it's silent on the topic. If NorwegianBlue's input file really is that huge, then this may be an issue.  87.114.18.90 (talk) 21:41, 16 September 2008 (UTC)
 * Ben's solution does indeed have that problem if the file is very large. The "<>" construct slurps all lines into memory. --Sean 23:23, 16 September 2008 (UTC)
 * I'll keep that in mind. I tried it out on a smaller (though still large) version of the file in question. If this turns out to be a problem, I'll use 87.114's script. Thanks again! --NorwegianBluetalk 09:50, 17 September 2008 (UTC)


 * A side effect of 87.114.18.90's bash method is that same-length lines are sorted. In BenRG's perl method same-length lines appear in an arbitrary order.  If that matters.  Saintrain (talk) 13:16, 17 September 2008 (UTC)
 * According to this documentation, you could add  to the beginning of BenRG's code, and it would guarantee stable sorting.  That is, lines of the same length would come out in the same order that they went in.  That's different from 87.114.18.90's output, where everything is sorted lexically.  Either one might be preferred, or it might not matter.  -- Coneslayer (talk) 13:27, 17 September 2008 (UTC)
 * In my code you can get lexicographical sorting within each line length by replacing  with  . -- BenRG (talk) 16:13, 17 September 2008 (UTC)
 * That's very cool. I've never perl'ed, it looks very handy.  Thanks.  Saintrain (talk) 18:23, 17 September 2008 (UTC)
 * In the event that huge really does mean "bigger than available RAM/swap" (and my suspicion that GNU/POSIX sort will only sort in memory) the following will do a bit better

 !/usr/bin/python

theGreatIndex=[] location=0

f = open('srcfile.txt','r') for line in f:   length=len(line) theGreatIndex.append((length,location)) location += length

theGreatIndex.sort

for (len,pos) in theGreatIndex: f.seek(pos) line = f.readline print line,


 * As it only remembers the line length (and its position in the file) it uses memory more efficiently (by how much depends on the distribution of lines in the source data). If NorwegianBlue's file has a pretty long average line length (e.g. apache log files) then this should allow you to sort a file perhaps 10 times bigger. But if the file is significantly bigger than available memory that (essentially random) seek will annihilate the disk cache, and this will all run eyewateringly slowly. The best thing to do when sorting astronomically large files (by line length or anything else) is probably a multi-stage windowed mergesort between a bunch of temporary files. 87.114.18.90 (talk) 16:51, 17 September 2008 (UTC)


 * Actually, I think GNU sort will operate on files larger than memory/swap. (I'm also surprised.)  The manual doesn't say explicitly, but it uses TMPDIR for something, and it lets you specify the size of a "memory buffer".  It also cautions that the "buffer grows beyond size if sort encounters input lines larger than size", implying that the longest line has to fit in memory.  -- Coneslayer (talk) 16:59, 17 September 2008 (UTC)

Gaming laptop
What is the cheapest laptop or netbook that could run a game like Team Fortress 2 well enough to play it very smoothly at its lowest-end graphics settings?--142.108.107.93 (talk) 18:40, 16 September 2008 (UTC)
 * You can get cheap laptops at every major brand for about $400-$600 that can easily run such a game, provided no other program is running. The cost goes up by $200 if you plan on using genuine Windows. You should go for 1 GB of RAM, getting a good CPU is very expensive. Admiral Norton (talk) 18:52, 16 September 2008 (UTC)


 * My friend's Toshiba laptop with a Core 1 Dual 1.8Ghz and Geforce 7300Go with 1GB of ram can play TF2 on lowest setting on its native resolution, 1280&times;800. TF2 isn't that demanding. --antilivedT 04:36, 17 September 2008 (UTC)


 * It's also important to have separate graphics memory (ie not "shared graphics memory"). This will markedly improve gaming performance. &mdash; QuantumEleven 12:54, 18 September 2008 (UTC)

Mobile MSN doesn't show contacts as online
Hi all,

I have a smartphone with WM6 and Windows Live Messenger. Why does Messenger show all contacts as offline even if they're for sure online?

Thanks, HardDisk (talk) 19:43, 16 September 2008 (UTC)

PS: How do I determine which program transmits how much data via GPRS/EDGE? I hate it that even if I'm connected via Wifi, that some app wants Internet access via GPRS :(
 * It's a known bug (see also, ). As far as I know, signing off and back on is the only way to "resolve" this problem.—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 20:31, 16 September 2008 (UTC)
 * There was a hotfix mentioned on one of the links, it solved the problem. Thanks!HardDisk (talk) 18:26, 17 September 2008 (UTC)

Sign-in secure for this online banking?
Hi all,

I am a member of Cambridge Savings Bank, and the website for signing in is https://www.csbwebonline.com/onlineserv/HB/Signon.cgi

Although the protocol is https, the little info button to the left of my Firefox browser says that the site 1) does not supply any identity information about itself, and 2) my information is not encrypted. Compare this to my Bank of America info button, which not only says the information is encrypted, but also that it's been verified by Verisign.

Is there any reason to not feel safe on the Cambridge Savings website?

Thanks, &mdash; Sam 21:08, 16 September 2008 (UTC)
 * The certificate itself is issued by VeriSign and is still not past it's expiration date. On FF3, I get a message saying that parts of the webpage are not encrypted before data is transmitted over the internet. The question that you should ask the bank is what is or isn't encrypted.  Thanks AreJay (talk) 21:34, 16 September 2008 (UTC)


 * The message about parts of the page being insecure usually means that it's including bits of content that are accessed through plain HTTP instead of HTTPS. In this case the little house icon at the bottom of the page (before "Equal Housing Lender") is an example of this; the page also includes some CSS stylesheets and JavaScript files called by HTTP. Poor page design, as it is confusing, or even scary, to the non-technical user, but probably not actually a security hole. I'd suggest you should point out the problem to the bank (though I have done the same with a bank I have accounts with, to little effect other than a patronising "everything is fine" reply). AndrewWTaylor (talk) 12:52, 17 September 2008 (UTC)


 * OP, I think you should still talk with your bank about it. Why should we put up with the laziness of their Web Development team? Kushal (talk) 20:04, 17 September 2008 (UTC)

Thank you all very much! &mdash; Sam 00:52, 18 September 2008 (UTC) —Preceding unsigned comment added by 146.115.120.108 (talk)

Dynamic Controls - Ideas please :)
Hi,

I'm writing a small program using vb express 2005. I currently have a form with a few controls (listview, comboboxes, a couple of masked editboxes and buttons here and there).. i need to make similar forms but with different controls, labels etc... So i thought it might be more efficient (and more flexible down the road) if the form / controls would be dynamically created from something like a list or database. But i'm afraid that might be making it a little too complicated (if need to store info like control type, location, etc) and take too much coding time. I was also considering just adding items (much like headers) to a list,and force all input to be done thru something like an edit box, but it would look too 'unprofessional'...

Anyone got ideas on how to implement this better? Am confident there are... Examples or links to articles pointing me in the right direction would also be much appreciated.

Thanks in advance 119.111.70.146 (talk) 21:17, 16 September 2008 (UTC)


 * Sorry, I don't VB but in C++ (I know you want VB but a C++ response might be better than none?) it would go something like this. In the constructor of the form where the button is to appear:

Button *pBtn = new Button(this); // make a new Button. pBtn is a pointer to the new instance pBtn->Parent = this;           // attach it to the form pBtn->Left  = 0; pBtn->Top   = 0; pBtn->Width = 25; pBtn->Height = 15; pBtn->Caption = "DynoButton"; pBtn->onClick = myDynamicButtonClickHandler; etc.
 * There's probably a list of properties like this in the button's "editor"? Don't forget to delete the button in the form's destructor. WYSIWYG is easier, huh?  Hope this helps.  Saintrain (talk) 20:29, 17 September 2008 (UTC)


 * You can do it programmatically in VB.NET but it isn't a trivial task. It's not really a trivial task in any language, frankly—even making forms in HTML (which takes care of most of the positioning issues automatically) requires a lot of tinkering, and for something like VB it's usually not worth doing unless there is a very strong need to have form information programatically assembled. Consider that in the end it would probably take as much if not more work than just doing it all in Design mode—even with a very robust system you're not going to be able to get off by just saying, "Oh, make me a few listboxes", you're still going to have to specify where they go, what data they have, what happens when you change them, etc. etc. etc.. --98.217.8.46 (talk) 00:39, 18 September 2008 (UTC)
 * It's incredibly simplistic. Create a new object of the type of control you want, set it's properties as you would any other class of control, even simpler if you load this data from a file, then add it to the specific Forms control collection. There's even a few simple examples in the MSDN library. - Jimmi Hugh (talk) 00:48, 18 September 2008 (UTC)


 * thank you all very much for your responses. i feel *.8.46 understood my question best, and was able to convince me to just hard-code it. 119.111.70.146 (talk) 17:59, 18 September 2008 (UTC)