Computing desk
< July 20	<< Jun \| July \| Aug >>	July 22 >

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.

July 21

Unfreeze KDE

When KDE freezes, my Kubuntu system often takes too long to respond to Ctrl-Alt-Esc, Ctrl-Alt-Del and Ctrl-Alt-F1, and I have to hard-reboot the computer. Is there any way to forcibly interrupt KDE and its apps, and get a console shell without the delay of Ctrl-Alt-F1? Once that's done, are there tools other than htop that can help me find and kill the process responsible for the freeze? Neon Merlin 00:33, 21 July 2009 (UTC)[reply]

Ctrl+Alt+Backspace kills X server and either leaves console or attempts to restart X. -Yyy (talk) 05:01, 21 July 2009 (UTC)[reply]

Also check out the Magic SysRq key. I've never had a lot of success with it, but it's supposed to work quite well for crash recovery. Indeterminate (talk) 00:43, 22 July 2009 (UTC)[reply]

Internet causing computer crashes.

Kinda long question-if this isn't the place could you recommend me a good forum at least?

I was cleaning a computer for virus and such (its pretty old and hasn't been running with any anti vir besides Windows defender). I ran Avira, Super Anti-Spyware, and A-2 Squared. Combined they found 5 virus (I don't have the computer in front of me so I can't say names). That solved one of the problems with the computer- that every time explorer.exe was opened the computer would crash and need to be restarted w/ power button.

However the internet still crashes randomly sometimes (regardless of browser) on random webpages. I can see the whole screen but it is completely frozen, so I need to restart it with the power button. Any ideas as to what could be causing it? Any way I could get the error report from my computer of the incident (I can't find Dr. Watson and I'm not sure how to use HijackThis! because it causes the computer to crash so I have to restart it). The problem is sporadic which really confuses me. I can't guarantee but I think it sometimes appears on sites with only text.

Secondly (not so important): what would you recommend as a good anti virus (pay or free-either way)? Out of Avira, Avast, AVG, Norton, McAffee, or anything else what would you recommend. I don't want it to be to hard on resources so I don't mind compromising some detections because I can run on-demand scanner (Super Anti-Spyware, A-2 Squared).

Thanks in advance! —Preceding unsigned comment added by 66.133.202.209 (talk) 00:38, 21 July 2009 (UTC)[reply]

I'd try Spybot and Ad-Aware on this - they may pick up spyware/malware infections missed by the other programs. Exxolon (talk) 01:45, 21 July 2009 (UTC)[reply]

Malwarebytes' Anti-Malware is pretty much a standard part of any anti-malware clean-up job. An additional tool that may be of use is Dr.Web CureIt! Ideally, when dealing with unstable, infected systems, critical files should be backed up and then bootable CDs used to remove malware without the hassle of active malware in memory causing problems with removal or system stability. Here's a good list of bootable anti-virus bootable CDs provided by various anti-virus vendors; most are free. As for an effective but light anti-virus, I find NOD32 fulfills that criteria rather well. I'd still recommend trying out the trials of various different anti-viruses, however. Hope this helps! :)--Xp54321^{(Hello! • Contribs)} 04:48, 21 July 2009 (UTC)[reply]

Limit the servers upload speed to ips

Is there a way to limit the download speed of people connecting to a web server, based on their ip address? Say for example the server has a poor upload speed, if one person downloads a lot of images at once they can make the entire site unreachable to others. Is there a way to do this? Thanks! —Preceding unsigned comment added by 82.43.91.128 (talk • contribs) 06:13, July 21, 2009 (UTC)

If you are using Apache, try mod_bandwidth - manya (talk) 10:14, 21 July 2009 (UTC)[reply]

I am but on a windows pc, and mod_bandwidth appears to only work for Unix

The term you want to search for is Traffic shaping or Bandwidth throttling, basically not so simple or cheap on windows. Use a separate linux box or possibly run a very small virtual linux instance on the same PC using Virtual PC, set it upo as the gateway and use traffic shaping on that. I also think there are new firmware for your router that might be able to help. --Stefan ^talk 03:05, 22 July 2009 (UTC)[reply]

DOS

hey can some one tell me how to make a batch file to del a particular FOLDER on computer.

is this correct

@echo off
C:XYZ/ABC/ del CDF
y
exit

(where ABC, XYZ, CDF are folders and i want to delete CDF...) —Preceding unsigned comment added by Sushil shenoy (talk • contribs) 14:44, 21 July 2009 (UTC)[reply]

Try this

@echo off         (Causes batch file to run silently)
c:                (Change to drive C)
cd \xyz\abc       (On current drive (now C:) move to folder \xyz\abc. Note \ not /)
del cdf /q        (delete files in folder cdf. Not all DOSes have /q (quiet))
rmdir cdf         (remove folder)
                  (batch file exits when it falls off the bottom.)

Make the batch file without @echo off. Make a folder, copy some files into it and try it. Alter to suit. -- SGBailey (talk) 14:53, 21 July 2009 (UTC)[reply]

Be very careful with deleting stuff via batch files; if one of the cd's doesn't execute correctly (due to typo or running the batch file from the wrong directory or whatever) you can easily inadvertently delete things. Many years ago I had a very bad accident with a batch file that regularly cleared out my IE caches (back when those needed to be manually cleared out!). SGBailey's looks fine of course but as you modify things, beware...! --98.217.14.211 (talk) 15:04, 21 July 2009 (UTC)[reply]

Yes, SGBailey's batch file is pretty unsafe because the "del" command will execute whether or not the "c:" and "cd" commands succeed. (Success might not occur if there was indeed a typo in either of the two lines, or if you simply forgot an intervening directory, or if for some reason your main drive mounted as the F: drive instead of C ... so it might not do what you want, which in this case would be bad if there happens to be a "cdf" subdirectory in the current directory.) A safer substitute would be

@echo off
del c:\xyz\abc\cdf /q
rmdir c:\xyz\abc\cdf

Tempshill (talk) 15:58, 21 July 2009 (UTC)[reply]

Removing or reducing reverberation

Is there a way using a currently available audio software package of removing or reducing the reverberation on a recording I have on cd? Thanks —Preceding unsigned comment added by 79.75.44.241 (talk) 16:50, 21 July 2009 (UTC)[reply]

I don't know of a package that does that - but it's certainly possible in principle. There is an analogous problem in computer graphics where you need to remove 'ringing' artifacts from lossily-compressed images. That problem is pretty much solved...so it must be possible in audio also. SteveBaker (talk) 02:07, 22 July 2009 (UTC)[reply]

What algorithm could in principle be used steve?--79.75.44.241 (talk) 13:45, 22 July 2009 (UTC)[reply]

The general class of algorithms you're looking for is probably Deconvolution algorithms. I'm not aware of specific ones used to reduce audio reverb. I suspect that this is not something you could do easily or with little artifacts, but It's entirely possible that I'm completely wrong. Artificial reverb (created in a computer) could probably be removed a lot cleaner than real, natural echos. APL (talk) 15:01, 22 July 2009 (UTC)[reply]

Indeed, reverb removal is an underdefined inversion problem. I have specifically worked on this topic for audio processing, and it is hardly a "solved problem." It is very similar to multipath interference (except that it is intentionally added because it "sounds nice"). Deconvolution, multipath detection, and nonlinear noise reduction are some keywords you might find helpful during a Google Scholar search, but I doubt there are any remotely-useful commercial or free software tools to do reverb reduction. If it is some other effect, and not reverb, your chances of removal improve significantly. If by a stroke of dumb luck or artistic creativity the original recording included the impulse-response (i.e. they have a really nice, solid, isolated snare drum beat that played through the reverb processor), you can use that to estimate the impulse response of the reverb. (This is very unlikely - reverb is not usually added to the entire composition; more likely it is added to individual instrument or vocal tracks, with different amplitudes, delays, "wet/dry" mix, etc., on each track). Also note that a sound engineer can easily change any of the reverb parameters throughout the track - so you have a nonstationary, nonlinear, underdefined multi-component inversion problem (the physicists in the audience are now cringing). Read up on reverb to get an idea of why it is complicated to remove. Nimur (talk) 15:25, 22 July 2009 (UTC)[reply]

Direct weblink

Is there a way to create a direct link to search results from this page? When I do a search for "keathley george", the results show up at http://www.abmc.gov/search/detailwwnew.php, but if you go to that URL directly, the information is blank. I'd like a direct link to the search results to use as a reference for an article. Thanks. — jwillbur 23:40, 21 July 2009 (UTC)[reply]

'Fraid not. They have opted for using a way of sending HTML variables that doesn't let you put them in a URL in any way. It's kind of unnecessary -- they would literally have to change one word in their code (albeit maybe a few times) to have it set up the other, more useful way. --98.217.14.211 (talk) 01:49, 22 July 2009 (UTC)[reply]

(edit conflict) I'm afraid there's no way to create a URL that goes directly to the search result. However, if your article is on a website, you can create a button that goes directly to the search result using the following HTML code

<form method="post" action="http://www.abmc.gov/search/detailwwnew.php">
<input type="hidden" name="ID" value="21473">
<input type="submit" value="Go">
</form>

Otherwise, it looks like you'll have to give the URL of the search page and explain what to search for to get the search result. --Bavi H (talk) 01:58, 22 July 2009 (UTC)[reply]

Thank you, Bavi. It's actually for a Wikipedia article, so a button won't work, unfortunately. But I am curious, how did you find the correct ID value? — jwillbur 05:08, 22 July 2009 (UTC)[reply]

Last time I checked WP:EL expressly discourages linking to search results. If you're trying to provide sources for others to expand the article you are better off pasting te URL onto the talk page with instructions as mentioned above. Zunaid 07:55, 22 July 2009 (UTC)[reply]

He doesn't want to link to a search results page—the site is a database, he wants to link to the actual entries. They don't let you do that. It's dumb. --98.217.14.211 (talk) 13:36, 22 July 2009 (UTC)[reply]

You can find the ID value by viewing the source of the page that has the "GO" buttons on it. Each of those is a little INPUT with an ID value. --98.217.14.211 (talk) 13:35, 22 July 2009 (UTC)[reply]

Yes, jwillbur, I did exactly what 98.217.14.211 said. Once I was on the page with the Go button, I used the View Source command in my web browser, then searched for the word "Go" (match case) to find the proper form code. --Bavi H (talk) 00:55, 23 July 2009 (UTC)[reply]