Wikipedia:Reference desk/Archives/Computing/2010 August 15
Computing desk | ||
---|---|---|
< August 14 | << Jul | August | Sep >> | August 16 > |
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. |
August 15
[edit]Online back up services
[edit]From what I've been reading, online backup services seem to be slightly pointless for anything more than a few GB worth of data. Let's say I have 1TB to back up and a home DSL connection. I haven't done the math but I've seen estimates that say that 3-4 GB can be uploaded in a day. On the slow end, that's nearly a year for the initial backup. So do any of these services offer the ability to physically send a drive full of data to them for the initial "upload"? Dismas|(talk) 05:24, 15 August 2010 (UTC)
- Some quick math yields an estimate of 7.9 gigabytes per day uploaded for a person with a 768 kilobit per second upstream DSL connection, which would be 129 days for the initial round-the-clock upload. Yes, that's hilariously slow if you actually have 1TB to back up and you have that slow of an Internet connection. Of course, you would choose to slash your amount of backed-up data by choosing to not back up your system and application directories. I disagree quite a lot with "slightly pointless", though. It's only pointless until the initial backup is complete. Then it's very pointful. Comet Tuttle (talk) 06:10, 15 August 2010 (UTC)
- I think I've read something on the RD before where someone mentioned such a service. Actually it may have it was something like some sort of hosting service which apparently do do this sort of thing or perhaps Amazon Cloud [1] which does as well.
- In any case for things specifically designed as online backups, a quick search reveals to me [2], [3], [4], [5], [6], (I think) [7], (I think), [8], [9] appear to have this feature. [10] do as well but I think they're a photo backup service not a general purpose one. [11] may as well (saw suggestions of it but can't find any mention on their site from a quick look).
- Comparison of online backup services has a tab for the opposite (sending you a disk to recover a backup) but does mention physical seeding for at least one provider. I would guess many more have such a feature this is just the results of a quick search. And do note I'm not recommending any of these, in fact one of them seems to be using free GoDaddy hosting which doesn't exactly inspire confidence that you should trust their service with your backups.
- If you're wondering, I found the first result from discussion of such a feature, found out it was called 'physical media upload' from that page and searched for more with such a feature. During that, I found out it was also called 'physical (media) seeding' (far better name IMHO).
- You may notice that many of these are more directed at businesses then home users, while this may seem counterintuative at first, I'm not really that surprised. Most home users likely don't have enough they want to backup to make it worthwhile particularly given the cost of keeping that much backuped online. I presume there's also a charge for physical seeding in addition to the requirement for a external hard drive and the cost and risk of sending it which also make it less worthwhile for home users who may not mind the wait since unless you live here in NZ or other places with very high data charges, even if it takes a few weeks it'll be basically free (well unless the service charges a bandwidth fee).
- Note that a number of them appear to require you to use their hard drives, it makes it easier I guess since they don't have to worry about funny drives with compatibility problems and also ensures you have the necessaring packaging and they're likely have less hassle if the hard drive is damaged when they get it.
- Nil Einne (talk) 06:56, 15 August 2010 (UTC)
- Wow! Thanks for that! I didn't know what to search for because I didn't know what it was called. Thanks again, Dismas|(talk) 07:43, 15 August 2010 (UTC)
Call to improve password security?
[edit]There was again much fuzz in the media about supercomputers becoming more affordable, and that it makes brute-forcing passwords much easier. However, there is a very easy way to make brute-force completely useless: have a maximum number of allowed tries before blocking the account for some time. This is why PIN codes for credit cards can be only 4 digits long. For online or other personal computer-related use this could cause the possibility of someone blocking the accounts of others by deliberately entering wrong passwords, but even this can be easily solved: don't block the account, but have a cooldown period between accepting consecutive tries. Even forcing to wait for one second after entering a wrong password can thwart any brute-force attack, one second is big enough to hinder algorithms which are based on trying millions of combinations every second, but small enough that it doesn't disturb human users. So why is there always so much panic around passwords, and forcing them to make longer and longer, instead of everyone implementing a system like this? --131.188.3.21 (talk) 09:41, 15 August 2010 (UTC)
- Blocking attempted access after a number if wrong tries is very common. However, brute force attacks rarely work like this - the normal target of brute force attacks is downloaded password files. These are stored encrypted, but are susceptible to long timescale attempts to break them. --Phil Holmes (talk) 09:57, 15 August 2010 (UTC)
- If you mean online passwords (e.g. for Facebook) then the speed and latency intrinsic to the system makes brute force impractical and supercomputers irrelevant. Attackers instead cast a broad net, trying dictionary attacks of common passwords. It's tempting to add a cooldown system to this, but with internet scales - saying "if you make three bad attempts, the delay before you can login goes to five minutes, and then 30, and then 300, and so forth", but that has two major problems - firstly the attacker attacks thousands of accounts concurrently, so they can be off attacking all the others while the timer runs on one (so you haven't slowed the attacker down at all) and secondly such schemes afford a great opportunity for denial-of-service - my cheapo botnet can occasionally try a random (wrong) password on your facebook account, meaning you essentially never again can login to facebook. Other online schemes, like ssh, already have a cooldown. -- Finlay McWalter • Talk 10:02, 15 August 2010 (UTC)
- On your last question, about "panic around passwords", Bruce Schneier generally seems pessimistic about password security; he famously advocates that people should have difficult-to-guess passwords but actually write them down and keep the piece of paper in their wallet, because that's hard to steal. Here is an essay of his about the insecurity of all passwords, and the observation that phishing attacks are the attack of today that avoids the need to ever crack a password. Comet Tuttle (talk) 22:15, 15 August 2010 (UTC)
wget
[edit]I mirrored a lot of pages with the -m -k options, but wget crashed just before it converted them to relative links, and the pages I downloaded are now gone from the site I downloaded them from, so I can't run the download again. Everything has been downloaded, it's just all the html files still point to online locations instead of the downloaded ones. Is there any way to make wget convert the links? Or some other program that might? 82.44.54.4 (talk) 11:17, 15 August 2010 (UTC)
- sed can almost certainly do it - rather easily, if all files are from the same domain.
sed -e 's:full_http_server:root_of_the_local_dir:g' infile > outfile
. Make sure the locations are properly escaped (using : instead of / saves you from having to escape all /s ;-). I'd make sureoutfile
is in some mirrored directory so as to leave the original files save and sound. --Stephan Schulz (talk) 07:12, 16 August 2010 (UTC)- That will work for some websites, but it's pretty fragile. It will fail if the site uses relative URLs of the form "/foo/bar", and in various other situations. -- BenRG (talk) 05:38, 17 August 2010 (UTC)
- I agree that that's a simple 95% solution, but I don't see why it will fail for the use case you described. Relative links should not be a problem to begin with, as they don't point to "online locations". Do you mean local absolute links? That's indeed not something I even knew existed, though yes, that would break a simple text replacement approach. --Stephan Schulz (talk) 07:56, 17 August 2010 (UTC)
- That will work for some websites, but it's pretty fragile. It will fail if the site uses relative URLs of the form "/foo/bar", and in various other situations. -- BenRG (talk) 05:38, 17 August 2010 (UTC)
- Thanks I'll try that 82.44.54.4 (talk) 11:24, 16 August 2010 (UTC)
What font you are viewing?
[edit]Is there any quick way of telling what font you are viewing in a web browser? (I prefer to allow sites to set their preferred typeface.) When I copy and paste into a word processor it just uses the word processor's default. I know you can view page source and dig into the css...was wondering if there's an easier way. 151.203.20.165 (talk) 13:29, 15 August 2010 (UTC)
- Firebug (web development)'s Inspect feature can do this - if you click on an element, it will list all applied CSS styles, making it easy to pick out the fonts. Other browsers may also have comparable extensions or Developer modes which allow this. Unilynx (talk) 13:36, 15 August 2010 (UTC)
- Thanks; Firebug seems to work quite well. 151.203.20.165 (talk) 14:04, 15 August 2010 (UTC)
Device drivers and hard disk data recovery
[edit]Firstly, a recent hard disk failure, I ordered a new hard drive and installed it in my computer, then proceeded to install Windows XP Professional (service pack 2) onto the new hard drive. This got my computer up and running again...mostly. Windows works, and I've been able to get some things functioning...the only problem is that none of my computer's hardware seems to be installed anymore. As far as my computer knows, it has no internal network card and no internal graphics card (or at least not the one installed and working before the crash) or sound card, causing me to have very limited graphics options and no sound or internet options/devices in Control Panel. I'm guessing that this is because the device drivers need to be reinstalled (correct me if I'm wrong). However, the Dell CD labeled "Drivers and Utilities" evidently (as I discovered upon running it) has nothing on it except a computer diagnostic problem and not the "drivers and utilities" I need. Where do I obtain the drivers (again, assuming this is why half my computer hardware doesn't seem to exist) and get all my hardware working again?
After this is done, I would like to recover the data from the non-working hard drive and copy it all (including installed programs, data within those programs, files, pictures, etc.) onto the new hard drive. The hard drive originally failed when I was running Windows Live Messenger, Mozilla Firefox, and GR2Analyst (see previous posts). I had installed the radar program that day, but it was running just fine for the few hours up until the hard disk failure. I'm fairly certain that the old hard drive did not suffer a head crash, because when I plug it into a SATA thing (this thing is the one I have) the hard disk spins without the click of death (it sounds like it's running normally), but I could be wrong considering my inexperience with such issues. Previously when trying to start up the hard drive while it was in the computer, it would start up to the Windows screen where the little blue bars scroll across the screen ([12]), which would sit there for about 5 minutes before displaying a blue error screen that said among other things "UNMOUNTABLE_BOOT_VOLUME", and automated diagnostics on the hard drive gave the error codes "Error 0142 Msg code 2000-0142 Unit 1 Self test status 79" (or something similar), "0F00:0750 Disk_0 Self-test Read Error" and "00F0:0244 Disk_0 Block 6468425 Can't read, replace disk or remove write protection". When I plug the old hard disk into the computer using the SATA cord, the computer recognizes it by saying "your new hardware is installed and ready to use", but it doesn't show up in Windows Explorer. Given all of this information, what might be wrong with the old hard drive, and what can I do to copy the information from the old hard drive to the new hard drive? Sorry about the huge wall of text, and a big thanks to anyone who makes it through it all and answers. There also might just be a barnstar or two in it for the answers that end up working (one for the first paragraph, and another for the second). Ks0stm (T•C•G) 16:35, 15 August 2010 (UTC)
- Assuming your actual PC is a Dell (based on the fact you have a Dell driver CD) the first part of your question should be quite easy. Just go to Dell's website and click technical support in the top right corner and then home users and finally drivers and downloads (or just click here!) Choose to enter a service tag and type in the code printed somewhere on your computer (either on the back or under a front flap are the usual places. They'll give you a page with ALL the drives for your computer for you to be able to download. Unfortunately sometimes the same model of computer would have come with a few choices so they'll offer multiple network card drivers so unless you know exactly which one is yours it'll be trial and error, but it shouldn't take too much longer.
- The second part of your question is a bit more tricky, but the fact Windows started to boot up previously is promising. When you plug it in and it says "Installed and ready to use" go into Control Panel and Administrative Tools and then Disk Management. You should see the new disk as on of the options in the bottom right and hopefully it'll be just as simple as it doesn't have a drive letter so you can simply right-clicking this and select "Add/change drive letters" and then just giving it a drive letter. After doing so HOPEFULLY it'll appear in "My Computer" ZX81 talk 16:55, 15 August 2010 (UTC)
- To see what hardware MS-Windows thinks your computer has got (from memory, might not be correct),
- Right-click on "My Computer" and select "Properties..." (or press Windows-Break)
- This should bring up a dialog box with several tabs, select "Hardware"
- There should be "Device Drivers" button near the top, press it.
- A new window will open, titled "Device Manager", with a tree view listing all the hardware.
- Any hardware with a question mark in a yellow circle does not have the correct driver.
- Right-click on the hardware, and select "Properties..."
- A new dialog will open, select "Install Drivers..."
- A file selection dialog box will open, select the .inf file you have downloaded for that device, and follow any instructions displayed.
- repeat for the other devices.
- CS Miller (talk) 18:45, 15 August 2010 (UTC)
echo time
[edit]![](http://upload.wikimedia.org/wikipedia/en/thumb/f/fb/Yes_check.svg/20px-Yes_check.svg.png)
How do you echo the time in a bacth file? I tried "echo time /t" and it didn't work. I just want it to display the time, not log it to a file or anything. 82.44.54.4 (talk) 19:58, 15 August 2010 (UTC)
- Try echo . | time /t ZX81 talk 20:20, 15 August 2010 (UTC)
- echo . | time (without the /t) was the old way of doing it, before the /t switch became available. -- 78.43.71.155 (talk) 21:10, 16 August 2010 (UTC)
- It said "the process tried to write to an non-existent pipe" 82.44.54.4 (talk) 20:22, 15 August 2010 (UTC)
- You don't need to echo anything: you just need to execute the "time" program: time /t in the batch script. Nimur (talk) 20:56, 15 August 2010 (UTC)
- That works 82.44.54.4 (talk) 21:17, 15 August 2010 (UTC)
- You don't need to echo anything: you just need to execute the "time" program: time /t in the batch script. Nimur (talk) 20:56, 15 August 2010 (UTC)
- Alternatively, you can do echo %time% --Bavi H (talk) 23:30, 15 August 2010 (UTC)
network cards
[edit]Do network cards and routers and such eventually wear out with use, like say a hard drive would? Or could you download at the maximum rate for years without any degradation? 82.44.54.4 (talk) 20:15, 15 August 2010 (UTC)
- Hard drive failures are nearly always mechanical in nature. With finely balanced parts rotating at high speeds and close tolerances between head and disk, it is hardly surprising that hard drives are suceptable to mechanical problems. Network cards and routers are made of solid-state components that, in a perfect world, would last indefinitely. However, manhandling the card, changes in temperature (including that caused by a build-up of dust), power surges, etc. can all damage electronic components. Astronaut (talk) 00:40, 16 August 2010 (UTC)
- Actually, electronic components do wear out as well. Apart from them suffering from manhandling damage, the moving parts in a circuit (the electrons) cause the compnents to wear out by a process called electromigration. How fast this happens depends on the current densities used in the circuits and how hot they get - so it's not easy to predict. It could be a few years or many tens of years.--Phil Holmes (talk) 09:37, 17 August 2010 (UTC)
- A common and well known point of failure of some electronic components in recent years is leaking electrolytic capacitors or as our article calls it, capacitor plague. This primarily affects motherboards, video cards and other things which deal with resonable amounts of power particularly I think if that includes DC-DC conversion. It's probably not that likely to occur in a network card or router I would guess Nil Einne (talk) 08:37, 19 August 2010 (UTC)
- Actually, electronic components do wear out as well. Apart from them suffering from manhandling damage, the moving parts in a circuit (the electrons) cause the compnents to wear out by a process called electromigration. How fast this happens depends on the current densities used in the circuits and how hot they get - so it's not easy to predict. It could be a few years or many tens of years.--Phil Holmes (talk) 09:37, 17 August 2010 (UTC)
Weird PDF
[edit]This PDF has some weird things going on it.
For one thing, I can't (on a Mac, with multiple PDF programs) search it at all. It doesn't seem to have any security bits relating to this set when I look at its properties in Adobe Reader.
When I try to copy and paste text, I get just gobbledegook as a result. I would paste some but it seems to be killing the Wiki editor with its crazy gremlin characters.
What's going on here? I'm both 1. just curious (I haven't run into this before), and 2. trying to search it for a reference and am frustrated! Is it intentional obfuscation, or is it an artifact of the PDF producing program? (Apparently it was created with GNU Ghostscript 7.05, which I wouldn't think would cause a problem.) --Mr.98 (talk) 23:21, 15 August 2010 (UTC)
- I downloaded it on my Mac using Safari (both current), and it opens fine. You seem to be seeing all of the formatting characters that Acrobat uses to structure text (that's the gobbledygook), which makes me think that (a) you have a corrupt download or (b) your version or reader is out of date or its plist is corrupt. try downloading it again. (though I'll say, from a brief glance at the text it reminds me of a cartoon I once saw: an alert dialog that says "This document has 2,432 spelling errors and is really boring. Print anyway?" - maybe your computer is engaging in civil disobedience...) --Ludwigs2 00:16, 16 August 2010 (UTC)
- Ludwigs2, I believe you're missing the point. Just looking at the pages in a pdf viewer doesn't reveal the abnormality. You have to try to do something that requires the actual text, and not just the appearance of text. This file seems to be using a technique similar to the one described at http://spivey.oriel.ox.ac.uk/corner/Obfuscated_PDF - it contains a bunch of custom fonts, in which each character looks like a completely different character than the one it's supposedly representing. 69.245.226.104 (talk) 00:36, 16 August 2010 (UTC)
- Right, I should clarify. I can read it fine. But I can't copy/paste or search. I, the human, can read it, but the computer cannot. (And it is not a scan.) Anyway, I suspected that something like what 69.245 describes is going on here, but it seems odd to me that this particular document would be obfuscated? --Mr.98 (talk) 00:45, 16 August 2010 (UTC)
- It does say on page 17, "No Derivative Works - You may not alter, transform, or build upon this work." The obfuscation is just typical copyright-fanatic behavior. "People might actually use this instead of passively reading it! Must add obstacles!" 69.245.226.104 (talk) 01:00, 16 August 2010 (UTC)
- It's not just obfuscated on Mac/Safari. I get similar looking crap in Windows/IE 8. IMHO, it does seem to show an overly paranoid approach (perhaps due to fears of plagarism). I notice it is a very long document of nearly 500 pages, so if you really need to work with that document why not email Dr Maret and ask her for a clean copy. Astronaut (talk) 01:04, 16 August 2010 (UTC)
- Ah, I see what you mean. Well, it doesn't seem that the document is protected in any way. however, I do notice that it is using custom fonts (I don't know why, since the fonts are not that attractive), and there may be some issue with translating to standard fonts. I don't know if that's accidental or intentional. --Ludwigs2 01:31, 16 August 2010 (UTC)
- It's definitely intentional. I extracted one of those fonts and looked it over. It's too abnormal to be anything but intentional obfuscation. All the glyphs are above U+10000, in an order almost but not quite like ASCII. 69.245.226.104 (talk) 01:35, 16 August 2010 (UTC)
- Yes, and the PDF markup inside it is also mechanically obfuscated (or is the product of some downright whack processing) - rather than the long paragraphs of text interspersed with occasional markup, it has a "move X emit char Y" line for every single character. While this breaks simple text extraction and copy-and-paste, it also breaks any hope of accessibility to the blind and partially sighted, loses any chance of search engines meaningfully indexing it, and breaks the embedded URL on p17. Duh. -- Finlay McWalter • Talk 16:26, 16 August 2010 (UTC)
- And all that for a result that any moderately useful OCR software can undo, especially given that the image is available in perfect electronic form... --Stephan Schulz (talk) 16:55, 16 August 2010 (UTC)
- Bizarrely, I was able to find it in Word format, here. Google Scholar, interestingly enough, can seemingly parse the data. --Mr.98 (talk) 22:45, 16 August 2010 (UTC)