Wikipedia:Reference desk/Archives/Computing/2007 November 19
From Wikipedia, the free encyclopedia
Computing desk | ||
---|---|---|
< November 18 | << Oct | November | Dec >> | November 20 > |
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] November 19
[edit] Cryptography: Which is more secure for authentication: RSA or DSA?
I'm doing some personal research learning about public key infrastructure and asymmetric cryptography. After Googling a bit, I thought I'd submit a question which I've been struggling to find a answer that satisfies me. I think the Computing category would be better suited to post this than the Mathematics category. Which cryptographic algorithm is more secure for authentication, specifically, key exchange and digital signing: RSA or DSA? -- PaperWiki (talk) 02:20, 19 November 2007 (UTC)
- I don't know whether one is in theory more secure but AFAIK neither one has been compromised, so they're both 100% right now --ffroth 15:38, 19 November 2007 (UTC)
- RSA can be broken by a quantum computer of sufficient bit size. No quantum computers with enough bits to break an RSA cipher of any realistic size could be created so far. —Preceding unsigned comment added by 84.187.67.90 (talk) 18:40, 19 November 2007 (UTC)
- I thought quantum computers don't work with bits =_= --ffroth 02:06, 20 November 2007 (UTC)
- They are digital devices - so they deal in bits - but they do their calculations using qubits which is a quantum superposition of many possible states. It's true that we don't yet have usable quantum computers - but when we do, RSA will become highly vulnerable to attack by even fairly small quantum computers. With conventional computers, you can double the complexity of cracking a code by adding one bit to the length of the key. With quantum computers, you have to double the number of bits to double the time it takes to crack it. So a 64 bit code takes a regular computer over four billion times longer to crack than a 32 bit code - but a quantum computer will only take twice as long. Sheer brute force makes our present codes unbreakable - but we're going to have to come up with something much cleverer in the future. SteveBaker (talk) 03:33, 20 November 2007 (UTC)
- As far as I know, DSA is also vulnerable to quantum attack, but I don't think any attack against ECC is known. There are definitely no known quantum attacks against any symmetric cipher (like DES or AES). Grover's algorithm is far too inefficient to be practical. But all of this is meaningless unless someone finally manages to build a large quantum computer. -- BenRG (talk) 12:46, 22 November 2007 (UTC)
- They are digital devices - so they deal in bits - but they do their calculations using qubits which is a quantum superposition of many possible states. It's true that we don't yet have usable quantum computers - but when we do, RSA will become highly vulnerable to attack by even fairly small quantum computers. With conventional computers, you can double the complexity of cracking a code by adding one bit to the length of the key. With quantum computers, you have to double the number of bits to double the time it takes to crack it. So a 64 bit code takes a regular computer over four billion times longer to crack than a 32 bit code - but a quantum computer will only take twice as long. Sheer brute force makes our present codes unbreakable - but we're going to have to come up with something much cleverer in the future. SteveBaker (talk) 03:33, 20 November 2007 (UTC)
- I thought quantum computers don't work with bits =_= --ffroth 02:06, 20 November 2007 (UTC)
- RSA can be broken by a quantum computer of sufficient bit size. No quantum computers with enough bits to break an RSA cipher of any realistic size could be created so far. —Preceding unsigned comment added by 84.187.67.90 (talk) 18:40, 19 November 2007 (UTC)
[edit] PCI video cards
I have an old Gateway PIII PC I'd like to get working. I pulled the graphics card a while back to clean it and see if it was PCI (to try to diagnose a video card or other failure in a new rig). It turned out to be AGP. The motherboard's AGP alignment is off with the case, so while I can get it to work by fiddling around when the case is open, once I close it it never works correctly. Whoever put it in and got it to boot was apparently blessed with divine powers, and now it just plain won't work. I'd like to get a cheap secondhand PCI card, just to make the box boot and be able to do basic things like word processing and Internet (there is no on-board video). Nonetheless, since PCI graphics cards are so old and slow, I figure I can get a pretty decent one at the same price as a mediocre one. Does anyone have any suggestions for higher-end (of their day) graphics cards that are PCI and not AGP? I was looking at the GeForce 5 cards but our article does not specify if they are exclusively AGP or have PCI versions. -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 05:15, 19 November 2007 (UTC)
- PCI graphics cards are a pretty rare commodity these days. Even AGP is starting to become outdated. Everything's moving towards PCI-E now. You might be able to find one used, perhaps on eBay or something. — User:ACupOfCoffee@ 05:45, 19 November 2007 (UTC)
- Yeah, I understand - I'm just trying to find the highest-end PCI card for its day, specifically a model name, because I figure as they're all old I should be able to get them at roughly the same price. -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 05:56, 19 November 2007 (UTC)
- I purchased a PCI video card from Wal-Mart just a few months ago to repair an older PC for a relative. [1] --— Gadget850 (Ed) talk - 15:22, 19 November 2007 (UTC)
- Personally - since this is a 'junker' PC - I'd take a hacksaw to the case and if necessary use duct-tape to get the AGP card to stay put! Enlarging the place on the back of the case where the video connector comes out should be pretty easy - and then you have $0 solution that'll almost certainly be a lot faster than any PCI card you could buy. Remember - it's not just the speed of the graphics card - it's the rate you can give it work to do that matters. The PCI bus is unbelievably slow compared to even 1x AGP (and you might have 2x, 4x or even 8x AGP). In all likelyhood, it's irrelevent how fast the graphics card is because it'll be spending most of it's time sitting there starved for data. Even a slow AGP card will likely beat out a fast PCI card. (Caveat: This is a gross generalisation - a lot depends on...um...everything really!) SteveBaker (talk) 16:26, 19 November 2007 (UTC)
-
- Check the clearance around the video card, it might press on a capacitor or something when the case is closed.Polypipe Wrangler (talk) 05:47, 23 November 2007 (UTC)
- Check, for example this search on Newegg. It's just a "Power search" of all video cards that match the criterion of Interface = PCI. Also, the Fx 5200 has something of a reputation in this field, as mentioned above. 68.39.174.238 (talk) 23:20, 24 November 2007 (UTC)
[edit] Windows XP: File | Save As
Windows XP: File | Save As
In the dialog box of any program in Windows XP when you select: File | Save As
There is an option in the "View Menu" to select "Thumbnails"
Is there any registry tweak to make "Thumbnails" the default choice?
If so, what is the tweak?
Also, is there any way to change the "default size" and "default location" of the dialog box?
[edit] multimedia
what are the server requirements of distributed multimedia systems
- presumably you mean video for your multimedia. It needs a high bandwidth for the network connection, and disk drive connections. If you need to run 24*7 365 days a year, you will need an operating system that does not need to be restarted (for whatever reason). You may need to take an analogue video input and convert it to mpeg2 or something like it. There has to be a way to load up the new content. And perhaps you will need digital rights management for your content. A distributed system will have a lower demand than a single central server, but it will be much more difficult to keep the content loaded. Graeme Bartlett (talk) 10:58, 19 November 2007 (UTC)
- Multimedia is far to fuzzy a term. You might mean some still images and some simple JavaScript to animate them - in which case your server-side requirements are minimal. You might mean still images plus audio or flash animations or host-side PHP or other programming - or you might mean full-up streaming video. The amount of traffic you expect to get is also a concern. My ancient home web server is a 600MHz PC with a single, very slow hard drive and nothing but a DSL connection to the net. You can get streaming video off of it if you're the only person using it - or it could manage dozens of simultaneous users for some JavaScripted game or something that only requires a few images to be downloaded. At the other end of the scale, consider something like YouTube that serves 100 million streaming videos per day and pays a million dollars a year in bandwidth costs alone! We can't possibly answer your question without MUCH more information. SteveBaker (talk) 16:14, 19 November 2007 (UTC)
[edit] Books about Web 2.0
Can someone tell me some books about Web 2.0??? I am from brazil so those books can be in Portuguese or english. Exdeathbr (talk) 14:05, 19 November 2007 (UTC)
[edit] Finding authors of deleted youtube videos
I have a youtube video bookmarked that was deleted by the user, i have tried delutube but to no avail, is there a way of finding user who uploaded the video just by looking at the video id code? thanks Jutwdev99 (talk) 14:57, 19 November 2007 (UTC)
- No. Try archive.org? --ffroth 15:37, 19 November 2007 (UTC)
- Depending on when it was deleted you could check the Google cache of the page. Exxolon (talk) 23:12, 19 November 2007 (UTC)
- Also you might try searching for the name of the video in YouTube or Google Videos. Often videos are mirrored by other users. --24.147.86.187 (talk) 23:28, 19 November 2007 (UTC)
[edit] Unix batch renaming of files to remove illegal Linux characters
I'm using Mac OS X connected to a Linux server. Some Mac file names have characters that Linux won't allow. I'm looking for some clever speedy Unix terminal command to look at a folder of files and batch rename all illegal Linux characters (like : \ " > ’ ? |) into normal hyphens. Any ideas? --24.249.108.133 (talk) 19:23, 19 November 2007 (UTC)
- The only illegal characters in a Linux file name are "/" and "\0". Everything else is legal, if ugly to work with. That said, I've used the following script for years to fix up unpleasant file names. Just save it to a file, and do a:
perl -w this-script.pl *
- in your directory of bad files. It tries hard to do the right thing, but you should probably back up your files first anyway. --Sean 19:55, 19 November 2007 (UTC)
#!/usr/bin/perl -w
use strict;
for (@ARGV)
{
unless (-e)
{
warn "$0: '$_' doesn't exist, skipping\n";
next;
}
my ($dir, $orig_file) = m#^(.*/)?(.+)$# or die $!;
$dir = './' unless defined $dir;
$_ = $orig_file;
s/%([\dA-Fa-f]{2})/sprintf '%c', hex($1)/ge;
s/[^\w._-]+/-/g;
s/[-=_]+/-/g;
s/^[-=_]+(.)/$1/g;
s/-*\.-*/./g;
next if $orig_file eq $_;
my $i = 0;
my $fname;
for ($fname = $_; -e "$dir$fname"; $fname = "$i-$_")
{
$i++;
}
$orig_file = $dir . $orig_file;
$fname = $dir . $fname;
print "rename '$orig_file' => '$fname'\n";
rename $orig_file, $fname or die "rename '$orig_file', '$fname': $!";
}
- I'm in bash mode and Terminal doesn't seem to like your code. What am I doing wrong? --24.249.108.133 (talk) 22:42, 20 November 2007 (UTC)
-
- As Sean said, there are very few truly illegal filename characters in Unix and Linux. (There used to be an issue when interoperating with Macs, since Macs used : as their pathname character but allowed / in ordinary filenames. But now that Mac OS X is Unix-based, that's not an issue any more.)
-
- With that said, filenames with "funny" characters in them can certainly be a real nuisance to work with. Here's a simpler script I just whipped up to take care of the characters you mentioned:
-
-
for f in `ls | grep "[:\\\\\\">'?|&]"`; do mv "$f" "`echo \"$f\" | sed \"s/[:\\\\\\">'?|&]/-/g\"`"; done
-
-
- Another character you might be bothered by is the space. Here's a variant of the above script that fixes spaces, also:
-
-
ls | grep "[:\">'?|& ]" | while read f; do mv "$f" "`echo \"$f\" | sed \"s/[:\\">'?|& ]/-/g\"`"; done
-
-
- Unfortunately, this second script can't deal with the backslash \ character.
-
- One more thought: you might want to be careful about blindly transliterating filename characters to '-'. For example, if you have a file named ":r", and you rename it to "-r", and you later type "rm *" in that directory, you're apt to end up running "rm -r" my mistake, which might not be what you wanted. —Steve Summit (talk) 18:01, 24 November 2007 (UTC)
-
- [Exercises for the student: Why can the first loop not work for spaces? Why can the second not work for backslashes? Does the second loop work for all spaces, or only some spaces? Why did I use sed, and not tr? Why are there six backslashes in a row at two spots in the first loop? There are some deep waters here... —scs]
[edit] Stripping an MP4
Does anyone know a program that can easily stip ALL tags and metadata off an MP4 (specifically audio only i.e. M4A) and leave just the stream in an MP4 container? The reason I ask, is when I convert a particular type of file (best not mention for legal reasons - it probably doesn't matter anyway) the output M4A doesn't work with my Nokia 6300 (normal ones do). When I use VLC to put the stream in a MP4 new container, the phone will then play the file, but if I then add tags (even with Nokia's own software), the phone won't recognise any tags on the file (which it's supposed to). Any help would be appreciated - EstoyAquí(t • c • e) 21:23, 19 November 2007 (UTC)