Wikipedia:Reference desk/Archives/Computing/2007 October 31

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Computing desk
< October 30 << Sep | October | Nov >> November 1 >
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.


October 31[edit]

Weird IE6 vs. Firefox issue[edit]

I'm trying to test out a page in IE6 and Firefox, both running on a Windows XP virtual machine running on Parallels Desktop in Mac OS X (whew!). Firefox has no trouble accessing the internet in this setup; Internet Explorer however gives me a "403 Forbidden" error:

Forbidden
You don't have permission to access http://asdkjaskdas.com/ on this server.
--------------------------------------------------------------------------------
Apache Server at asdkjaskdas.com Port 8100

I included the message to show that it gives the same error no matter what crazy, non-existant site you put in (including, of course, realistic ones). It thinks everything is an Apache server and every Apache server rejects it. So, uh, what's up? Again, Firefox connects fine (yay), so it isn't the connection. Note that "just use Firefox" is not a useful answer; I don't normally use IE but I need to test out to make sure a page I am working on renders reasonably correctly in it, since it has, you know, +90% of the marketshare. Note that I'd rather use IE6 here (not IE7) for the purposes of testing the page. --24.147.86.187 00:06, 31 October 2007 (UTC)

OK, fixed! It turned out that IE thought it was trying to connect with a proxy. God knows why. The 8100 should have tipped me off. --24.147.86.187 00:18, 31 October 2007 (UTC)

Booting stuff[edit]

I'm having a bit of a problem booting Windows. About a week ago, my computer was really having trouble getting going - it said "keyboard error" for some reason, but it was just Windows not wanting to load. After restarting it a couple times, I finally got it to load, with the intent of buying an external hardrive, moving all my files there, and then re-installing Windows. Well, the computer got turned off by accident, and I can't for the life of me get the stupid thing to start again. All the files are still there - I can see them via an Ubuntu live disk, although I do not have the privileges to access them - and the hard drive itself is fine. It just won't boot. Every time I start the computer, it says that there are no active partitions. I've looked them over through Ubuntu, but I can't manually activate my main NTFS thingie.

Any idea of what to do? I've gone through the whole "setup" thing with Dell, but it hasn't done any good. Am I supposed to be booting directly from the NTFS section, or is there something I'm missing? Can I fix this on my own? (I really don't want to pay Computer Geex twenty bucks to fix it...)

Any help would be greatly appreciated. --69.144.233.96 03:11, 31 October 2007 (UTC)

[1]. Run from the Windows setup disk in the recovery console:
  1. diskpart
  2. list partition
  3. select partition n
  4. active
? -- Consumed Crustacean (talk) 04:27, 31 October 2007 (UTC)

If I remember correctly I think a knoppix live disk will let you access your files. So I'd try that before you make any attempt to fix windows. Theresa Knott | The otter sank 10:00, 31 October 2007 (UTC)

When I had trouble getting my computer to boot because of a flawed windows program, I simple slaved the not working hard drive to another working hard drive that had a good copy of windows on it, moved everything I wanted to keep to the working hard drive, reformated the flawed disk and reinstalled windows. Or you could simply wipe windows off the flawed disk and keep it working as a slave with all your files still on it. These things worked for me, don't know if they would be much help to you but good luck. --NannySamdi 20:07, 31 October 2007 (UTC)

pfSense traffic shaping not (really) working[edit]

Is there any way to alter the traffic shaping of pfSense so that my ping on TF2 doesn't go to 500+ instead of the ~100 range when there's other traffic on there? I've gone through the wizards and put all steam ports into priority 6 queues, and the traffic are showing up there while the other traffic is in a priority 1 queue. However my ping still goes to an unacceptable level. Is there any way I can improve this? --antilivedT | C | G 07:45, 31 October 2007 (UTC)

Old hardware![edit]

I have a old 66mhz 486DX that's been sitting around for a very long time (obviously!). Turbo button, 5 1/2 floppy drive, the works. I used it when I was a kid then deserted it for a brand new 800mhz Pentium III (oh my god!) quite a few years later. Does anyone have any suggestions for creative uses of such old hardware? I can't turn it into a file server or anything for obvious reasons. Also, the ubiquitous question: can it run Linux? I've got Win 3.1 on there now, not sure if such an old machine can keep up no matter what distro I use. Is there one suited to really, really old computers? Or does the refdesk in its infinite wisdom have some other ideas? -Wooty [Woot?] [Spam! Spam! Wonderful spam!] 10:25, 31 October 2007 (UTC)

If it has 16Megs of memory, there are certainly old versions of Linux that it'll run. If you only have 8 Megs then it'll run old versions of Linux without the X-windows interfaces. Linux did run on 4Megs right at the dawn of time - but it was never very good with so little. Modern versions of Linux need vastly more memory than it's likely to have - so you need to dig out a version of Linux from at least 10 years ago. The tricky part is going to be getting the software onto the thing. Without network or CD drive, you're going to have to put Linux on a bunch of floppies and install it like that. I used to run Linux on a machine of about that vintage - so it most certainly can be done if you have enough RAM. SteveBaker 13:19, 31 October 2007 (UTC)
Linux will run on 386s and up. If you have low memory the trick is to get swap enabled before the install procedure starts, here is how you did that in slackware. Of course the easiest thing would be to install the HDD in another computer, do the setup, and return it to your 486, then you won't have to worry about disks and stuff. I know Slackware will work, just install the recommended packages in the A and maybe AP disk sets and pick and choose the other stuff you want. FOr fun you can try to compile the kernel and see how many days it takes. -- Diletante 16:44, 31 October 2007 (UTC)
I've installed a recent version of Gentoo Linux on a Pentium 233 with 16MB of RAM. Keep in mind that Gentoo installs software by compiling it from source code -- you should only use it if you don't mind it taking a week or two to install the OS. Gentoo is also fairly disk-hungry: you need either a local or a network drive with at least 2GB of disk space. --Carnildo 23:32, 31 October 2007 (UTC)

I think the first thing you should think of is how you want to use your computer. Does anyone have some ideas on how and where this computer can be used? --Kushalt 05:46, 1 November 2007 (UTC)

Check if any local museums want an actual working antique... other than that, I am pretty sure the Slackware mentioned above will work, I have something that is more legacy then that running it to just mess around with. Dureo 09:24, 1 November 2007 (UTC)
I used to know someone who turned old computers into firewalls for his personal network. x42bn6 Talk Mess 14:32, 1 November 2007 (UTC)
Generally, any task that needs a computer, but not a fast/new one. I once made a MAJOR improvement to a government LAN's performance by using several old Zenith Z-248's (Intel 80286-based PCs) as print servers. This offloaded the buffer (fast print file coming in from workstations, slow print file going out to printer) and half of the network traffic (I put the P-Servers at the printers, instead of in the server room) from the servers and network backbone. -SandyJax 16:37, 2 November 2007 (UTC)

Help with TLDs?[edit]

I have a website which i wish to register a domain name for. The particular domain name in question is already registered, but is not being actually used, ie "parked". (???) I tried to contact the owners of the domain by the email provided (it does say its for sale) and ive gotten no response... this was probably a year ago.

Is there any way that the domain can be opened up on the basis that the owner seems to have no intent to actually put the domain to use? I want to be clear and say that i dont own any sort of trademark that is being violated, but it makes sense to me that if a domain is registered it should be put in use if there is any demand by others to have the name... Ive looked around at ICANN for a few mins but havent found anything yet.

Can anyone help? :)

137.81.112.140 13:59, 31 October 2007 (UTC)

When does the domain expire? --— Gadget850 (Ed) talk - 14:15, 31 October 2007 (UTC)

"Expires on: 26-Jan-08"

however, of course, this doesnt mean that the contract will end then, its only the end of the term. Ive watched it expire last februrary and they obviously renewed. 137.81.112.140 15:09, 31 October 2007 (UTC)

Yes, they can renew any time before it expires. I don't see much recourse if they won't talk to you and they keep renewing it. --— Gadget850 (Ed) talk - 15:41, 31 October 2007 (UTC)
If your email doesn't get a response, try to send snail mail to the address in the WHOIS record. Also consider registering under other country codes/TLDs. -- Diletante 16:18, 31 October 2007 (UTC)

I realize that i can register another.... possibly similar TLD... thing is the one thats taken is the actual name of the organization i lead, so it would be nice to have the same name. Elsewise i would have to add something to the name that makes it different, but not harder to remember. :o hum......

137.81.112.140 21:13, 31 October 2007 (UTC)

If it's a country-level TLD, they may have rules against squatting, or may require proof of residence in the country; either of which could potentially be used to oppose their ownership. If your organization has legal rights to the name, you may want to go the legal route and ICANN's Uniform Domain-Name Dispute-Resolution Policy. There is no other mechanism to have their ownership removed, besides waiting for it to expire as mentioned above. 130.179.28.118 20:23, 1 November 2007 (UTC)

can any one help ? with MB to[edit]

It takes 1 minute to download a 7.5 MB data file. At what bit rate (to 2 significant figures) is the data being transferred? Weare2good 14:56, 31 October 2007 (UTC)

This looks entirely like a homework question, so I'll stick to general points. Your known conditions use bytes and minutes. Your answer needs to use bits and seconds. If your teacher hasn't specified the system, you might want to get nitpicky and point out that "byte" is system-specific (though by all means, use the most common bit/byte conversion value when you solve the problem). Don't forget about your SI prefixes, and then remember to keep your answer to the proper number of significant digits. — Lomn 15:39, 31 October 2007 (UTC)
Remember the distinction between real SI prefixes and computer industry pseudo-SI prefixes. In the past, it was sometimes uncertain which was meant, but most of the time it was the computer industry prefixes. Real SI prefixes go in powers of 1000, computer industry pseudo-SI prefixes go in powers of 1024. The recommended way nowadays is to use real SI prefixes when you really mean powers of 1000 like in everywhere else, and use a new set of prefixes when you mean powers of 1024. These are kibi- (Ki), mebi- (Mi), gibi- (Gi), tebi- (Ti) and so on. GNU/Linux seems to have caught on with them very well but unfortunately MS Windows, which 99.999999% of the computing world uses, ignores them. Data transfer speeds are commonly measured in real SI prefixes but storage sizes on disk are measured in computer industry pseudo-SI prefixes. JIP | Talk 18:42, 31 October 2007 (UTC)
60Mbit/minute.--Dacium 04:11, 1 November 2007 (UTC)

UK Mobile Automated Texting Services on the cheap[edit]

Can anyone help me with teh following. I am the editor of a charitable magazine and we wish to set up an texting service whereby people can send short messages to be printed. At the moment it's simply a mobile phone with no credit, it picks everything up OK but that's about it.

What sort of software / hardware would be needed to send an automated text back saying something like 'thank you for submitting your text, your message will be printed in next month's issue'.

I have a feeling this is one of those simple ideas that is going to be way outside our non existent budget! I've seen lots of free texting websites and wonder if there is anything free or for a small subscription that may be useful to us. Thanks —Preceding unsigned comment added by 88.144.30.22 (talkcontribs) 11:30, 31 October 2007

Example 3D art gallery?[edit]

Polyhedron with 12 24 vertices, none of which can be seen from the center.
Same polyhedron with the blue faces removed, showing the construction better.
View from the inside towards one of the corner; observe lack of visible vertices.
180° panoramic view of the inside.

Art gallery problem says that there exist simple polyhedra with interior points to which every line from a vertex crosses a facet. Anyone have an example of this (to add to the article)? I have thoughts about a "room" with trapezoidal side walls that manage to obscure the corners of the ceiling and floor and yet have their own corners concealed by the ceiling and floor, but I can't seem to make it work out. --Tardis 17:48, 31 October 2007 (UTC)

I'm skeptical that this is true. We don't have a reference for this supposed 'fact' - and it seems really unlikely that it's true. Let me think about it for a while - I bet I can disprove it. SteveBaker 18:38, 31 October 2007 (UTC)
It does seem absurd. Perhaps "vertex" means "convex vertex"? —Tamfang 19:44, 31 October 2007 (UTC)

Actually, I think I can see how to construct such a polyhedron. It will be rather ugly, but I see no reason why it wouldn't be possible. The construction I have in mind relies on two facts:

  1. a long prism with its ends hidden out of sight has no visible vertices, and
  2. if you cut the tip off a blunt pyramid and look in through the resulting hole, from most vantage points you won't see any interior vertices.

So, start with a fairly round convex polyhedron — a dodecahedron or an icosahedron should do — and cut off all the corners. You now have a polyhedron with small holes, surrounded by a few closely spaced vertices, where each original vertex used to be. Now plug the holes by gluing truncated pyramids, as described above, to the holes; if you do this right, this step will not increase the number of visible vertices. You now have a polyhedron with "bags" attached to each corner. (The bags don't have to be pyramidal, you just have to shape them so that no internal vertex is visible through the hole connecting them to the main polyhedron.)

Finally, add some long, thin prisms, each with one end hidden inside one "bag" and the other in a different one, arranged such that they hide all the vertices around the bag mouths from view when seen from the center of the main polyhedron. Note that you don't have to worry about the prisms intersecting: where that would otherwise happen, you can always add a slight bend to one prism in such a way that it's hidden behind the other. Since the ends of the prisms are hidden from view inside the bags, you're also free to arrange for them to be connected to the surface of the polyhedron any way you like.

Actually, while writing this description I managed to come up with a prettier and more elegant construction based on slightly tweaking the the edges and vertices of a tetrahedron (or any polyhedron, really), but I haven't come up with a clear way to describe that yet... —Ilmari Karonen (talk) 22:55, 31 October 2007 (UTC)

You can't "bend" your prisms without creating new vertices that you can see - polyhedra have to have planar surfaces and straight edges. Tempting though it seems: When you try to remove a vertex by cutting it's corner off and fitting a 'bag' over the end, you create at least three new vertices around the edges of the hole that are still very visible. SteveBaker 12:36, 1 November 2007 (UTC)
Regarding the prisms, yes you can. Just bend one of the prisms away from the viewpoint: the vertices that have to be added to produce the bend will be hidden behind the other, unbent prism. As for the bags, as you say and as I wrote above, adding them doesn't directly reduce the number of visible vertices (in fact it multiplies it by a factor of 3 or more); what the bags do is create hidden niches in which the ends of the prisms can be hidden in the next step. —Ilmari Karonen (talk) 12:55, 1 November 2007 (UTC)
Okay, here's a more elegant construction. Start with a truncated tetrahedron, with four small triangular faces and four large hexagonal ones. Split each hexagonal face into four triangles — one equilateral one and three slimmer ones — by adding a new edge between every second vertex; do this consistently so that five faces now meet at every vertex. Now take each of the original triangular faces and rotate it 90 degrees around its center normal, dragging the attached edges and faces with it (we can do this since all the adjacent faces are now also triangles). There are two possible directions to rotate the triangles; you want the one that makes the adjacent "slim" trangles flip over, making the polyhedron non-convex. If you did this right, and if I'm not mistaken, you should now have a simple, non-convex, chiral polyhedron with 20 triangular faces and 12 vertices, none of which can be seen from its center. —Ilmari Karonen (talk) 23:46, 31 October 2007 (UTC)
You're mistaken. That doesn't work. When you rotate the triangles 90 degrees - you continuously move the parts of the former hexagons out of the way so you can still see the vertices from the center of the object. SteveBaker 12:36, 1 November 2007 (UTC)
You're right, I realized that this construction won't work. Mind you, it's not for the reason you give; the problem is that, while the "twist" works just fine to hide the corners of one triangle, trying to apply it to all four simultaneously makes the surface self-intersect at exactly the moment when the vertices vanish from view. (Also, I made a mistake in the description above; the rotation should be by 120-90 = 30 degrees, or slightly less, not 90 degrees as I wrote.) The earlier "bag-and-prism" construction still seems perfectly feasible to me. —Ilmari Karonen (talk) 12:55, 1 November 2007 (UTC)

Okay, I finally got around to modeling and rendering a simple polyhedron satisfying the requirements, shown on the right. POV-Ray source code is available on the image description page. Feel free to use it in the article if you like. —Ilmari Karonen (talk) 04:18, 6 November 2007 (UTC)

I like, I like. Thanks for solving this. I would suggest that it be included (although it should be described properly as having 24 vertices). --Tardis 16:04, 6 November 2007 (UTC)
Damn, apparently I've forgotten how to multiply small integers. :/ 3 × 8 = 24, as you say. Thanks for the notice. —Ilmari Karonen (talk) 21:30, 6 November 2007 (UTC)

Low-level programming[edit]

How would one go about making a program which is run directly on the processor, instead of in an operating system? And as an aside, what operating system did Linus Torvalds use to write his first experimentary kernel? JIP | Talk 17:54, 31 October 2007 (UTC)

On a regular PC? It's easy to run a program directly on the processor if you can find a way to get it started. The easiest way to do that is to put it on the boot sector of a disk drive and boot from it. If your PC will boot from a memory stick - then that's the way to go. In essence, your program IS the operating system in that case. You could probably pursuade a standard boot loader (such as GRUB) to boot your program from a file on disk if you want to make life REALLY easy. Of course if you want to read files, display stuff, use the keyboard, mouse or sound card - then your program is going to need some device drivers - just like a real operating system.
On a 'dedicated' or 'embedded' processor (something like the BASIC Stamp board), you can load your program into PROM or flash memory and boot it from there. I've been playing around with a kit called "Games'n'Music" that lets you run stand-alone programs on a little Nintendo DS games machine. It's rather fun to work so closely to the hardware.
As for Linux - I don't have Linus's biography handy right now - so this is from memory: He didn't start off with the intention to write an operating system. He just wanted a little program that he could boot from that would turn his PC into a dumb terminal so he could log in to a remote computer system via modem. That program included keyboard and screen drivers, then interrupts, then multitasking...it grew more and more features until it became able to spawn other processes - and became an operating system. He initially used a Minix environment to get himself up and running - so early Linux used Minix file system structures. That part was actually fairly easy - there were lots of similar things in existance at the time. I had even written a small disk operating system (called 'EDDA') for my ancient TRS-80 which had a hand-built (and thus unsupported) floppy disk interface. EDDA was burned into a UV-erasable PROM chip. That was back in the early 1980's - but back then there was no Internet so EDDA never got shared and never 'took off' in the way that Linux did. Minix was already a fully-fledged, working, UNIX-like PC operating system that you could pick up for free - it would have destroyed Linux if only it's author had been prepared to license it under the GPL (or something similar) and let the world hack on it. I was running Minix on my PC (and on my Atari ST)when Linus announced Linux - so I picked up a very early version to play with. If it hadn't been for the ability to improve it and play with it, I'd have stuck with Minix. SteveBaker 18:18, 31 October 2007 (UTC)
And not to get too involved in the Linux question but Torvald wrote "just" the kernel; much of the rest was developed by the GNU Project. There are some that have rather hard feelings about Torvald getting so much credit for developing one component (albeit an important one) while the rest of the people seem to just fade into the background. I don't know enough about it to have much of an opinion, personally. --24.147.86.187 02:41, 1 November 2007 (UTC)
You are absolutely right about that. Without the GNU code, Linus would have had nothing to run under his new Kernel. However, the GNU team had been working for years (and are still working) on their kernel (Hurd) and never did get it to take off. Linus's contribution is not the code he wrote - but the way he runs the project. His 'benevolent dictatorship' model works - the GNU way did not. GNU's approach works well for creating and maintaining a bazillion little command-line tools - but for a bloody great monolithic kernel, it fails. The Hurd kernel tries to solve that management issue by changing the architecture to make the kernel itself be a bunch of smaller modules - but it's not working out...just as it didn't work out with Minix either. So forget Linus's code contributions - praise him for his managerial skills! SteveBaker 21:14, 1 November 2007 (UTC)
Yah, linus's main contribution is his leadership, which is natural for him to have due to his extensive code contributions. The microkernel is far superior though; we just need a bit more OS complexity before the macrokernel gets too bloated and developers are forced to use the more extensible model --ffroth 05:50, 4 November 2007 (UTC)


  • I think the dead-easiest way to do it would be to write a driver for Linux, insert it, turn off interrupts, and then you basically own the processor. Something like this should do it:
#include <linux/init.h>
#include <linux/module.h>
#include <linux/kernel.h>
    
static int take_over(void)
{
    cli(); // turn off interrupts

    // at this point, you can do whatever you want with the system,
    // such as ... warm up the room!
    for (;;) ;
    
    return 0;
}
    
module_init(take_over); // arrange to call take_over() on insmod
That's the general idea, and much easier than twiddling with GRUB. --Sean 18:45, 31 October 2007 (UTC)
That's true - but it requires distributing an entire working Linux distro in order to have people run your program. There are all sorts of reasons why you might not want to do that in the kinds of specialised applications where running on the bare metal is useful. SteveBaker 22:39, 31 October 2007 (UTC)
  • Sure, but my impression is that the OP just wants to try being the master of his domain. --Sean 13:13, 1 November 2007 (UTC)

Photo timeline software[edit]

Hi all,

I'm looking for a good piece of software that will display my photos in a timeline (using the photo's own metadata). My dream software would feel something like the roll-over effect on the Mac Dock -- the photos would be small and get larger as you mouse-over them. Does anything like this exist? Thanks! — Sam 18:40, 31 October 2007 (UTC)

Doesn't Google's Picassa do this? —Preceding unsigned comment added by 88.144.30.22 (talk) 18:48, 31 October 2007 (UTC)

This Adobe Software has a timeline feature, I think. 69.95.50.15 20:46, 31 October 2007 (UTC)

Email fields[edit]

Is there any practical difference between the "To:" field and the "CC:" field? If I put your email address into either field, and send a message:
- you will receive a copy of the message
- you will receive a reply from anyone else who chooses "Reply to All"
- you will not receive a reply from anyone else who chooses "Reply"
- your email address will be visible to all recipients (unlike BCC)
Is the only difference, then, a social rather than technical one: to distinguish who the letter is really written "to"? jeffjon 19:54, 31 October 2007 (UTC)

Your third point is slightly off, or irrelevant. Since in your scenario "I" have sent the email, then of course "I" will be the only one who receives a reply from anyone who chooses "Reply". Of course "you" will not get a reply, since "you" are the one doing the replying. On the other hand, "you" might have the option to keep a copy of all sent mail enabled.
Other than that, I'd have to agree, there seems to be no practical difference. --LarryMac | Talk 20:55, 31 October 2007 (UTC)
They exist purely for business reasons. An email directly to you is for you attention and should require your action. A CC (carbon copy) is for your information, a BCC is a blind carbon copy so you are given this information without the knowledge of other people (say HR copied in for data gathering reasons). I guess some mail clients might allow alterations in the reply-to systems to make it reply only to those in the main list (default) and reply include CC-recipients also. I know in my work system we can only setup max distribution lists using individual names so often the CC field is used to get the email out to a lot of people at once. Basically because it's too easy to include more people than need it than it is to think about which audience need the mail (work rant over!) ny156uk 23:21, 31 October 2007 (UTC)

HTTPS vs man in the middle[edit]

Hello everyone, I just read recently the article Transport Layer Security, HTTPS and AES

I wondered what would happen if you happen to know part of the message if not the message itself that is being transmitted and you got the capability to be a man in the middle. And how this protocole defend against that.

Here is an example :

Bob the hacker connect to the mybank website where he got an account, and connect to his own account. That way he knows how much https request he must send before coming on the screen where the vital information permiting him to transfer monney reside, and his html code.

Now Bob go to the isp of Alice Themilionaire, shoot the technical administrator with a shotgun and place a tap on the line of Alice. Go home and light up his computer.

Alice connect to the mybank website, get on the log in screen, type her password, and get on that vital screen.

Can bob when he intercept the reply packet from the server which contain the reply from the server containing "the golden web page" try randomly to XOR part of the packet with the data : <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> Until he find that the xor output will be not a random output but a cyclic 256 bit number (the doctype is bigger than 256bit especially if the webpage is encoded in UTF-8) which happen to be the key that repeat itself, permiting him toi crack the rest of the packet and find his golden information.

Is it "possible" to do that way ? I got inspired by the way you can "crack" the Vigenère cipher if you happen to know part of the document, then do a guesswork on the rest with the information you have discovered on a long message. — Esurnir 20:17, 31 October 2007 (UTC)

  • Short answer: no. Alice and the bank's whole session is encrypted with a secret key. The HTTP headers are even more predictable than anything in the web page, but unless Bob has the key, he's going to have to try zillions of combinations to see if any match those known plaintexts. Since he's already shown he's not above shooting people, he should point that shotgun at Alice and demand her password instead. --Sean 21:01, 31 October 2007 (UTC)
I don't understand, I thought that the public key cipher is used for authenticating and exchanging the shared secret, and that the rest of the communication is done with a 256 bit AES (let's take that one who is reputed as "secure") cipher. Am I wrong ? — Esurnir 21:59, 31 October 2007 (UTC)
That's right (except for the 256 bit part, usually it's just 128). And you would right about the known text attack of just XORing the http headers out to find the key... but Rjindael is not just a simple xor, it's a very complex algorithm that's immune to the known text approach --ffroth 22:09, 31 October 2007 (UTC)
Hmmm... Let's say that I got at least the certainty that some byte of the encrypted packet is a known text (the http header for example). If I choose a 16 byte part, shift row, shift collumn, and xor it with the encrypted text at the exact place where I know my text is. Wouldn't that create me the key ? Or where would that operation "fail" ?
From what I understand in both case, I take a number A (my message), when I encrypt it, I get A, do some stuff on it to get stuff(A), then make stuff(A) XOR K (the key) = E (encrypted message). If I do the exact same stuff(A) operation on my known text, then make stuff(A) XOR E wouldn't that produce the key ? Or where does that logic fail ? Doing the operation stuff(A) is supposed to be the same operation, doesn't it ?
P.S. I assume that trying to find the decipher using the known text but not knowing the exact position of is much more difficult.—Esurnir 23:08, 31 October 2007 (UTC)
What you're describing is a known-plaintext attack. The cipher used for encrypted web communication is not a simple XOR, but is far more complicated, so that the only known way to conduct a known-plaintext attack is to try each and every one of the 340,282,366,920,938,463,463,374,607,431,768,211,456 possible keys until you find one that works. There is no other known way to recover the key given the message and the encrypted message. --Carnildo 23:42, 31 October 2007 (UTC)
Oh shame on me, always read the litle print. I just discovered the Rijndael key schedule which is the part where the egg get scrambled.−Esurnir 00:08, 1 November 2007 (UTC)
That's a little O_O-inducing isn't it? :) --ffroth 18:01, 1 November 2007 (UTC)

Installing mac progams while on a PC[edit]

I am trying to install some Mac programs, for use in my school (e.g.openoffice) on my flash drive, but since I'm a PC user at home, I cant open the .dmg files. The Macs at school will not open dmg files due to admin blocking. How can I install these programs? --Omnipotence407 22:01, 31 October 2007 (UTC)

The admins are blocking that for a reason - they (probably) don't want unauthorised software installed on those machines because of any number of real or percieved risks. You should ask them to install it for you. SteveBaker 22:29, 31 October 2007 (UTC)

I am not trying to install them to the computers, I want to install them to my flash drive. I want to do this in order to show my Impress Presentations w/o having to convert them to Powerpoint. --Omnipotence407 22:37, 31 October 2007 (UTC)

Well the tough thing is that DMGs are not just like ZIP files, they are disk images. So you can convert them into PC-readable disk images (ISO, using DMG2ISO) and then open it with a program that can read ISOs on the PC (blah), but that probably isn't going to do the trick. If it I were me I'd try to find a Mac that didn't have the restrictions on them and then use that to open the DMG and copy the file to the drive (I'd be worried that if you did it on a PC the Mac OS would not recognize the copied file as an executable application). I'm not sure if Impress will run just off of a flash drive (usually these programs need to install things in the User/Library/Application Settings folder) but that might work. --24.147.86.187 02:36, 1 November 2007 (UTC)

Audacity as a dB meter[edit]

What is the best way to use Audacity as a dB meter? Basically, I am looking for a way to measure the maximum dB level. Thanks --Omnipotence407 22:23, 31 October 2007 (UTC)