Wikipedia:Reference desk/Archives/Computing/2008 October 21

= October 21 =

Booting Windows XP off of ext
I've read about these drivers that enable Windows to read ext partitions, but I was wondering, woudl it be possible to boot Windows off of an ext partition? If so, how would I go about doing this? Thanks.

63.245.144.77 (talk) 04:13, 21 October 2008 (UTC)
 * I am almost sure it is not possible, at least for now... It is not even possible to boot systems like Ubuntu directly from a NTFS partition - and this is a system that can be "hacked", since it's open source, unlike Windows XP... SF007 (talk) 19:40, 26 October 2008 (UTC)

Help finding Adobe software
I was asked to look at a software package and see if it would fit a given application. I know it's made by Adobe, it is basically for designing web sites and it possibly works in conjunction with Photoshop. I believe the cost is somewhere around US$100. I've been trying to trawl around Adobe's site but the combination of heavy graphics and the ancient computers here at work is making the task rather daunting. So, does anyone know what I'm thinking of? I'd prefer a direct link to the Adobe page that explains the software. Thanks, Dismas |(talk) 05:12, 21 October 2008 (UTC)
 * Probably Dreamweaver, although it costs $399: http://www.adobe.com/products/dreamweaver/. There are other Adobe web-design tools, but they are devoted to specialized technologies like Flash and web slices, and none of them cost $100: http://www.adobe.com/products/. --Account created to post on Reference Desk (talk) 08:26, 21 October 2008 (UTC)


 * Maybe Contribute? It is often used to maintain sites built with Dreamweaver. http://www.adobe.com/products/contribute/ MaxVT (talk) 15:29, 21 October 2008 (UTC)
 * My mistake. It is not made by Adobe.  It's SiteGrinder.  Dismas |(talk) 09:55, 22 October 2008 (UTC)

.rar file
how can I open a .rar file type?Shraktu (talk) 05:13, 21 October 2008 (UTC)
 * What OS are you using? Have you read over the .rar article?  It goes over the software needed.  Dismas |(talk) 05:54, 21 October 2008 (UTC)

You need to download a free copy of winrar, (http://www.rarlab.com) install it and use it to read/extract all .rar archive files. Additionally, winrar will allow you to also read various other kinds of archive types such as .zip, .tar, .tar.bz2, .tar.gz. etc.

IE Keeps crashing/ Hanging when loading into Wikipedia
Hi i have a group of users who frequently access to wikipedia site. However lately they have been encountering this problem:

IE hangs when trying to load www.wikipedia.org. No responses from IE. I would need to terminate the session and restart again. Sometimes after shutting down and restart it works, but some other times it will just not response.

Is there any solution for this? or advise you could give relating to this problem? thanks —Preceding unsigned comment added by 82.118.67.5 (talk) 11:03, 21 October 2008 (UTC)


 * Have you tried loading the page in Firefox or another browser? Are there any other pages that cause it to hang? Does it crash when loading http://en.wikipedia.org or just the language portal http://www.wikipedia.org? Can you kill it with Control-Alt-Delete or does it freeze the whole system (entirely possible on Windows)? « Aaron Rotenberg « Talk « 15:35, 21 October 2008 (UTC)


 * I increased the number of simultaneous connections IE makes to the server, and the problem went away for me. http://support.microsoft.com/kb/282402 --bjm

8800 GS not working in Bootcamp (iMac).
A friend of mine just bought an iMac, but upon installing Vista 32 bit, his 8800 GS is identified as a Standard VGA Adapter in dxdiag. Nvidia drivers fail to install (don't recognize the card), and I've looked through various websites claiming to have solved this, but their solutions haven't worked. Could this have been an installation error, or maybe hardware? · AndonicO  Engage. 11:14, 21 October 2008 (UTC)
 * I take it he installed the Boot Camp drivers? What if you try installing an official nVidia driver instead with a modded .inf file from ? Past those idea, not a clue. -- Consumed Crustacean (talk) 00:59, 22 October 2008 (UTC)

Excel 2007
Hi, my problem is in Excel 2007 using Windows XP SP3.

I have a column full of cells with dates in them. I want the word "DUE" or similar to appear in another cell when today's date is in the week leading up to the due date, inclusive. If today's date is before the aforementioned range I want a blank cell, and I want "OVERDUE" if today's date is over that range. That is:

Cell A = Due Date

Cell B = Notification message

TODAY is less than A-7, then B = ""

TODAY is in [A-7, A], then B = "DUE"

TODAY is > A, then B = "OVERDUE"

I would like to see that as a valid cell formula. Any suggestions? —Preceding unsigned comment added by 81.187.252.174 (talk) 13:34, 21 October 2008 (UTC)


 * Use the IF function - look it up in on-line help to see how it works. IF on its own produces one of two possible outputs, but if you nest IF functions you can extend the number of possible outputs. As your problem has three possible outputs, you will need two nested IF functions, one inside the other; see here for a tutorial. Gandalf61 (talk)


 * That's fine, I can nest the IF - but I need to know how to see if TODAY is within [A-7, A]. —Preceding unsigned comment added by 81.187.252.174 (talk) 14:04, 21 October 2008 (UTC)


 * Make that your final "else". If we label your conditions in the order you presented as X, Y, and Z, then you'd have something like...
 * . -- LarryMac  | Talk  14:38, 21 October 2008 (UTC)

Easiest solution...Have the rseult of =today in a cell of its own. Then test against that cell... (assuming today is in A1, due date is B1)

=IF(B1>A1,"Overdue",IF(B1<=(A1-7),"","Due"))

Make A1 a fixed reference ($A$1) and you can copy down provided you have setup the table correctly. 194.221.133.226 (talk) 14:54, 21 October 2008 (UTC)


 * You've got the signs backwards. Make a list of dates in column B and put this into cell C1:


 * =IF(B1<=TODAY,"Overdue",IF(B1<=(TODAY+7),"Due",""))


 * Then drag or double-click to get the formula to fell the column. Days before Oct 24 are overdue; days between Oct 24 and Oct 31 and due. Dates from November are blank. Matt Deres (talk) 17:42, 24 October 2008 (UTC)

Opinions on keyboard w/o separate number pads?
I use a notebook computer. I have a stand into which I can insert the notebook, raising the screen to a more ergonomic level, but that means I need a separate keyboard (the notebook's is covered by the stand). I've gotten used to the notebook's key layout -- specifically, no separate numeric keypad off on the right, as with most desktop machines. I'm thinking about buying a separate mini keyboard -- full-size keys, but no keypad, like this one (used as an example only). Trackball or touchpad is not important; I'm too accustomed to a mouse. (That's another reason for not wanting the standard keyboard: it's too far from the typing keys, across the numeric keypad, to the mouse.) I'd appreciate any suggestions, especially those based on experience. Thanks. --- OtherDave (talk) 14:17, 21 October 2008 (UTC)


 * How much time do you spend inputting financial/numerical based information? If that answer is anything apart from "very little" you will feel the benefit from having a dedicated numerical keypad. There'll be a 'lock' key to put the 'numerical' pad on on the laptop and from experience I can say you can get used to it, but i've never been able to get anywhere near the speed and accuracy I can manage with a dedicated numerical pad. By the by don't be surprised if you're told off, the reference desk isn't a place for 'debate' (though plenty goes on) 194.221.133.226 (talk) 15:04, 21 October 2008 (UTC)


 * I appreciate the opinion. I rarely enter numerical data beyond updating my own information in Quicken.  The point is that when the laptop's in the holder, I can't use its keyboard at all (See [here] for an example.)  I'm currently using a wireless keyboard I had, but if I could saw six and a half inches from the right site -- space for cursor-control key clusters and the numeric keypad, it would look a fair amount like the example in my original question. --- OtherDave (talk) 17:24, 21 October 2008 (UTC)


 * A Google search on compact keyboards yields many sources. Many of those sources also have stand-alone numeric keypads that you can use with the main keyboard when needed and move out of the way when not needed.  -- Tcncv (talk) 00:50, 22 October 2008 (UTC)


 * I've been using a super-sexxy ergonomic keyboard (without number pad) for many years - and because I had the same worries as you, I bought a separate USB number pad. They are pretty cheap and easy to find on the Internet (eg on Amazon.com - for $10.99).  However, I have to say that since the day I bought it, I've never once found the need to use it.  Since so many people work on numberpad-less laptops these days, pretty much all modern software can cope well without the numpad. SteveBaker (talk) 00:56, 22 October 2008 (UTC)


 * Can't you get a regular keyboard and just ignore the numeric keypad? Alternatively, why not dump the stand and get an external screen to go with your external keyboard.  It'll be just like a desktop PC then.  Astronaut (talk) 20:56, 22 October 2008 (UTC)

Different web addresses with same target
Hello all. Sorry for a potentially vague question. What I was wondering about is when two different web addresses point to the same webpage, differing only by the addition of a number in the URL. For example: "www.law.cornell.edu" and "www4.law.cornell.edu" both go to the same website. What is the reason for the presence of a "4" in the URL? Thanks! --67.159.86.84 (talk) 15:49, 21 October 2008 (UTC)


 * The "www4" probably means that they've got several different web servers providing their website, and are doing load-balancing so that no one server gets overloaded. The "4" in the URL simply means that the page you're viewing is being provided by web server #4. --Carnildo (talk) 20:56, 21 October 2008 (UTC)


 * It's worth noting as well that technically you shouldn't have to do it so crudely—there's ways of making it all look identical without much effort. Alas, academic IT is pretty poor. ;-) --98.217.8.46 (talk) 21:37, 21 October 2008 (UTC)


 * Thanks, everyone!--67.159.87.120 (talk) 14:01, 22 October 2008 (UTC)

Floppy drive necessary these days?
Do I really need a floppy drive in my Dell computer running windows XP in case anything goes wrong and I have to boot from floppy. I would like to replace it with another optical drive. Is this easy/possible?--GreenSpigot (talk) 17:21, 21 October 2008 (UTC)
 * No, you don't need one, but it makes things easier. Put the floppy disk into an existing floppy drive and copy the files onto a CD, then you can get rid of the floppy drive. Den dodge  Talk Contribs 17:26, 21 October 2008 (UTC)


 * (ec) If your PC's BIOS permits booting from a CD drive, then you can do without a floppy drive. Unfortunately, most floppy drives are only 3.5" wide while optical drives are 5.25" wide.  Take a look inside your PC to see if the 3.5" floppy drive is fitted into a 5.25" bay using some removable brackets.  You will also need to consider the exterior plastic case moulding - sometimes the floppy has a removable blanking panel next to it, sometimes not (in which case you could take a sharp knife to it).  Of course, you might already have a free 5.25" bay, in which case just install your new optical drive in there and leave the floppy drive alone.
 * Note that you won't be able to use the floppy's data cable to connect to an optical drive, but there could be a spare connector on an existing optical or hard drive that you could use once you have set the drive jumpers correctly. Otherwise you can get a new cable from your local PC components store.  Astronaut (talk) 17:38, 21 October 2008 (UTC)


 * Very useful info! But I remember with windows 98 the optical drive drivers were not loaded after a H/W change and you had to reload them from floppy before you could get the optical drive to work. Is it different with XP?--GreenSpigot (talk) 17:47, 21 October 2008 (UTC)


 * Actually, it is newer BIOS-es that can read (and boot from) a CD without waiting for the OS to load drivers. -- LarryMac  | Talk  17:58, 21 October 2008 (UTC)


 * "Newer", in this case, meaning "almost anything built in the past ten years". --Carnildo (talk) 20:57, 21 October 2008 (UTC)

OK thanks to everyone who replied. It would seem that I can get rid of my FDD and boot from existing optical if necessary. But how can I fit another optical drive if I already have one optical drive and 2 HDDs connected internally? Can I have one optical drive as master and the other as slave connected to IDE1?--GreenSpigot (talk) 16:50, 22 October 2008 (UTC)


 * Both IDE0 and IDE1 support connecting two devices. This works for optical drives in the same way as for HDDs. MTM (talk) 21:03, 22 October 2008 (UTC)

Neo Office Help
I am in the middle of a project and suddenly my document is read-only. How do I switch this off? I have tried Command-I and changed the permissions there, but to no avail. I think it must be something in Neo Office. Can anyone help?--ChokinBako (talk) 18:10, 21 October 2008 (UTC)


 * As an emergency panic button, try saving the document in your local machine with another name and continue editing. Then, later, when you can give us more details, we will investigate the matter further. HTH, Kushal (talk) 18:36, 21 October 2008 (UTC)

Collision detection
I'm writing a spacewar clone, and I'm currently doing the collision detection code. All my game objects are represented by a collection of lines, and if any one of those lines cross with another object's lines, a collision has happened. I know about using circles to find likely collisions first, etc, for efficiency, but I'm having some trouble with the exact, final test.

My lines have both velocity and angular velocity, and only the endpoints are known. So for some points P_1 and P_2 on the two lines, with vector velocities V_1 and V_2 and angular velocities θ_1 and θ_2 about centres C_1 and C_2, a collision at time t implies (oh, and let R(θ) = the rotation matrix through θ)

R(tθ_1)(P_1 - C_1) + C_1 + V_1t = R(tθ_2)(P_2 - C_2) + C_2 + V_2t

Now this is a hairy equation, because R(θ) involves trigonometric functions with an unknown (t) on the inside. P_1 and P_2 are also not known, though they are bounded (like t).

The time slices are small (1/30th of a second per frame is what I'm aiming for), so the small angle approximations could be used, linearising it (which would be easy to solve, comparatively at least). But first I want to try and do it exactly, and see if it's possible, before I fall back to that. 79.72.153.17 (talk) 19:45, 21 October 2008 (UTC)

Oh, I should also add, because it's lines intersecting, it suffices to check if any of the endpoints can intersect with the other line, so the equation will need to be solved 4 times for different values (or not) of P_1 and P_2. 79.72.153.17 (talk) 20:36, 21 October 2008 (UTC)


 * If you really do mean line (as in straight line) then can't you a) generate the "slope intercept" forms for both lines b) from that calculate where they join c) back-substitute that location into the time-parametric forms of the line, which gives you the time each particle arrives at the intersection. If the difference between their t is relatively small, they're there at much the same time, so you have a candidate collision. Now you want to do pixel-correct collision detection. If the objects are sprites then a collision occurs if the bitwise-AND of their shifted collision masks is nonzero. If the objects are 2d vectors then you can (almost) do collision detect by seeing if any of the line-segments that make up object A intersect with any of the line-segments that make up object B ; the only flaw to this is if one object is large enough to entirely contain the other, in which case it could (depending on speeds) "skip" inside without colliding (but fixing that is easy enough to hack).


 * However, you're talking about having more angles in there than I really know what for, so maybe you really mean the objects aren't moving along straight lines but instead along curves. If the curves are predictable then it should (I guess, my math isn't good enough at this point) be possible to predict their intersection(s) as above. -- Finlay McWalter | Talk 23:34, 22 October 2008 (UTC)


 * If you really want to do this exactly (and I doubt very much that anyone would really need to - aside from the intellectual pride of doing it "right") then you should start off by transforming the coordinate system so that one line stays at (say) X==0...the Y axis and the other line moves relative to it. Also, don't use trig and angles - use matrices - that means that everything is in nice linear math.  But nobody EVER goes to that much trouble for collision detection.  What I would do (and I'm explaining this kinda roughly) is to calculate the position at the end of the current frame for all of the lines - and just do a static collision test between their end-of-frame positions.  HOWEVER, if the motion of either of the objects during a single frame is more than about half the size of the object - then I'll split the time into two half-frames and test those separately.  If the motion during either half-frame is more than the size of the object - I'll split again into quarter frames...and so on.


 * In effect, we are breaking the motion into two halves each time that the speed that things are moving is fast enough that we might 'miss' a collision because one object might jump over the other during the frame.  Since most objects are moving less than their size in each frame - we only rarely have to split the motion up - so this extra testing only very rarely comes into play and it hardly affects the time it takes to run the testing at all - the simplification of only having to do a static test (no trig - no differential equations!) more than pays for the few extra tests you actually have to do.   Another observation is that the faster things are moving - the less people will notice if your collision detection errs on the side of collisions when there should strictly have been a miss...especially if the collision can be covered with a nice juicy explosion!


 * If I find a collision - and I need to know to more precision when during the frame and where in space the collision actually happened - then, again, I'll do a 'binary search' through the duration of the frame to find where the collision must have happened to some arbitary temporal precision that I can specify.


 * I would also observe that it's better to test convex polygons (or perhaps triangles) than lines. This makes it impossible (for example) for a small object to wind up entirely inside a large object because you 'missed' a collision with the infinitely thin skin of the larger object.


 * Testing circles instead of lines or polygons is even better - and it's worth looking to see if you can approximate the shape of your objects with a few dozen circles and be 'good enough' because circle-versus-circle testing is insanely fast compared to line-versus-line (if the distance between the centers is less than the sum of their radii...KABOOM!)...also circles are rotationally invarient - so you only need to transform the center of the circle - not two ends of a line. This means that you can replace a 40 line vector description of an object with maybe 80 to 100 circles and STILL run faster.  Most shapes that have 40 lines look pretty good when looked at as 100 circles...and since this is only for collision purposes - it doesn't have to be that accurate.  This works even better in 3D where sphere-on-sphere tests are hundreds or even thousands of times cheaper than polyhedron on polyhedron tests.


 * SteveBaker (talk) 04:13, 24 October 2008 (UTC)


 * Ooh, some responses. Finlay McWalter: yes, I mean literally straight lines (or, more specifically, line segments) colliding. The difficulty is these lines are rotating as time goes on. If there was no rotation, this would be a much much simpler problem to solve. (The curve formed would be a Cycloid, or some combination of several cycloids). Unfortunately, being able to name the curve is about as far as my knowledge goes in that direction, although it'll be an interesting new topic for me to read about, like the harmonic addition of sines and cosines that I found the other day.


 * SteveBaker: a lot of this is pride, yes ;) I'm not expecting a usable exact solution to drop out. The problem with using matrices is that there are two independent centres of rotation. I could use affine transformations and their matrix representations, but those are trickier. I also can't see how to directly linearise it, because the unknown t is inside the trigonometric functions. There is also no guarantee of convexity or even polygonicity. Some of the shapes could be broken down into convex polygons (asteroids, etc), but most can't. An approximation using circles to find precise collision times probably also won't perform well. 79.78.52.62 (talk) 12:33, 24 October 2008 (UTC)


 * It seems like you would have to solve a transcendental equation which likely doesn't have an analytic solution (given sufficient pride, you could try to prove that!). Regarding problems with non-convexity: Try binary space partitioning. The trees can be pre-computed if the object's shape is always the same (although modern computers are likely fast enough to do it in real time for your game). Icek (talk) 11:29, 27 October 2008 (UTC)

design and logo,s
hi can you help i have a site which won t show design or logo,s i was told reinstall would correct this but hasn t. what would cause this in a new site and where can i check if reinstall was carried out. i,ve been banging my head of a wall for days trying to get answers. hope you can help thanks Scivver2008 (talk) 21:35, 21 October 2008 (UTC).


 * I'm sorry, but you haven't given us enough information to really make heads or tails of what you are asking about or what the technical problem is. Can you perhaps send us the URL, or describe in more detail what's not working? --98.217.8.46 (talk) 21:39, 21 October 2008 (UTC)

Dual pointers in windows
Hello. Is it possible to use multiple pointers in windows xp/vista? For example, using one usb tablet and one regular mouse. I want to use them in graphic, editing and 3d packages. I'm already using all the keyboard shortcuts but there are moments where it would be easier to have 2 pointers. I've looked quite a bit on the net but haven't found anything that works. Anyone uses two pointers efficiently and would recommend a method or software? Thank you. 190.220.104.35 (talk) 22:02, 21 October 2008 (UTC)


 * Currently, Multi-Pointer technology is not available for xp/vista that I am aware of. About two years ago I followed this technology being made but haven't seen it mass produced or even released for average users to mess with. it looks like there was an attempt at this being done for linux called MPX however I can't speak from experience in that. —Preceding unsigned comment added by Forai (talk • contribs) 22:07, 21 October 2008 (UTC)

communicating between two windows with Javascript
Let's say I have two windows open both pointing to pages on the same server. Can I communicate between them using Javascript? Basically I'd like it if one window could query the other to find out what information it currently had displayed. Again, it's on the same server (it is not cross-site scripting). How would I do this, if it is possible? I'm looking for a Javascript-only solution primarily but if a little PHP was involved I could live with that. The only thing I've come up with is having the page I want information from set a cookie each time it is changed that has the information in it. Good idea? Workable? --98.217.8.46 (talk) 22:24, 21 October 2008 (UTC)
 * Looks like the cookie method you describe is patented ;( -- Finlay McWalter | Talk 13:17, 22 October 2008 (UTC)
 * Wow, what a dumb patent. I mean, seriously, if that's the first thing that came to mind for me, a non-professional coder, how can that possibly be something unobvious to a practitioner in the art? Anyway, I'm not going to start respecting lame software patents for things I'm writing for my own personal use, in any case. --98.217.8.46 (talk) 13:54, 22 October 2008 (UTC)