Wikipedia:Reference desk/Archives/Computing/2011 August 10

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 10

Cannot access Windows Update

I need help. I need to access Windows Update, and I keep getting this error: "The website has encountered an error and can not display the requested page." I'm doing this on a computer that not only isn't mine, but is entirely in French, so I can't really navigate much due to my inability to read the language. Anyone who can help me out on this, please, I need your guidance. Windows XP Service Pack 2, if it helps. --70.29.252.46 (talk) 00:10, 10 August 2011 (UTC)

Well, the 1st update you should run is to install service pack 3, right? You can download the enire SP3 package in French here and install it manually. Vespine (talk) 04:25, 10 August 2011 (UTC)

Problem long solved, but thanks for the suggestions nonetheless. --70.29.252.46 (talk) 08:14, 10 August 2011 (UTC)

computers?

would it cost more to buy a computer or make one? what would be needed to make a computer? 70.241.16.91 (talk) 00:22, 10 August 2011 (UTC)

Commonly, computer hobbyists assemble a computer by purchasing pre-fabricated parts: a computer case, a power supply, a CPU and heatsink, a main-board, and a video card (for example). It is possible that this will save you money, but it depends on your skill as a shopper and your technical expertise. If you want to make a computer from more elemental parts, it will require far more expertise, and a lot more time and money. Nimur (talk) 00:34, 10 August 2011 (UTC)

There's heaps of websites that go into a lot of detail about what's involved with building your own PC. Yes it's typically cheaper but most people do it because it is fully customizable, you can put in only and exactly the components you want. I've never bought a pre-made computer, but if you've never even opened up a PC case and swapped a video card or upgraded a disk or anything, I would not recommend you start with building a whole computer, unless you have easy access to a friend who has built computers and is willing to help you. Things can go wrong and if you can't work out what it is, you'll be stuck with an expensive and large door stop. Vespine (talk) 01:49, 10 August 2011 (UTC)

I doubt if you'd save a great deal of money building a PC from parts, but it may well be worthwhile if you can reuse parts from old PCs. I managed to cannibalise two duff PCs to make a working one without too much difficulty - not exactly state of the art, but it cost me almost nothing, and enabled me to keep online long enough to afford a proper upgrade. As Vespine says, you'll need some backup from someone with a bit of experience to attempt a new build, so why not see if you can get hold of some obsolescent hardware first, to tinker with - too often, PCs get scrapped while they are still usable. AndyTheGrump (talk) 02:03, 10 August 2011 (UTC)

It depends on the type of computer you're looking for too. If you want a high performance "Gamer Machine" you may well find that you save a significant amount of money assembling it yourself, because the markup on those machines is rather high. But if you're looking for a "bare bones" cheapest computer you can buy, the price difference probably won't be significant. 76.28.67.181 (talk) 06:49, 10 August 2011 (UTC)

Also note that there are web sites that allow you to specify every component in the computer, and then they will assemble it for you. This provides the flexibility you might want without having to be an expert at computer assembly. Price-wise, though, it's in the same range as buying a computer off-the-shelf, although if you pick all premium components this approach will cost significantly more. StuRat (talk) 02:08, 10 August 2011 (UTC)

Another note not brought up above... If you are not in a rush and can plan well, you can save a lot of money by buying the parts you need only when they are on sale. I put together a computer that would normally run around $2,000 for $500 by purchasing each part when it was heavily discounted with large rebates. It took about 6 months. If you plan to do this, there has to be some order to your purchase. This is how I did it:

1. Buy a CPU. The CPU will limit what kind of motherboard you can use.
2. Buy a motherboard and CPU fan. The motherboard will limit the kind of memory you need, the power supply, and the case.
3. Buy memory, hard drive, and case at any time. Buy the video card (if you want a high-end one) next. That will put a power requirement on your power supply.
4. Buy the power supply.
5. At any time, pick out the monitor, keyboard, and mouse that you like.

Then, once you have all the parts, you are ready to go - except you don't have a Windows license. I don't use Windows, so I never factor that into my costs. -- kainaw™ 12:29, 10 August 2011 (UTC)

I see a couple problems with that strategy:

1) Your components may remain untested until after their warranty has expired, making it impossible to return any defective parts.

2) Technology improves and prices fall so quickly in this area that 6 month old components might cost more and do less than those purchased today. StuRat (talk) 14:37, 10 August 2011 (UTC)

Just because it irks me only a little, more then one person has now said buy a "cpu fan": If you buy a boxed cpu, you do not need a cpu fan, they come with a perfectly decent cpu fan. Today's CPUs are NOT the 100degC monsters of several years ago, pretty much all the stock coolers are perfectly fine.Vespine (talk) 22:59, 10 August 2011 (UTC)

Yes and no. Again, it rather depends on your PC. If you are building an overclocked gaming machine, you may well want a heatsink/fan combination the size of a halfbrick - if you look at the power consumption, even a standard CPU at stock clock speeds and voltages is producing enough heat for it to be worth looking at as a potential problem, so any extra (passive) cooling can't be a bad thing IMO. AndyTheGrump (talk) 03:05, 11 August 2011 (UTC)

Having built quite a few comptuers of various types including with aftermarket HSFs I agree with Vespine here. Any modern retail CPU comes with a HSF that is up to the job under normal settings and conditions. If the stock HSF isn't up to the job, the product is clearly defective and you should be entitled to return it in many countries. Obviously if you are overclocking or you find the HSF too noisy or for a variety of other reasons you may want to consider an aftermarket HSF. And similar if you buy a non retail CPU it probably won't come with a HSF. But for a retail CPU the stock HSF should be fine. Case cooling may need some consideration but that's a different matter. BTW, I would personally avoid any component other then perhaps HSFs and cases which come with only a 6 month warranty. Not saying I agree with the above strategy, in fact I would agree with StuRat's second point that most of the time it would be a mistake from a price standpoint (well for the HSF, keyboard/mouse, case and PSU it may be ok) but I've never lived in the US only Malaysia and New Zealand and we rarely have the sort of rebates that seem common there. Nil Einne (talk) 13:00, 11 August 2011 (UTC)

Yes and yes :P We ARE talking about a person here considering building their FIRST PC. They should not be thinking of building a pimped up overclocked gaming rig, and we should not be talking them into it. Once they become an ENTHUSIAST they will learn and consider aftermarket CPU cooling for themselves, it's a (completely) unnecessary complication and expense for a first build. As for noise, I've bought and build almost every intel CPU architecture since 486 days and since core 2 duo 4 or 5 yeas ago, the stock sinks have been very cool AND quiet. Vespine (talk) 23:13, 11 August 2011 (UTC)

o-p here, thanks for all the help! :) 70.241.16.91 (talk) 19:36, 12 August 2011 (UTC)

MS Office on a tablet

I should be very grateful if someone tells me can I open the MS OFFICE 2011 files in a tablet PC.I have many typing works and I want to know can I connect a bluetooth keyboard to a tablet? any proposal or answer is appreciated in advanc.--Chavosh (talk) 05:58, 10 August 2011 (UTC)

I've changed the title of the section to something more descriptive than simply "Question". As it says at the top of the page, things like Question, Query, or Problem are frowned upon. I'm sorry I don't have an answer for your question though. Dismas|^(talk) 06:07, 10 August 2011 (UTC)

It depends on the tablet. If the tablet runs MS Windows, then you can install and run MS Office. On an Apple iPad you can use applications such as Documents to Go and Apple's iWork[1] to view Office files, though they may not have the full functionality of MS Office. Various file viewers are also available for Android tablets. As for keyboard, the iPad supports bluetooth keyboards[2], and many other tablets will too, but you'll have to check for the specific tablet you have or are thinking of buying. --Colapeninsula (talk) 12:05, 10 August 2011 (UTC)

Hi-Grade ftp site for laptop drivers

I've been trying to help with somebody else's old laptop, which would be handy for using like a netbook. It works, but has no display adaptor installed, so it's stuck at an annoying resolution like 640x480. I assume it needs its drivers, which I assume are at the manufacturer's FTP site, linked to from this page: http://www.higrade.com/nqcontent.cfm?a_id=10956 (or the direct link is ftp://ftp.higrade.com/2010/Drivers/)... but the link never loads. I've tried just clicking it in Firefox, and I've also tried using FireFTP, and I've tried from two different connections (his house and mine). Is it just that their FTP server is down, or is there some way in? Is there another source of these drivers (for the Hi-Grade 4400) which has a moderate to reasonable chance of giving me the right files and not some malware? Is there some source of a generic driver which might be suitable instead?  Card Zero  (talk) 13:09, 10 August 2011 (UTC)

If you can find out what kind of graphics card it has, or the integrated card model / chipset it is, you might be able to find the driver from a different source. Is it Intel? TheGrimme (talk) 17:04, 10 August 2011 (UTC)

I don't have access to it right now, but I remember the processor is a Celeron, if that helps. Not sure what software to use to identify the make of the graphics card.  Card Zero  (talk) 20:34, 10 August 2011 (UTC)

OK, I got my hands on it again and found out:

GPU: SiS 315 Intergrated

Chipset: Northbridge: SiS 650

so I did the obvious thing and found the SiS site and downloaded the SiS 315 series GPU drivers for the right system, 32-bit XP, and ran the setup program ... and got the message "The setup program could not find a suitable driver, the setup program will terminate". Huh? Why on earth didn't that work? What else can I try? (Update) I found I could install the SiS 315 driver via the Windows control panels - it must be in a CAB file somewhere I suppose - but then I get "this driver cannot start" once it's installed. Perhaps the system profiler I used (Everest) got it wrong, and the GPU isn't SiS 315? I also tried hw32_384 and SiSoft Sandra (two other profilers) but couldn't get either to run (the first said something about missing setup files, and the second just triggers a reboot for some reason whenever I try to run it). Is there a simple way to find out what the GPU is?  Card Zero  (talk) 22:56, 12 August 2011 (UTC)

Got it! I didn't need the SiS 315 GPU driver, I needed the SiS 650 IGP driver. Sorted.  Card Zero  (talk) 23:52, 12 August 2011 (UTC)

undefined reference to vtable

I'm trying to compile a program I've been writing, for which I have received much helpful advice on this ref desk, and now it's almost done, I'm getting a compile-time error, "undefined reference to vtable". Needless to say, I'm using virtual functions, but there doesn't seem to be anything wrong with the program, and the error message seems to be from the underlying aspects of the compilation process (such as the linker). Ordinary compiler errors state a line number in the .cpp file, but here there are no line numbers, just a lot of weird information. Here's some of the info it dumps:

/tmp/ccTEIg2P.o(.gnu.linkonce.t._ZN4NodeC2Ev[Node::Node()]+0x7): In function `Node::Node()': diffs.cpp: undefined reference to `vtable for Node'

/tmp/ccTEIg2P.o(.gnu.linkonce.t._ZN4NodeD2Ev[Node::~Node()]+0x7): In function `Node::~Node()': diffs.cpp: undefined reference to `vtable for Node'

I'm using a rather old version of Linux (Fedora Core 4) and g++, with the compile command, g++ -Wno-deprecated -o diffs diffs.cpp. Is there a fault in my program, or is it g++, and can I get around it, or do I need to upgrade? Is there by any chance a free compiler around somewhere online, that can take text and produce an executable, as a temporary stopgap that will enable me to double-check it is the computer not me? Thanks again in advance, It's been emotional (talk) 17:23, 10 August 2011 (UTC)

Try defining the parent class functions as pure virtual functions, by specifying their definition this way:

  virtual void myFunction(int myArg) const = 0;

This is a common error. Make sure all virtual methods, including the constructor and destructor, are either pure virtual ( =0) or are defined. I find the = 0 syntax to be one of the most unintuitive aspects of the C++ language. Java uses the abstract keyword, instead of setting a pure virtual function to 0. Nimur (talk) 22:20, 10 August 2011 (UTC)

Thanks, because I never would have worked that out, and I don't think I ever would have found that link. I'm now getting a segmentation fault, which is glory, because I know how to deal with that one. It's been emotional (talk) 02:45, 12 August 2011 (UTC)

Different Voltage slower computers?

I have a Dell Studio laptop and I use it in North America with 100-110 V. Recently, I travelled to a country where the AC voltage is 240 V. After using my computer for a few days, the computer would freeze after been used for a few minutes and could only run properly in safe mode. My adapter says it can handle anywhere between 100-240 V voltage. After returning home to NA, the computer is running properly again. Could the AC voltage have had an impact on the performance of the computer? Acceptable (talk) 17:50, 10 August 2011 (UTC)

"Theoretically," the voltage should have had no effect on the computer. Typically, by the time the computer's internal ACPI power-supply sees the power, it's already been conditioned by the brick-shaped thingie in your cord (the AC/DC power supply). Analog circuitry inside the AC/DC supply should have produced a clean, standard DC current and voltage level (the same output, no matter what the input voltage); and the computer shouldn't have had any "knowledge" that its original input was 220 volt mains power.

In practice, if the AC/DC supply is imperfect (!), it may have supplied a different DC current or voltage when its input was 240 V. Any number of software or firmware level "glitches" may have occurred as a result: ACPI drivers tend to be flaky and can contribute to system instability. Nimur (talk) 18:13, 10 August 2011 (UTC)

I work in IT support and although I'm not a "hardware person", I think it is extremely unlikely that voltage had anything to do with your laptop being slow and crashing. It might have just been a coincidence, or maybe there is something else that you do when you are traveling that you don't do at home, like don't start up shut down properly all the time, or boot up not on your regular network, or something else. If it was a desktop computer, maybe you could convince me of some strange power filer effects or something, I've seen desktop power supplies cause "gremlin" issues, but on a laptop the adapter is essentially a glorified battery charger. Having said that, I'm not saying it's IMPOSSIBLE, I learned long ago never to absolutely rule out any hypothesis without solid proof, no matter ridiculous or implausible it sounds, but I would investigate other more likely things first, like look through your event log.. Vespine (talk) 22:55, 10 August 2011 (UTC)

And for reference, there's Power_supply_unit_(computer)#Laptops. It doesn't say much except to add that the power that comes out of the adapter is conditioned a second time by the laptop it self. Vespine (talk) 01:06, 11 August 2011 (UTC)

Mains voltages in the UK can legally exceed 250v, and do so regularly in country areas, but even this should not, in theory, change the voltage output within the laptop. (My "power brick" is designed for 100 to 240v AC and runs fine on 252v without causing problems.) As others have said, "shouldn't" doesn't mean "didn't". I would have thought that fluctuating voltages would be a more likely cause. My laptop behaves very oddly when I (foolishly?) try to run it in my car using a 12v to 240v inverter power supply. The two switch-mode circuits in series evidently interact in some strange way. Dbfirs 08:59, 11 August 2011 (UTC)

explicitly forbid a Windows 7 process from consuming more than 95% of the CPU

Sometimes, weird things happen where a weird browser plugin malfunctions and starts consuming 100% CPU. I've always wondered why OS developers let that sort of thing happen -- shouldn't it always allow for a "small reserve backup" to allow you enough CPU cycles to decently call up a process management program to kill to deal with the problematic process in question? Are there any rules you can implement that allow noncritical programmes to behave normally, but will never give them more than 95% of the CPU cycles? If not, why doesn't such a programme exist? elle _{vécut heureuse} ^{à jamais} (be free) 18:54, 10 August 2011 (UTC)

Windows' kernel allows you to set preemption properties: Scheduling Priorities. You can change your program's shortcut to always start with a certain priority-level. Here are some tips for different ways to set that up.

If you're asking the more philosophical question: why doesn't the Windows GUI (explorer.exe, or its child processes) forcibly preempt user-space programs? This is a design feature; Windows permits a user-application to use the entire system's resources. In certain other operating systems - most notably, a true Unix such as Solaris, your system administrator may specify hard preemption limits on specific processes (including the GUI or X Server). However, the details of making such a kernel-scheduling tweak are very complicated - undesirable side effects, like system instability and deadlock, may occur. For this reason, most computer platforms, such as certain Unix and all Linux and Windows, are designed to be "interactive with users." Their schedulers allow you to "pleasantly ask the Kernel to please preempt" a specific program. (No hard-guarantees). If you are a kernel engineer, you can easily design your computer system to never deadlock; but most users are not kernel engineers; so this is the better design-decision for a desktop computer. Technically, your computer never crashes when Firefox goes haywire - it's just sluggish. If your computer were a flight-control hardware for an airplane, it would be better to crash Firefox than to crash the plane because of sluggish response. Nimur (talk) 19:25, 10 August 2011 (UTC)

There is never a need to keep "spare" CPU cycles: at any given point in time the CPU might as well be doing something. There already exists a notion of priority in scheduling, which allows for programs like BOINC to soak up 100% of unused CPU time without affecting normal operations. The problem is probably, as Nimur said, that the UI just doesn't have high enough priority (looking at my Ubuntu system, it looks like X is running at normal priority, so the same thing can happen to it). I don't know enough about these issues to explain why this is the case, though. Paul (Stansifer) 02:45, 11 August 2011 (UTC)

There are still priority inversion-like scenarios on non-real time systems, even if CPU/IO priorities are being used. For example, a rogue process might allocate a lot of memory or access a lot of data on the disk, forcing the data of other programs and the UI, even if running at a higher priority, out of memory. When these programs are accessed again, their data needs to be read back from disk, and during that time the lower priority processes can continue to run (and perhaps make matters worse). See Thrashing_(computer_science). Unilynx (talk) 10:06, 13 August 2011 (UTC)

Where to find certificates in Windows 7?

Our application at work uses certificates for WCF authentication. My former boss made a great program which automatically creates a certificate, places it in the "My" store in the "LocalMachine" location, and writes its thumbprint in the WCF <serverCertificate> configuration. But my question is, how do I actually find this certificate in Windows 7 so I can view what it contains? I tried the Certificate Manager (certmgr.msc) and it found a certificate with the same subject name and creation date, so I guess it should be it, but neither the key ID or the thumbprint were anything like what my boss's program had written in the <serverCertificate> configuration. Yet the WCF service manages to find the certificate. How does all this work? Have I found the actual certificate in certmgr.msc? If yes, then why is its thumbprint different? If not, then where is it? JIP | Talk 19:51, 10 August 2011 (UTC)