Wikipedia:Reference desk/Archives/Computing/2009 November 30

= November 30 =

How the algorithm to fade from one image to another is implemented
Is it usually done by

1)For each pass in a loop depending on how quickly the fade is to be, completely replace every nth pixel of image1 with the corresponding pixel of image2 and decrease n so slightly more pixels will be replaced the next time around until all pixels have been replaced.

2)For each pass in a loop depending on how quickly the fade is to be, modify every single pixel value according to a calculated shift so every image1 pixel value transforms into its corresponding pixel2 value with many values belonging to neither image occurring during the transformation.

3)Something else I didn't think of.

If it's either of the two I could think of, I'm betting it's #2 since #1 would probably appear spotty even if the resolution was sharp, but it (#2) sure seems like a lot of calculation looping through every pixel. Then again, I know computers are just very fast. 71.161.61.41 (talk) 00:10, 30 November 2009 (UTC)


 * Normally, it is done by completely overlaying one image on top of another image. Then, through a loop, change the opacity of the top image until it is invisible and only the bottom image is visible.  The end result is that the video output will average each pixel of both images together based on opacity of each image - but the user programming the fade isn't doing that.  The person who programmed the video driver (or similar) does that part. --  k a i n a w &trade; 00:18, 30 November 2009 (UTC)


 * Yes, #2. What you're doing is generally linear interpolation (although there might be circumstances where you'd flatten the two ends of the line horizontally a bit, to give the effect a "soft landing"). Yes, it's a lot of bit shovelling, but you're right, modern CPUs are way fast. This process is generally called alpha blending (with you changing the alpha value to make the fade effect). Some very resource-constrained platforms to perform alpha blending using the #1 method you describe - Quake 2 did this (if you didn't have a decent video card) and called it "stipple alpha". -- Finlay McWalter • Talk 00:21, 30 November 2009 (UTC)


 * It's irrelevent how fast your CPU is - blending is best handled with shader code in the GPU. Linear interpolation happens at ungodly speeds in the GPU.  In order to get a nice-looking cross-fade, you generally want to ramp the alpha of one image against the other with a smoothed off ramp - not a straight linear ramp. I generally use a sine-wave shape to vary the alpha over time.  What that does is (a) look smoother because there is no abrupt onset and ending of the fading and (b) it gets over the 'confusing' part where the two images are roughly equally represented in less time...it just looks way better. SteveBaker (talk) 02:31, 30 November 2009 (UTC)

Refilling an HP printer ink cartridge
I am very experienced in refilling black ink cartridges. All the instructions I've seen for refilling an HP45 cartridge say that you should make a hole underneath the cartridge, where it looks as if the original factory filling hole was. Is there any good reason not to ignore this and make the hole above the cartridge, where the ink is less likely to leak out and filling is so much easier? The Lexmark printer I previously used had cartridges where it was easy to remove the loose top and pour ink onto the sponge inside, but does the HP45 have - I speculate - some sort of siphon arrangement that requires an air tight seal? Without destroying a cartridge, does anyone know what the HP45 would look like if cut in half? 78.147.183.186 (talk) 00:55, 30 November 2009 (UTC)


 * One reason to fill that cartridge from the bottom is that you already have a hole there, rather than needing to drill one. Of course, leaking is possible in that location, but maybe that's why they put the hole on the bottom, to discourage home ink refills. StuRat (talk) 03:09, 30 November 2009 (UTC)

Chatzilla DCC automation
I use Firefox with Chatzilla 0.9.85. On some channels, public DCC file servers accept requests with the syntax ! Can I set up a script so that if I type a request with this syntax in Chatzilla, it will automatically accept any offer from that user to send a file with that name? Also, can I set one up so that if certain usernames (who are search bots) offer .txt.zip files, they will automatically be accepted and unzipped and their contents displayed in the chat tab? Neon Merlin  02:08, 30 November 2009 (UTC)

Uniloc Patent
Does the Uniloc patent cover product activation in general or product activation with a key? Either both seem very broad to me and cover basic things that don't seem patentable. Although I can see a specific method of product activation using a key with a couple different protocol steps as patentable. --Melab±1 &#9742; 02:10, 30 November 2009 (UTC)


 * You can read the patent here. SteveBaker (talk) 02:16, 30 November 2009 (UTC)

Fractional bit of entropy from /dev/random
In Linux, is it possible to get less than one bit of entropy at a time from /dev/random, in the form of a bit whose probability of being 1 is something other than 0.5? (For example, one could then simulate a six-sided die using a first bit with probability 1/3 of 1 and 1 to 2 bits with probability 1/2 of 1, and then interpret 000 as 1, 001 as 2, 010 as 3, 011 as 4, 10 as 5 and 11 as 6.) Neon  Merlin  02:54, 30 November 2009 (UTC)
 * I do not use Linux, but couldnt you just use /dev/random three times to get three bits, giving you eight different numbers with an equal probability? If the number obtained was seven or eight then re-do the proceedure. I recall it is possible to get a Guassian distribution for example by adding (I think) several instances of the more usual 0 to 1 probability function, due to the central limit theorem, and you should be able to do this with /dev/random as well. For example if you added /dev/random 100 times you would get a mean of 50 - cannot remember the formula for the standard deviation. 78.146.171.75 (talk) 11:18, 30 November 2009 (UTC)


 * Disclaimer: I haven't the faintest idea how /dev/random/ works . However, I was under the impression that it behaves like a file. If so, reading individual bits would be impractical anyways. As 78.x.x.x indicates, a common algorithm is to use a modulo operation to obtain a number in the desired range, discarding outputs from the random number generator that would result in an uneven distribution. decltype (talk) 11:35, 30 November 2009 (UTC)


 * You cannot read "less than one bit" from /dev/random. Obviously, a bit is as small as it gets.  Further, you will have difficulty reading less than a byte (8 bits).  So, your best bet is to read one byte and mod by 6. That will randomly give you 0-5.  Add one to the result to get 1-6.  If your problem is that /dev/random is blocking, use /dev/urandom.  It is slightly (very slightly) less random, but won't block. --  k a i n a w &trade; 13:14, 30 November 2009 (UTC)
 * Yes, but as noted above, it is important that you then discard outputs that result in an uneven distribution (252-255 inclusive). Otherwise your die will have a lower probability of getting a 5 or 6. decltype (talk) 13:24, 30 November 2009 (UTC)


 * The idea is to avoid wasting more entropy than is needed to generate the number, in cases where entropy is in short supply. The above method, if it's to be unbiased, will spend slightly more than 8 bits of entropy on average (since 252-255 will have to be rerolled), while a 6-sided die is theoretically (and on average by the method I suggested) only 2.585 bits. Neon  Merlin  13:30, 30 November 2009 (UTC)
 * In that case, I would suggest generating multiple dice rolls from a single read of random bits, and buffer the results. To generate five rolls with regular six-sided dice, you can get away with thirteen bits, which is close to the optimal 12.925. If you can find a power of 6 that is closer to a power of two, you could do even better, but that would require a large number of bits and a lot of computation. Even generating three rolls from a byte would be an improvement (generate a number 0-215 and discard the rest). decltype (talk) 14:04, 30 November 2009 (UTC)

Regarding SGML
I am working in a SGML file using EPIC editor software recently I have faced a problem while trying to print the document. It shows some error messages 1. some unrecogonized characters were present and 2. Invalid revision Bar point

I cant able to rectify this error if any one have solution for this are welcome to give their solutions.

Thanks and regards

DINESH KUMAR B —Preceding unsigned comment added by Dineshbkumar06 (talk • contribs) 04:14, 30 November 2009 (UTC)

Where Do The Aquamacs .el Files Live?
Hi, I’m just getting started with Aquamacs on OS X 10.6 and I’m not sure where I’m supposed to add customizations. For example, I’m trying to install a markdown mode (http://jblevins.org/projects/markdown-mode/) but I don’t know where to put the actual .el file. Also, I’m just putting my .emacs files in my home directory. Is that the best place for it? —Preceding unsigned comment added by 72.234.155.128 (talk) 04:31, 30 November 2009 (UTC)

PHP - database to form
For existing records in a database that are input via form, to allow an authorized user to edit those records, could I:

1. make an edit button run a query on that record's full database entry and return all current values

2. slide those values (& nulls or no datas) into the original data entry form as the default values

3. have this edit form trigger an UPDATE SET query

and be done with it? Is it really that simple? I'm wondering about records that have database values which have never been entered - so not nulls, just no data - how would they return in the query array, just empty fields? Does anyone see a problem with this? It seems a little too simple...218.25.32.210 (talk) 07:47, 30 November 2009 (UTC)
 * I was unable to understand your question. Could you perhaps include an example to make it clearer?  --Sean 14:25, 30 November 2009 (UTC)


 * If I remember correct, different data types (different INPUT types, really) return differently depending on what happens if they are blank. So if it is a text string, you get a blank text string in your return array. But if it is, say, a checkbox, then you get NOTHING in the query array when you submit it empty, which means you have to then look for it to know that it was returned as blank. (That is, you don't get a variable with nothing in it — you get no variable. So you only see the variable when it has been positively checked.)
 * But other than that -- it's pretty straightforward, yes. You just have to make sure that you are careful about how the different datatypes return, and of course all of the different INPUT types have different ways of indicating default values (irritatingly enough). --Mr.98 (talk) 14:41, 30 November 2009 (UTC)


 * OP here. Sorry, sean. This was all back of the envelope so I didn't have any code to provide. Thanks, Mr. 98 - I'll keep a close eye on each input type's details, hadn't thought of that. 61.189.63.183 (talk) 23:05, 30 November 2009 (UTC)

C# abstract classes
In C#, is it possible to specify that a given nested class A inside an abstract class B is abstract, but that in non-abstract subclasses it must not be? Also, is it possible to specify that where said B.A implements IEnumerator&lt;B&gt;, the C.A inside any given subclass C must implement IEnumerator? Neon Merlin  07:56, 30 November 2009 (UTC)


 * Not as far as I know. Why would you possibly want to do either of those? Some reflection-based framework? « Aaron Rotenberg « Talk « 20:28, 30 November 2009 (UTC)

Use phone as bluetooth repeater from laptop to headphones?
Can I configure my Samsung VICE R561 phone to serve as a repeater for music my laptop sends to my headphones over Bluetooth A2DP? The VICE itself supports A2DP, its application platform is J2ME, and I'd be willing to install third-party firmware if necessary as long as I can still use the phone functions that I consider to be core (talk, text, WAP, alarm clock, appointment calendar, camera, phone's own music player). Neon Merlin  10:16, 30 November 2009 (UTC)

Is it good programming practice to use state variables?
I am only an amateur programmer. Currently I am writing a subroutine that checks that the data in a text file is in the right format when loaded into the program. For example I want to ensure that the command word "start" is followed by the command word "finish" exactly five lines later, and if not give an error message. Is it best practice just to have a variable that goes to a value of 1 when "start" is found, and then counts up every file line until it gets to 6 when "finish" should appear, or is there some better way to do it? There are other state variables as well. I'm thinking that state variables may be difficult to follow if I want to revise the program in the future, and perhaps lead to spaghetti code. Thanks. 78.146.171.75 (talk) 14:38, 30 November 2009 (UTC)


 * It's difficult (and often too subjective) to be so definative about what is good (or, perhaps, bad) practice, particularly in such an abstract circumstance. In some circumstances it may be best to have state implicit in the program structure (e.g. we can't get to this line unless we've already seen the start condition). In other cases it might be better to explicitly have a state machine. In general, when it comes to variables, states, and function names, I have a (rather whorfian) theory - if you can give things a sensible name (e.g. STATE_AWAIT_END_TOKEN, wait_for_next_line, bPacketHeaderReceived) then you still understand your program, but when you define some thing that owes its existence to the exigencies of your specific implementation, where that variable means this-is-true-but-not-that-unless-something-else (where you can't give that condition a plain english definition), then your understanding of what you're coding may be slipping. -- Finlay McWalter • Talk 15:06, 30 November 2009 (UTC)


 * I would say that state variables are a good concept. The implementation of "state" can range from great to terrible.  Compare, for example, Entity Java Beans, which are basically persistent state variables; against Global variables in C or C++.  In the first case, the persistent state is well managed by the programming language syntax, the runtime environment, and a lot of the abstractions are properly and completely handled.  In the second case, global variables in C++ tend to result in a leaky abstraction because they don't enforce certain requirements (like thread consistency, concurrent access, locking, validity checking, scope, etc).  So, while a state variable is probably a good thing in general, its implementation determines whether it is a good design choice for your particular problem.  Nimur (talk) 15:41, 30 November 2009 (UTC)


 * It's good that you're thinking about these issues, but I wouldn't sweat the style/understandability of anything that can fit in just a few lines of code like what you're describing. Lexically scoped variables in short functions are usually obvious even without decent names. --Sean 18:06, 30 November 2009 (UTC)

Thanks. I'm curious, would it be possible to write a program entirely as a giant state table, apart from where the program interfaces with other things? 89.242.99.245 (talk) 19:54, 30 November 2009 (UTC)
 * It is always possible in principle because computers are finite-state machines, but the resulting program may not fit in memory because the number of states tends to grow exponentially with the amount of storage the program would require if written normally. A gigabyte of RAM can be in any of 28589934592 different states. State tables are frequently used in practice, though, usually in machine-generated code or as a poor man's substitute for coroutines (which would be supported by all major programming languages in a sane world, but generally aren't in the real world). -- BenRG (talk) 20:38, 30 November 2009 (UTC)
 * You might be interested in Boost Statechart, which helps you model your program as an FSM. --Sean 20:53, 30 November 2009 (UTC)

System admin blocks access to C: drive
My system administrators at work have blocked access to the C: drive on my office computer. I can only save files on the network. As a result nothing is saved (cookies, passwords, MS Office customizations), it all disappears overnight and has to be laboriously reinput each morning. As you might imagine, this is incredibly annoying – and also bad for productivity. I've raised the issue with the IT department but to no avail. Is there any way for me to unblock this access? Thanks. --Bluegrouper (talk) 16:02, 30 November 2009 (UTC)


 * If you actually value your job I really wouldn't recommend trying to bypass the access restrictions. They've been put in place specifically and depending on your contract, trying to get past them could be seen as a breach of contract. It sounds very much like they don't want people customising the computers so in reality you're better off either not changing them back every day (which is the loss of productivity) and if there is something specific that would help you out, you're best off explaining that change rather than asking for complete access. Incidently, regarding only able to save files on the network, this is probably for backup purposes as the individual machines probably aren't backed up, but the network fileserver is. ZX81  talk  16:32, 30 November 2009 (UTC)


 * Have a look at PortableApps and load up a USB stick. At the very least you can browse the internet with a saved personal history, passwords, cookies. 61.189.63.183 (talk) 23:07, 30 November 2009 (UTC)


 * All these settings should be saved to the network as a roaming user profile. If your I.T. dept. does not have you (and other users) properly set up, they are negligent in their duties. --Nricardo (talk) 02:02, 1 December 2009 (UTC)


 * It's probably a mandatory user profile (same link you gave) which is a read-only profile by design (to stop changes being made). ZX81  talk  02:15, 1 December 2009 (UTC)

Google Earth
I lay out a path on Google Earth, do the whole mileage thing, and it all works great. But when I go to close out, I really need a way to save this (my plot is over 600 miles); I'm trying to plan out a major trip and need Google Earth to pull through for me. Thanks! Hubydane (talk) 16:42, 30 November 2009 (UTC)
 * You could try using some different software. Map24 is pretty good for this sort of thing. --Richardrj talkemail 16:53, 30 November 2009 (UTC)


 * When you have your directions in the "Directions" window, just drag the direction name (e.g. "New York, NY to Niagara Falls, NY") with a little blue globe on it into the "Places" window, and it will save it there. Next time you want to view it, just click on it in the "Places" window. --Mr.98 (talk) 17:30, 30 November 2009 (UTC)

My apologies; I'm not actually doing a road trip. I'm kayaking a river, so I can't exactly do from point A to point B and get directions. —Preceding unsigned comment added by Hubydane (talk • contribs) 18:29, 30 November 2009 (UTC)


 * Don't know about Google Earth (never used it), but in Google Maps, there's a "Link" button in the upper right which gives a URL which encodes all the complete state of the window when you click it. Perhaps there is something similar in Google Earth (not a "save" feature per se, but an email/link/send type option.) -- 128.104.112.95 (talk) 23:48, 2 December 2009 (UTC)


 * You could just copy the screen and paste it into paint, that's what I always do. then you can do anything you want with it and print it out easily. It helps if you copy it section by section and put them together afterward, then you get a closer zoom and more detail.148.197.114.158 (talk) 18:57, 4 December 2009 (UTC)

C question
I have a structure like that given below:

How do I access aP (and set its values) inside main, using variable obj? Please help. I've tried everything. Out of luck today.--117.196.129.72 (talk) 16:50, 30 November 2009 (UTC)


 * you'll have to allocate space for the structure and the array first of course. Does aP really need to be a pointer to a pointer?&mdash;eric 17:20, 30 November 2009 (UTC)
 * Or obj->aP[j][i] or *obj->aP[j] depending on what you're trying to do. What is aP? -- BenRG (talk) 17:45, 30 November 2009 (UTC)


 * It appears that aP is "array of pointers". This would be better as a class (or a struct if you want to abuse structs) that required the size when you created the array of pointers and automatically declared/destroyed memory as required.  But, you'd just be doing a lot of work to remake the already available vector of pointers. --  k a i n a w &trade; 20:29, 30 November 2009 (UTC)
 * Opie said "C", not "C++". --Sean 21:01, 30 November 2009 (UTC)

Only using 50%
On Windows 7 some programs only ever use 50% max cpu when they clearly need more, but Windows is apparently limiting them to only 50%. How can I let them use the full 100% ? —Preceding unsigned comment added by 82.44.55.75 (talk) 17:33, 30 November 2009 (UTC)
 * Most likely you have a dual-core CPU and the program is using 100% of one core and 0% of the other. Some programs may have a "number of threads" or "number of cores" option that you can set to 2 instead of 1. Otherwise, there's nothing you (or Windows) can do to make the program use the second core. -- BenRG (talk) 17:42, 30 November 2009 (UTC)


 * To elaborate a little further: whether a program can use multiple cores depends on how it is programmed. Many programs these days can seamlessly use multiple cores (games are pretty good at this, for example), but many cannot. Here is an article from a couple years back now that discusses how some programs can really make good of multi-core processors, while others cannot. It is primarily about going from 2 to 4 cores, but the basic concept is the same. Programming for multiple cores is not trivial, depending on the type of program it is—some are really amenable to having multiple threads working in parallel, but some are not. --Mr.98 (talk) 18:23, 30 November 2009 (UTC)


 * Another possibility is that you are on a laptop and the CPU is limited by the power settings to 50% to conserve energy.F (talk) 03:28, 1 December 2009 (UTC)

Another question regarding this
I (not the original OP!) have another question about this: I know that if you write an app that will consume as much CPU time as possible (e.g. by entering an infinite loop), it will only use 50 % of my CPU power (my CPU has two cores). However, when I read the indicators on the CPU usage Vista sidebar gadget I use, it says that Core #1 is used to 50 %, and Core #2 to 50 % as well. Shouldn't it be 100 % and 0 %? --Andreas Rejbrand (talk) 18:54, 30 November 2009 (UTC)
 * I'd question the accuracy of the sidebar gadget, because I'd also expect it to be 100% and 0% if it's only using one thread. It could be the gadget is just taking the total load and assuming it's shared across two processors although that's a pretty bad assumption if it is. The way to really check would be to look on Task Manager's "Performance" tab which will have a graph for each processor. Run your loop and see if only one of the processors greatly increases (I just tried a simple "do loop" in VB6 and on my quad cord only one processor increased as expected. ZX81  talk  19:10, 30 November 2009 (UTC)


 * It appears like the gadget is correct. See cores.png @ privat.rejbrand.se for a screenshot. I know that AlgoSim only use one thread for the factorization (because I wrote AlgoSim), and the CPU is almost not doing anything besides the factorization. Both cores were almost at zero before the factorization began. --Andreas Rejbrand (talk) 19:15, 30 November 2009 (UTC)


 * It's also possible that it is rebalancing the load between cores frequently enough that the split ends up being roughly 50/50. I'm more inclined to suspect a poorly written gadget though. There would be little to no benefit in rebalancing the task across cores, and you'd throw away cache every time you did so. &mdash;ShadowRanger (talk 19:17, 30 November 2009 (UTC)


 * Posted that without noticing the response you made. Looks like my speculation was correct though. &mdash;ShadowRanger (talk 19:18, 30 November 2009 (UTC)


 * I would think whether a program uses 100% of one core or 50% of both will depend on how it is running at the time. If it's single threaded and running in a tight loop, then I would strongly expect it to use only one core and to consume 100% of the CPU of that core.  However, if it yields in any way (a simple Sleep(1) would do this) then it will be rescheduled and will use both cores approximately equally.  Perhaps that is what's being observed.  --Phil Holmes (talk) 10:19, 1 December 2009 (UTC)


 * Check that. I've got a quad core and an app that takes 25% of the total CPU (i.e. all of one core) in a tight loop and it doesn't max out one core - it's evenly distributed.  Windows is obviously pre-emptively scheduling and distributing the load across the cores.  --Phil Holmes (talk) 12:27, 1 December 2009 (UTC)


 * Yes this behaviour is well known and can cause problems with some processor power saving features that turn off or limit one or more cores independently. You can of course set affinity (e.g in the task manager) but I'm not sure if this always works as it should and most people are not going to know how or bother (plus if the application is slightly multithreaded it's going to cause problems).
 * But for example AMD K10 (the original Phenom's) had the ability to throttle cores independently but it was known to cause problems because applications would be switching between cores which had been throttled and which were going at full speed so that they would go significantly slower then they were supposed to because would be running on throttled cores half the time (or whatever). Alternatively they just couldn't throttle the cores but I think throttling then the process switching to the throttled core was the more common phenomon and it lead to people completely switching off AMD's Cool'n'quiet feature. This was removed in the Phenom II at least temporarily with the problems it caused usually given as the reason.
 * Similarly the Intel Core i7 and Core i5 processors are able to completely switch off cores if they aren't in use which obviously wouldn't work if the process is switches between cores. Windows 7 changed this behaviour and also introduced core parking where processes are loaded onto one core if they can  and Microsoft even seem to have filed patents . Not surprisingly that can cause problems as well and I've seen numerous discussion about turning it off.
 * Incidentally, that claims it doesn't work on AMD but I'm not sure of that. They seem to be reintroducing some sort of independent core power management at least for their Opterons (actually I'm not sure if it ever disappeared from there) and are discussing it in relation to Windows 7 e.g.  (it appears to be a feature of the Istanbul which is similar in design to the Phenom II AFAIK e.g. ) and people discuss it in relation to AMD . Perhaps the new ones are better and introduce a C6 state (although I'm not sure why C6 is necessary particularly for the "ideal core" tech, obviously for "core parking" AMD may or may not be better off without it) or perhaps the Tom's Hardware writer is simply misinformed or biased (it's not uncommon in my experience).
 * I guess the obvious question is whether anyone here is using Windows 7 (or Windows 2008 R2)? The OP was but wasn't referring to this issue. The screenshot looks like it was from Windows Vista to me but maybe the poster just changed their desktop because they preferred the Vista look.
 * Nil Einne (talk) 07:57, 2 December 2009 (UTC)


 * Thank you for your explanation. Yes, the screenshot is from Windows Vista. --Andreas Rejbrand (talk) 16:35, 2 December 2009 (UTC)

Internet Explorer has stopped working
It's the same thing that happened here but on a web site. Now the web site is working fine, and Explorer is obviously working, except when I go to the one screen and I get the Geejo cursor and it looks foggy.

I had two activities going on and the other continued as normal. At the bottom of the screen there is a rectangle that says "Smallville - page 38 ..." which, if I click on it, leads me to the foggy screen. I come back here by clicking on the black rectangle that says "Editing Wikipedia:R ..."

I did make one mistake: on the screen with the Geejo cursor I got a pop-up message with the title of this topic and a place to click to shut down Explorer, and I would receive information on how to get the problem fixed. I didn't want to quit anything, so I clicked on the Red X in that pop-up. I should have left it because I could probably have gotten an answer. I'm ready to quit now (temporarily) and see what happens. Vchimpanzee ·  talk  ·  contributions  · 22:21, 30 November 2009 (UTC)

I closed using Alt-F4 and I got a message saying I would be sent a solution to the problem if there was one. I did get something. It said "Address a problem with Flash Player". I'm reluctant to uninstall or install anything right now. I'll see if I can get by without doing it. Vchimpanzee ·  talk  ·  contributions  · 22:27, 30 November 2009 (UTC)
 * I would disable any un-necessary add-ons by pressing ALT + T and selecting "Manage Add-ons." I've fixed this issue on other people's computers by disabling add-ons like the SSV Helper, Messenger, etc. You probably have a bunch of toolbars that you never use, too, right? If you do, then uninstall them from the Programs and Features dialog in the Control Panel. That's all I can say because all you provided for an explanation is a link to a thread that says "Geejo strikes again."--Drknkn (talk) 22:28, 30 November 2009 (UTC)
 * Well, that thread provides a description of what happened to me. User:Geejo once had a signature that simulated a cursor that gets stuck. Vchimpanzee ·  talk  ·  contributions  · 22:32, 30 November 2009 (UTC)


 * I don't actually know what I'm not using. None of those look unimportant. Like I said, maybe the best thing to do is see if it happens again. Installing stuff scares me because I don't know if something might get messed up. Vchimpanzee ·  talk  ·  contributions  · 22:35, 30 November 2009 (UTC)


 * Is this one of the library computers you use, or is this on a computer you own? Could you explain what you mean by the "foggy screen"?  Does this mean the window grays out and when you try clicking its "X" close-box, a message appears saying that the application is not responding?  The workaround, VChimpanzee, is to use another browser for a while, like Firefox, Safari, Opera, or any of the other browsers in the world.  There is something wrong with your computer setup, and the workaround is to try another browser.  If this problem goes away, then keep using the new browser choice, and don't use IE anymore.  Imagine how much nicer your life would be without this giant assortment of difficult-to-describe and difficult-to-diagnose annoyances.  While I'm at it, I will remind you that you should switch away from Yahoo Mail, too.  Comet Tuttle (talk) 06:33, 1 December 2009 (UTC)


 * I'm not changing browsers or email, so please drop it :). You have to understand that 99 percent of the time it works fine, and I'm just trying to figure out what's going on. As it turned out, I did get an informative message on my computer, but I want to see if I can get by without it happening again rather than go through a change that would be more trouble than it's worth. Firefox is something I will never choose--it causes problems of its own, including the AT&T ad--but it's what I use at this one library while seeking to avoid possible problems with most web sites at home.


 * As for the email problem, it will frustrate me, yes, but I have to remember not to go to Yahoo email unless I don't intend to go back where I was. I usually keep one screen reserved for whatever I need to look up, or a "notepad" that I'm saving. But anytime I go to an email, I should expect to click "Inbox" rather than "back" to go to another one. I've made that little adjustment and, while occasionally frustrating, it works fine. Vchimpanzee ·  talk  ·  contributions  · 23:23, 1 December 2009 (UTC)
 * Again, I swear to God that I saw this exact problem on a customer's computer. I also want to repeat one more time that I fixed it by uninstalling all the toolbars on his computer and disabling the SSV helper add-on. Just look in the "Currently loaded add-ons" screen and disable everything except Java, Flash (AKA "Shockwave Flash" or something like that) and the PDF add-in. 99% of the add-ins in that screen didn't come with your browser, and you can always re-enable them! Even if the problem doesn't happen that often, you should disable the add-ons, anyway, because it will speed up the browser and make it more secure. These add-ons are often attacked by malicious web sites. An up-to-date Internet Explorer browser is very secure, but if it has an insecure plugin, hackers can use the plugin to hijack your browser!--Drknkn (talk) 11:59, 2 December 2009 (UTC)

I looked under "Manage Add-ons" and got this list:


 * Shockwave Flash Object
 * Adobe PDF Reader Link Helper
 * Windows Media Player
 * Java Plug-In SSV Helper
 * Show Norton Toolbar
 * Norton Confidential
 * Symantec Intrusion Prevention
 * Yahoo Toolbar
 * Yahoo Toolbar Helper

I noticed there were other categories. I found Bing and Yahoo under search providers and disabled Bing.

Accelerators were Blog with Windows Live, E-mail with Windows Live, Map with Live Search and Translate with Live Search. I don't use any of those but don't see anything about disabling them.

There was a fourth category which didn't have anything listed.

Regarding "Malicious web sites", at home I only intentionally go to four web sites plus three email addresses, and if the college library is closed on a weekend or on a holiday when everything is closed, I go to four newspaper web sites which CAN cause pop-up ads even though those are blocked. This is because I want to minimize potential problems, and I had a spyware problem at a library and the person in charge said if I had been at home (at the time I was too scared to get my own computer) it would have gotten into the computer, but the library's software stopped it.

I blame an ad for what happened when I asked the above question. I did say that when I closed the Internet I got a message to upgrade Flash, but unless it becomes a frequent problem, I'd like to avoid that for fear I'll cause new problems. I believe an ad messed up a library computer last week, but I was pleased that the solution did not shut down the Internet like it usually does. I was able to keep going. Vchimpanzee ·  talk  ·  contributions  · 17:50, 2 December 2009 (UTC)

Oh, one more thing. Go here and click on Geejo's name and you can see what's happenng to me, or at least what would happen if the same thing happened on your computer. Vchimpanzee  ·  talk  ·  contributions  · 17:52, 2 December 2009 (UTC)


 * Update: I got that Geejo cursor for about five seconds and then all was fine. I looked and found an hhgregg ad with snow falling. That could mean something, since I was never able to see what caused the problem Monday. Vchimpanzee ·  talk  ·  contributions  · 20:36, 2 December 2009 (UTC)


 * Another update: on a library computer this morning, I don't recall what the cursor did but that spinning circle with the name of the web site I was on never stopped. I tried going to another site and got a pop-up with the URL of the site followed by "is not available". Which wasn't true. I did notice a yellow triangle and a message about errors on the page at the bottom of the screen, and now that I recall, I could have clicked on that yellow triangle and found out what was wrong. I was still able to use the computer, however. Because I had several rectangles at the bottom of the screen. Vchimpanzee ·  talk  ·  contributions  · 19:38, 3 December 2009 (UTC)


 * Shockwave Flash Object: keep this
 * Adobe PDF Reader Link Helper: keep this
 * Windows Media Player: keep this
 * Java Plug-In SSV Helper: disable this
 * Show Norton Toolbar: disable this
 * Norton Confidential: disable this
 * Symantec Intrusion Prevention: disable this
 * Yahoo Toolbar: disable and uninstall this
 * Yahoo Toolbar Helper: disable and uninstall this

Uninstall everything from Yahoo! in the Programs and Features dialog. All you need is a browser to get to your Yahoo! e-mail (unless you use one of the local programs -- which you don't, right?) As for the accelerators, if you don't use them, disable them: ALT + T --> Internet Options --> Advanced --> Browsing --> Display Accelerator button on selection. Perhaps you should also tell Norton to leave your browser alone completely. Open up Norton's options and disable browser protection. Like I said above, IE is very secure, especially in Windows Vista and Windows 7. If you were using XP, I'd say otherwise, but it's safe to disable browser protection in this case.--Drknkn (talk) 00:02, 4 December 2009 (UTC)
 * My Internet provider has protection for my computer, and I was told to uninstall Norton to use it, but I was afraid to try. I felt confident enough to try this year, but Norton renewed my subscription before I had a chance not to. I'd really rather have protection. I've never had a virus or anything like that (that I know of), and Norton is doing its weekly scan right now. All it ever finds is cookies, but I feel better. I disabled the toolbar, but it tells me I'm at risk for phishing. So I enabled it again. You may feel secure, but I don't.


 * I like my Yahoo mail button. I can look at some of the things you say to disable, but I need that simple way to get to email. This is one reason I mainly use Yahoo at home.


 * That leaves the Java helper. Let's see what that does. Vchimpanzee ·  talk  ·  contributions  · 19:10, 4 December 2009 (UTC)

Caps lock function like typewriters?
The caps lock key on at least some typewriters locks into place and is released by pressing the shift key (or the caps lock key again). This was much superior to it staying on until caps lock is depressed again for a number of reasons. Is there any way to make a computer keyboard do this? I am using a Logitech Cordless Desktop EX110. I tried going to control panel → keyboard but there's nothing relevant in there.--Fuhghettaboutit (talk) 23:27, 30 November 2009 (UTC)


 * If I remember correctly, this functionality is actually implemented in Microsoft Windows, and activated by making a small change in the registry. Hopefully I am right, and that someone remembers the registry key and value to be modified. --Andreas Rejbrand (talk) 00:37, 1 December 2009 (UTC)


 * Now I found a simpler solution, that works in Windows Vista (and probably XP too). Go to the "Regional settings" control panel applet, select the "Keyboard and Languages" tab, and then click the "Change keyboard layout" button. In the "Advanced Key Settings" tab in the new dialog box, select the "Press the SHIFT key" checkbox. (I do not have an English language copy of Windows, so the actual labels might differ.) --Andreas Rejbrand (talk) 00:47, 1 December 2009 (UTC)


 * By the way, if you search Google for "typewriter caps lock windows", the first match is , and the first link on this page leads to , where the solution is presented in a very clear way. Google is your friend! --Andreas Rejbrand (talk) 01:00, 1 December 2009 (UTC)


 * It's ironic how many times I've said those exact same words to people and here I am being told the same thing. Sigh. Somehow, for this, I didn't think it would be a Google search away but obviously I was wrong. The link with explanations for Vista isn't really useful for XP which I'm running (it says you might have to do something different for XP but nothing I've tried work). In any event using my good friend, Google I've found the registry solution here. Thanks.--Fuhghettaboutit (talk) 04:08, 1 December 2009 (UTC)


 * Note that on typewriters the caps lock key would stay depressed until released, while on a computer keyboard it doesn't have a depressed position, it pops right back up. Thus the need for a caps lock light to show the current status, since you can't just look at the button to see if it's up or down.  (It always seemed odd to me that they put the light on the opposite side of the keyboard from the key, though.) StuRat (talk) 07:01, 3 December 2009 (UTC)

Suspected Malware
Does anybody know what vu9aijpsq5rbd5w is? According to Auslogics StartUp Manager, it runs two instances of jre-6u13-windows.exe (Java) on start up, which then start two instances of internet explorer (even though I've uninstalled IE and Firefox to try and get rid of the thing). Whenever I try to stop IE via the task manager, jre-6u13-windows.exe restarts it; when I try to delete jre-6u13-windows.exe, it just comes back. I cannot uninstall Java via the add/remove software tool or the auslogics uninstall manager. Any help would be greatly appreciated! Crisco 1492 (talk) 23:42, 30 November 2009 (UTC)


 * That's definitely the behavior of an unwanted nasty. Google "Malwarebytes" and install it. It's a free & well-built app that should be able to handle getting rid of that thing. Please report back after trying it so we know if you've been successful or need further help! 218.25.32.210 (talk) 01:08, 1 December 2009 (UTC)


 * I didn't do that, I ended up using eRecovery to wipe the drive. All's well now, except I need to reinstall Word 2007 before I can continue working. Oh well, I'll Google "Malwarebytes" for the future. Thanks! Crisco 1492 (talk) 00:46, 2 December 2009 (UTC)


 * Wiped the drive? Damn. You used a chainsaw to cut your toenails! 218.25.32.210 (talk) 01:07, 2 December 2009 (UTC)


 * Well, those toenails were breaking my clippers left right and centre! Rather get rid of them before they start growing inward. But I will get some better malware removal tools. AVG doesn't cut it for some things. Besides, it wasn't that bad. I backed up My Documents beforehand. Crisco 1492 (talk) 04:12, 3 December 2009 (UTC)