Wikipedia:Reference desk/Computing

Welcome to the computing section
of the Wikipedia reference desk.

skip to bottom

Select a section:

• Computing
• Entertainment
• Humanities
• Language
• Mathematics
• Science
• Miscellaneous
• Archives

Want a faster answer?

Main page: Help searching Wikipedia

   

How can I get my question answered?

• Select the section of the desk that best fits the general topic of your question (see the navigation column to the right).
• Post your question to only one section, providing a short header that gives the topic of your question.
• Type '~~~~' (that is, four tilde characters) at the end – this signs and dates your contribution so we know who wrote what and when.
• Don't post personal contact information – it will be removed. Any answers will be provided here.
• Please be as specific as possible, and include all relevant context – the usefulness of answers may depend on the context.
• Note:
  • We don't answer (and may remove) questions that require medical diagnosis or legal advice.
  • We don't answer requests for opinions, predictions or debate.
  • We don't do your homework for you, though we'll help you past the stuck point.
  • We don't conduct original research or provide a free source of ideas, but we'll help you find information you need.

Ready? Ask a new question!

How do I answer a question?

Main page: Wikipedia:Reference desk/Guidelines

• The best answers address the question directly, and back up facts with wikilinks and links to sources. Do not edit others' comments and do not give any medical or legal advice.

See also:

• Teahouse
• Help desk
• Village pump
• Help manual

June 7

Local printer to internet printer

What is good print server software to turn a local printer (e.g. USB connected on Windows 7) into something that can be printed to from anywhere on the internet? I can handle the IP / port setup, I'm just wondering what the best software approach would be for making printer available? Dragons flight (talk) 03:48, 7 June 2011 (UTC)

I was hoping someone would write back with a magical answer, but I honestly don't think you can. Windows 7 can only work as an IPP client, not as a server. I'll be very happy to be proved wrong, but as far as I know the server components only run on Windows server (2003 onwards, although possibly 2000). I think it might be possible on Windows XP Professional though, but I'm not 100% sure on that either. Sorry for the vague response and perhaps someone will know of a 3rd party product to replace the print spooler, but I hope this is of some help anyway!  ZX81  ^talk 17:19, 7 June 2011 (UTC)

You could probably do this with OpenVPN. The server- and client-side is supported on Windows, and with the proper configuration, you should be able to share printers. It also provides a layer of authentication (and encryption), so you don't have to worry about exposing your printer to random people on the Internet. I've used OpenVPN before, but not for printer sharing, so I'm not sure how to do this or whether it's even possible, but it's worth a try if no one else comes up with a better solution.--el Aprel (^facta-_facienda) 03:08, 8 June 2011 (UTC)

MP3 tag editor for "genre" and "composer"

Is there any program that could fix my the "genre" and "composer" tags of my MP3 collection, WITHOUT touching at the same time the titles, album, artist info that are already to my liking? Most programs I have found overwrite everything at once. 66.131.99.39 (talk) 04:25, 7 June 2011 (UTC)

Yep, Mp3tag can do just that :) (I use it for just that reason all the time)  ZX81  ^talk 04:40, 7 June 2011 (UTC)

Is it possible for it to retrieve the actual data from the Internet and ONLY fill in genre and composer? Raskolkhan (talk) 12:12, 7 June 2011 (UTC)

I don't have a copy installed at the moment to try, but their website says "Import from Amazon, discogs, freedb, MusicBrainz - Save typing and import tags from online databases like Amazon, discogs, freedb, MusicBrainz, and more.", so I'm guessing probably yes.  ZX81  ^talk 17:21, 7 June 2011 (UTC)

iTunes does this. Select a bunch of tracks. Do a "Get Info". Select the checkboxes of the categories you want to modify. It won't alter categories that aren't checkmarked. --24.249.59.89 (talk) 21:50, 28 June 2011 (UTC)

Inkscape convert text

I can't find this in Inkscape, but I feel it should exist... I have an svg that has some text in it. I want to convert the text to a bunch of objects, each with nodes and edges. I know that the text is currently represented as nodes and edges, but I cannot select a single node on a single letter and change it. So, I need to get rid of the in-between text layer and use the actual objects. Is this possible? -- kainaw™ 15:13, 7 June 2011 (UTC)

As far as SVG is concerned, text is text, not nodes and edges (yes, the text renderer that gets called by an SVG renderer does emit nodes-and-edges, but that's not how SVG stores them). Inkscape can convert text to nodes-and-edges (but not vice versa); select the text object and do path->object_to_path -- Finlay McWalter ☻ Talk 15:21, 7 June 2011 (UTC)

Thanks. I thought that placed text on a path - to make curved text. So, I didn't select it. -- kainaw™ 15:33, 7 June 2011 (UTC)

Yeah, in general object-to-path takes a structured object like a circle, star, hexagon, spiral, etc. and turns them into a path, where you get the point-editing you want, but lose the higher-level control of the original object. To make a curved text object you first create the text object and a bezier curve for the path, select both, and then do text->put-on-path. This does keep the text object as text, so you can still edit the text after it's done. But beware that (at least the last time I tried it) the SVG->bitmap renderer used by Mediawiki doesn't handle text-on-paths correctly. -- Finlay McWalter ☻ Talk 15:45, 7 June 2011 (UTC)

Still broken: compare File:Wfm bendy test.svg as viewed by Mediawiki and in Inkscape or Firefox's own SVG renderer. -- Finlay McWalter ☻ Talk 15:52, 7 June 2011 (UTC)

From memory, the MediaWiki interface doesn't do curves and lines unless they've also been converted to paths. ╟─TreasuryTag►ballotbox─╢ 09:18, 9 June 2011 (UTC)

IP Address

Is it possible to know somebody's home address through their computer's IP address? B-Machine (talk) 15:18, 7 June 2011 (UTC)

It depends entirely on who you are and who the person is. In general, the answer is "no." You (specifically you) cannot know the home address of any random IP address you find on the Internet. -- kainaw™ 15:34, 7 June 2011 (UTC)

Since the above poster didn't answer fully, I'd like to explain a bit better. You can use various tools to get the approximate location of an IP address. Many servers report geolocation, and tracert can tell you the hops between you and the subject. This is unethical, but adertisers do it, so there's no reason why regular citizens shouldn't. If you have a pressing legal need for this information(if for example the ip in question was involved in illegal activities), you can subpoena the ISP that owns the the address, and they likely have records that would tell who that is. If your goal is to stalk or harass someone you know from the internet, this solution won't work for you(thank goodness). i kan reed (talk) 16:06, 7 June 2011 (UTC)

Unless you are very good at bluffing, social engineering, or have a friend on the inside of the ISP... --Mr.98 (talk) 17:57, 7 June 2011 (UTC)

Adding a 3rd answer, possibly. If the IP in question was allocated by RIPE (Europe) and the person in question has a block of 8 IPs or more then they're required to be registered to that persons name/address (whois on the IP will be that person, not their ISP). Not all ISPs actually do this, but they're supposed to. I have no idea about the rules for other Internet registries though (and as mentioned above this is only applicable when the end user has 8 or more IPs).  ZX81  ^talk 17:15, 7 June 2011 (UTC)

Covering another part of the possible answers. It is easy for some employees of your Internet Service Provider to find out your home address if they see your IP address. If an employee of, say, Comcast, a cable TV and Internet service provider in the United States, sees that your IP address is 68.87.98.101, he or she could look up that IP address in the Comcast database and see if it's a Comcast-provided IP address. If it is, then he or she could look up whose home address is registered to that number. Comet Tuttle (talk) 18:53, 7 June 2011 (UTC)

Skyhook Wireless, like many similar services, sells access to a comprehensive and up-to-date database of IP addresses, WiFi hot-spot MAC addresses, and other network-location -to- geographic-position mapping tools. The methods that they use to acquire such data could be considered somewhat malicious and invasive: driving around and sniffing for public metadata about private wireless networks, but these techniques are neither illegal nor technically difficult. Nimur (talk) 20:37, 7 June 2011 (UTC)

Do any of you know any tools I could use to do this? I'm just curious, that's all. B-Machine (talk) 17:27, 8 June 2011 (UTC)

As pointed out above, whois and tracert are the common tools to approximate geolocation of an IP address. -- kainaw™ 17:33, 8 June 2011 (UTC)

June 8

cable provides telephone service

I get my internet and TV from a cable TV provider; and my telephone from the telephone company. Today the cable company called me up offering me a package including telephone service, without changing my telephone number. If the cable TV company provides my telephone service, does it go over their lines or does it continue to go over the telephone company lines? Bubba73 ^{You talkin' to me?} 02:19, 8 June 2011 (UTC)

It would be over their lines. Basically, it's a VOIP service like Vonage and others sell, only it's from your cable company. Grandmartin11 (talk) 03:02, 8 June 2011 (UTC)

Thanks. So when my internet connection is down, the phone will be down too, right? Bubba73 ^{You talkin' to me?} 04:33, 8 June 2011 (UTC)

Maybe. I've had phone service through the local cable monopoly for years, and I believe that occasionally I've lost Internet service but not phone service. Comet Tuttle (talk) 05:38, 8 June 2011 (UTC)

In many jurisdictions anyone providing residential phone service has to meet strict requirements for reliability and availability due to the need for dependable emergency service. For example, in many cases the cable telephone box is required to have a built-in battery sufficient to provide phone service for several hours after a loss of mains power. I would expect cable telephony to be quite reliable as long as the physical connection to the cable company is intact, even if internet service is interrupted; however, I don't have any direct experience with it. Of course, not much can be done if there are physical disruptions in the cable provider's network. Dragons flight (talk) 06:00, 8 June 2011 (UTC)

The real answer is that it could be either, and so the best option would be to ask your cable provider. --Phil Holmes (talk) 15:59, 8 June 2011 (UTC)

My internet over cable is so unreliable that I don't want to trust the phone over it. (The cable repairmen just left here a half hour ago. I have the cable company trouble line on speed-dail - the only one that isn't a family member.) Bubba73 ^{You talkin' to me?} 18:06, 8 June 2011 (UTC)

World Mode G3

After doing a search on this topic I received many link discussing the use of World Mode G3, but no specific definition of what it is or who provides it.99.4.183.96 (talk) 10:36, 8 June 2011 (UTC)

Could you clarify what you're talking about? Googling "World Mode G3" gives me no clue. AndrewWTaylor (talk) 11:01, 8 June 2011 (UTC)

Game programming languages?

Hi, what languages are used for programming games like world of warcraft, runescape, and for ones on consoles (handhelds, etc.)? Could anyone direct me to good sites/books/tutorials with useful info (general and/or code) on said languages? 3879orcs (talk) 14:30, 8 June 2011 (UTC)

A great number of professional games, for all platforms, are written in C++. It's fairly common for scripting inside games to be written in Lua. Both articles contain links to Wikibooks (and C++ to Wikiversity too) tutorials. -- Finlay McWalter ☻ Talk 14:39, 8 June 2011 (UTC)

If you are asking because you want to choose your first programming language, I think Lua is the best choice. It's easy to learn and use, easy to deploy and very fast. You probably want to start with Löve, which has a lot of support for things typically needed in games.

A lot of commercial games are written in C++ but have Lua embedded, and in Manufactoid you can even program a little in Lua to solve levels. (However, I believe it's not strictly necessary, and it assumes that you already know Lua.) Hans Adler 14:51, 8 June 2011 (UTC)

Lua is written in C. C++ is C with classes. http://www.iso-9899.info/wiki/Main_Page /Books ¦ Reisio (talk) 08:07, 10 June 2011 (UTC)

Breaking out of alert() loops

Do any modern browsers offer satisfactory ways to break out of Javascript alert() loops?

Consider the following code:

while(i=1) {
alert("Annoying!");
}

How does one break out of such a page? When I've gotten trapped in loops like that, I've often had to force quit the entire browser, which seems like overkill.

(In general, I find alert() boxes pretty irritating and wish they were easy to disable entirely, or turn into something more innocuous!) --Mr.98 (talk) 17:26, 8 June 2011 (UTC)

Firefox: lets you close the individual tab responsible.

Chrome: the 2nd occurrence has a "prevent this page from creating additional dialogs" checkbox.

Opera: lets you close the individual tab and has a "stop executing scripts on this page" checkbox on the dialog

Konqueror: nothing

-- Finlay McWalter ☻ Talk 17:37, 8 June 2011 (UTC)

Out of curiosity: does Firefox allow you to stop executing scripts at such a time, too? My thought is those pages that have the alert boxes in their close action, which are doubly irritating. The Opera/Chrome options seem like the most favorable of your list. --Mr.98 (talk) 17:43, 8 June 2011 (UTC)

Firefox doesn't offer me such an option (but there's probably an extension somewhere to limit this). Firefox and Opera treat alert dialogs as modal only to the tab; Chrome (at least on Linux) treats them as app-modal, which means the browser becomes effectively unusable until you check that box. Opera's is the best, and most responsive, solution. I also tried Windows Safari on wine on linux (so that's not really very representative) and it behaved as Konqueror. -- Finlay McWalter ☻ Talk 17:51, 8 June 2011 (UTC)

For the most part, Firefox is still designed to be extended for your needs. So, you can install NoScript. Then, you can press ctrl-backslash (if I remember correctly) to turn on/off scripts on the page. -- kainaw™ 17:52, 8 June 2011 (UTC)

But would that work if you are stuck in a loop like that? I ask because many browsers seem to interpret "alert()" as something that takes total focus (so you can't even go to Quit the browser — hence the force quit). --Mr.98 (talk) 18:55, 8 June 2011 (UTC)

More: IE9 and Safari on Windows 7 both block (app-modal), chrome and firefox as above. I also tried a developer build of Opera Mobile on Linux (which has no concept of tabs); it doesn't have the "no more scripts" box, but the back button still works inside the dialog (so does reload, which fixes the loop unless it's called from the onload event). -- Finlay McWalter ☻ Talk 18:13, 8 June 2011 (UTC)

Looking for a Programming language

Hello. In French (the non-programming language :), you negate a verb by surrounding it with 'tags' of sorts; the "opening tag" is always 'ne' but by itself does very little, it just indicates that you are going to begin a negative; the "closing tag" may be any of several (pas, point, jamais, guère, aucun, etc.) and is where the real information is, telling you what sort of a negation you're going for. For example, if I said, « Je n'ai pas faim », it would mean "I am not hungry". If I said, « Je n'ai jamais faim », it would mean "I am never hungry". This distinctly reminds me of a programming language, though I cannot pinpoint which. I am not looking for negation in particular, but rather sort of the opposite of HTML, where the opening tells you only something general but the closing tells you something specific. Does anyone know a programming language like that? Thanks. 72.128.95.0 (talk) 17:34, 8 June 2011 (UTC)

In general, it sounds like anything that contains simple Negation operators, e.g. Boolean NOT, or the "!" used in many C-like languages. (HTML is a markup language, not a programming language, which is why it operates in the way you've described — it is about describing the characteristics of something in a positive way, not a negative way.) --Mr.98 (talk) 17:41, 8 June 2011 (UTC)

The closest thing I know of is reverse Polish notation. Looie496 (talk) 17:47, 8 June 2011 (UTC)

Actually, thinking about it, any stack-oriented programming language more or less fits the description. Looie496 (talk) 17:50, 8 June 2011 (UTC)

MUMPS is a rather distasteful programming language that allows for all kinds of nonsense - such as having a FOR loop that changes behavior based on if you end it with a QUIT or GOTO statement. In general, a programming language should not leave you in the dark about the context of what you are currently viewing while you wait for more information. It should precede what you are viewing with all the information you need. That is why HTML places all options in the open tag and not the closing tag. -- kainaw™ 17:48, 8 June 2011 (UTC)

I think Kainaw is describing what we call a context free grammar; but the way he phrased it, it sounds like he's conflating this with a strict rule about statement-order. As Looie496 has pointed out above, a language can be constructed that is "backwards" (i.e., a valid statement in the language can provide important structural information about a term, even after the term was introduced). Nonetheless, even with this ordering, the language can still be defined using a context free grammar (CFG). For a rigorous language that is useful to machines, it is important to be context-free, so terms may not overlap. But there is no requirement that terms appear in any specific order. Kainaw's suggestion for strict ordering of terms definitely does help humans to understand the language, but there are many reasons why machine languages should have "backwards" orderings. Human languages (including French) may include statements whose terms are not strictly context-free; i.e., whose sub-statements overlap; this is because human language does not follow the rules of a CFG. (Or, more precisely stated: most efforts to define CFGs for most human-languages have been unsuccessful to date. Once this definition-task is accomplished, we should be able to translate French and English into GIMPLE and compile natural statements into machine code - but this appears to be far off on the horizon of Computer Science). Nimur (talk) 18:52, 8 June 2011 (UTC)

The standard language doing everything backward is Forth. --Stephan Schulz (talk) 19:04, 8 June 2011 (UTC)

Quibble: I'm not sure what you mean by "overlap", but context-free grammars are permitted to be ambiguous, which is probably counts as overlapping. Now, if you want to return a single parse (and you don't have any after-parsing precedence rules), you need to supply the grammar with a string that produces an unambiguous parse, but that's a much less strict restriction. CFGs are quite powerful, though the traditional parsing tools tend not to provide the full power, settling for something like LR(k) or LL(k) instead. If you reverse a production in an LR(k) or LL(k) grammar, you might get something that can no longer be parsed as LR(k) or LL(k), but no matter what you flip, it'll never stop being a CFG. Paul (Stansifer) 02:11, 9 June 2011 (UTC)

Install DMG programme on Windows 7

Hi, I'd like to install the demo of this software on my Windows 7 laptop. How would I go about this? Thanks! ╟─TreasuryTag►Acting Returning Officer─╢ 19:18, 8 June 2011 (UTC)

.dmg files are Apple Disk Images. That article lists a few archivers on other platforms that may read them; but if the archive contains a MacOS-X application then it won't work on Windows. -- Finlay McWalter ☻ Talk 19:22, 8 June 2011 (UTC)

That software appears to be mac only. Sorry. --Thekmc (Leave me a message) 20:41, 8 June 2011 (UTC)

You can run Mac OS in a virtual machine. You can get cheap Apple computers secondhand. You could ask the ImageMagick people how they'd do such things. ¦ Reisio (talk) 08:02, 10 June 2011 (UTC)

Best word to describe "Software which is run by an operating system"

Is there one single term for describing the difference between a software running on a microcontroller, and alone responsible for the setting of pinouts, handling interrupts, processor registers, etc., and those that run under an operating system? When I casually say "software", it is usually understood as a desktop application, something running under an OS, but how to specify it? "Application software" is not accurate either, I think, because it's more about the role in regard of the user.

Maybe embedded vs non-embedded? Is there something more accurate? — Preceding unsigned comment added by 87.169.30.70 (talk) 19:32, 8 June 2011 (UTC)

It is common to refer to the programs that manage the hardware as a collection of firmware and device drivers even though there are components of the operating system that are not in either of those categories and there are user applications which allow you to directly access the hardware. -- kainaw™ 19:44, 8 June 2011 (UTC)

The distinction is blurred anyway; even on a very simple microcontroller one frequently runs a primitive micro-executive which handles some very basic things like threading and messaging, and it's very common to end up with a two-stage loader (smartcards run tiny H8-3xx or 8051 microcontrollers, yet still contain things their authors optimistically call "applications"). -- Finlay McWalter ☻ Talk 19:56, 8 June 2011 (UTC)

Yes, I know... what I'm more interested in, how to describe the difference between an application which runs on an operating system, maybe even has a GUI, works around with threads, has memory management and everything nice provided by the OS, etc, and the one which runs on an embedded system and has to rely on interrupts, direct setting and reading of portpins, because there is no OS under it. I cannot find a good word or very short phrase to say it, without describing it at greater length, just like I did here. --87.169.30.70 (talk) 20:46, 8 June 2011 (UTC)

There is no single word to encompass all variants of the type of program in question. "It depends." It depends on the CPU architecture; the operating system that is already resident on the machine; and a slew of other things.

Hosted code, managed code, and user space program all refer to specific aspects of software that is limited in capability (and therefore relies on an operating system, managed code runtime, or hardware abstraction layer). In some contexts, such terms refer to specific program software or hardware features (like an execute-permissions bit mask in a 386 processor). In other contexts, these terms refer to specific tools and environments (like the Microsoft CLR). In general, most CPU hardware has a clear distinction between kernel space and user space; this dividing line is more instructive and more precise than the distinction between "GUI" and "not-GUI," because it specifies exact restrictions on executable capabilities.

If we want to be more precise than this, we need to start specifying exact machines and architectures. Many CPUs do not have user-mode virtual memory protection - so there is no distinction between "operating system" and "application program," except a matter of notational convenience for the programmer. Here are some examples for common platforms you might be familiar with:

• On Unix, applications that are resident in user-space are hosted by the kernel, who may preempt, dynamically schedule, and allocate resources. A UNIX kernel typically also controls other permissions: for example, we can then talk about programs that require super-user permission on the file system to perform certain functions; as opposed to (software) user-space programs.
• On some hardware platforms, especially tightly-knit, syncretic hardware / operating-system combos, file system permissions are managed separately from other resource permissions (such as access to physical memory or peripheral I/O devices). An example of this is VAX/VMS which ran on Digital computers.
• On Java platforms, there is a clear distinction between Java bytecode and JVM internals. Bytecode is inherently sandboxed and may not perform action without permission from the JVM.
• On Windows computers, code compiled with the Driver Development Kit has kernel privileges, while code compiled with the standard Visual Studio SDK does not. Only Microsoft may write Windows kernel code, but you can bluescreen the machine (or otherwise violate implicit, protective security contracts) by writing a kernel-space driver.
• On an IBM computer with a Trusted Platform Module, code may either reside inside or outside the TPM security zone. Code that resides outside might be considered "privileged" and/or "unsecure," whereas code that resides inside must authenticate and present a valid credential and signature to hardware.
• Using ucOS, there's nothing separating your code from the ucOS code. Most microcontroller hardwares do absolutely nothing to stop you from inserting a program-jump, memory-write, or memory-load; the kernel will do nothing to stop this from happening either; and you could disable software preemptive interrupts, denying return of control to the kernel.

As you can see, the term you seek is highly variable depending on the type of computer hardware and operating system you're working with. Nimur (talk) 21:28, 8 June 2011 (UTC)

Thank you. However, in this case not the highest scientific accuracy is needed, just an oversimplified general description, even if not correct in each and every special case. --87.169.30.70 (talk) 21:47, 8 June 2011 (UTC)

Then, I would use "hosted code," and if anyone needs more specific details, you can say that it is hosted by (e.g.) Linux. Nimur (talk) 22:15, 8 June 2011 (UTC)

I would call that "kernel space code", "kernel code", or say that it runs "in kernel land", as opposed to "user space code", "user code", or saying it runs "in user land". --Sean 15:55, 9 June 2011 (UTC)

Microcode runs on CPUs.
Sleigh (talk) 23:32, 10 June 2011 (UTC)

how do most php programmerse program the html part (mac)

in the same way that a couple of years ago people who were web desginers in html used dreamweaver, what would someone today use? especially if they're also working in php? Dreamweaver still or something else? Especially interested on the most common solution on a mac...

note: I do code in php and html, don't need something totally wysiwyg, more interested in what designers are doing these days. thanks.

I can't specifically answer what designers are doing nowadays, but I can say that Dreamweaver can code in php, too, and many jobs for webdesigners specify ability to use Dreamweaver as a prerequisite. --KägeTorä - (影虎) (TALK) 00:18, 9 June 2011 (UTC)

Most people I know who call themselves "web designers" use Dreamweaver, but do not use PHP. They do use some simple drop-in PHP modules (or even Ruby modules). But, they couldn't write a simple "Hello World" script. After they do the design, it is sent off to a programmer to add functionality. When I do a website myself, I do everything in Kate. It is far easier to type the HTML/PHP code by hand than to spend days searching the web to try and figure out how to make Dreamweaver do what I want it to do and not what it thinks I should do. -- kainaw™ 12:06, 9 June 2011 (UTC)

There are a number of smaller, but widely used IDEs such as those by JetBrains. Also have a look at List of PHP editors and List of PHP editors.Smallman12q (talk) 11:17, 10 June 2011 (UTC)

Using my choice of icon for each desktop shortcut, folder, or URL

I use WinXP in classic mode. All the folders, shortcuts and URLs on the desktop look the same as each other and just have different text below them. Is it possible to use a particular icon for each thing? For example folderA could be a blue square, but folderB is a red circle. Thanks. 92.24.128.171 (talk) 21:09, 8 June 2011 (UTC)

Yes - right-click the icon and select "Properties". Under the "Customize" tab you'll see a "Change icon" button. You might need to Google around to find the exact icons you want. AndrewWTaylor (talk) 22:39, 8 June 2011 (UTC)

Thanks. I'd be gratful for any information about how to make desktop icons. Such as what size they need to be, what format, etc. 2.101.15.113 (talk) 10:47, 9 June 2011 (UTC)

The basic format is ICO (file format), but icons can also be embedded in other files (e..g .exe and .dll). There's some information at Computer icon#Icon creation. Here's a tutorial on how to make one using GIMP. AndrewWTaylor (talk) 10:52, 9 June 2011 (UTC)

I'd recommend IcoFX, from personal experience. You can easily import an existing image and turn it into an icon. IcoFX has a lot of functionality, and yet is easy to use. Rocketshiporion♫ 23:56, 14 June 2011 (UTC)

what might a beginner without formal training learn wrong about web design?

for example, I would consider it a mistake to edit html files on the server directly instead of using any kind of version control. what else might a beginner without formal training learn wrong about web design? (I mean something not obvious, bad habits etc, that migth still "work"). Thanks. --188.29.130.143 (talk) 21:50, 8 June 2011 (UTC)

Oh, where to begin...

• not testing on all reasonable browsers (even the old versions of IE we all wish didn't exist)
• not taking adequate care that the site prints out nicely (heck, almost no-one does that, not just beginners)
• not taking adequate care that users with disabilities can use the site successfully (this can be hard, as there are lots of different disabilities and access-methods).
• writing the whole thing in flash, because menus that go "whoosh" is kewl.
• writing non-validating markup that happens to work on the one machine they tested it with
• doing stuff in javascript that can be done better with CSS hover
• not testing for cases where the user has javascript, java, or flash disabled
• not worrying about users who don't have a the same screen size, or same fonts installed, as they do
• serving old versions if IE transparent PNGs
• not thinking about, supporting, and testing for mobile platforms (where there's a different browser, tiny screen, different fonts, and a touch interface).

... but that's mostly true for lots of people who claim to be experts too. -- Finlay McWalter ☻ Talk 22:05, 8 June 2011 (UTC)

The absolutely #1 mistake that everyone from angsty tweens to major corporations make, however, isn't really a design mistake per se. It's spending a bunch of effort on the website technology and appearance (fancy graphics, fading menus) and not nearly enough on providing decent content that people will want to read, view, listen to, or play. -- Finlay McWalter ☻ Talk 22:16, 8 June 2011 (UTC)

To make it clearer... The most common problem on the web is that websites tend to follow the design before content model instead of the content before design model. So, you get a pretty website that has no content. Nobody wants to waste time on a website without content regardless of how pretty it is. However, people will spend hours on an ugly website if it has the content they want. -- kainaw™ 12:11, 9 June 2011 (UTC)

Two problems, in three words, with the first often leading to the second...

1. Clutter

2. Slow performance HiLo48 (talk) 22:32, 8 June 2011 (UTC)

Much of this is covered in Nielsen's Alertbox newsletters http://www.useit.com/alertbox/ -- SGBailey (talk) 08:49, 9 June 2011 (UTC)

I'd say the #1 biggest, most enduring, most damaging mistake people make is not including a proper doctype declaration properly. You might think merely validating would cover this, or that using popular HTML cleaning software would ensure against it, but you'd be wrong. ¦ Reisio (talk) 07:29, 10 June 2011 (UTC)

could you explain what makes it proper, or give a proper example? You also seem to be saying that it's not just how it's written, but how it's included... --188.29.60.182 (talk) 09:24, 10 June 2011 (UTC)

Indeed, as the second link I gave details, how it’s included is quite important. The simplest proper example is <!doctype html> with absolutely no character before the first <, and this includes invisible characters some editors insert by default. ¦ Reisio (talk) 23:07, 10 June 2011 (UTC)

I agree that the lack of content and overemphasis on graphics is the biggest problem for web designers. To put it simply "less is more", when it comes to graphics. Any NEVER start playing an animation/movie when somebody enters a web site. You can provide them with a play button, then they can decide if they want to grind their system to a halt to play your stupid animation. Just today, I tried to find out how to change the oil in my lawnmower, but the manufacturer's web site was only interested in providing animated ads for their products, and not in providing any useful info. StuRat (talk) 07:36, 11 June 2011 (UTC)

To add to my previous comment:

1) Anything which can be represented as simple text should be.

2) Anything which can't be represented by text alone, but can as a still picture, should be. Present a thumbnail initially, and let users click on it to enlarge.

3) Anything which can only be represented by animation, should be, but should not play automatically. An example might be a cross section of a running internal combustion engine, but only when trying to explain it's operation. Don't show such an animation when trying to sell cars, for example. StuRat (talk) 07:47, 11 June 2011 (UTC)

USB Devices No Longer Working (Vista)

Resolved

Hello all,

I have a problem. Since removing my USB devices earlier today (by just taking them out, and not deactivating them through the icon in the taskbar, which incidentally wasn't there anyway), none of them works. I have a mouse, an external HDD, a headset, and also a Flash drive, and three USB ports. Of course, I only ever have three of them connected at one time. Power has not been a problem, so we can rule that out. I believe the problem was caused by my taking the USB cables out before deactivating them to 'safely remove' them. Therefore, is there any way to fix this?

I have rebooted several times, to no avail. The USB external HDD is not even detected before boot-up (I tried to boot into my Linux partition on the drive to see if it made a difference). I am unable to perform a system restore, as there are no restore points (I need to fix that, too).

Anyway, as I know the specific cause of the problem, maybe someone can give me some help specifically to that?

Cheers, Kage. --KägeTorä - (影虎) (TALK) 22:23, 8 June 2011 (UTC)

EDIT - No worries. I did what any self-respecting genius would do and put it on the road outside, then ran over it several times in my neighbour's jeep. Then I took the battery out and put it back in and everything works as normal now. Cheers. --KägeTorä - (影虎) (TALK) 22:37, 8 June 2011 (UTC)

June 9

Laptop wont recognize FAT32 Flash Drives

Hi All,

Weird, it just started happening recently. Anytime any FAT32 flash drive is inserted I am told it needs to be formatted before use (tried with 4 drives from different people). Running chkdsk reveals no problem found, output below:

C:\>chkdsk g:
The type of the file system is FAT32.
Volume Serial Number is E68E-CB52
Windows is verifying files and folders...
File and folder verification is complete.
Windows has checked the file system and found no p
    8,002,240 KB total disk space.
            8 KB in 2 hidden files.
        1,012 KB in 230 folders.
    3,664,296 KB in 1,559 files.
    4,336,920 KB are available.

        4,096 bytes in each allocation unit.
    2,000,560 total allocation units on disk.
    1,084,230 allocation units available on disk.

trying to change path to g: using cmd informs me I need to install a driver for it? Anyone have a clue what's going on and how to fix it? TIA PrinzPH (talk) 01:20, 9 June 2011 (UTC)

ADDENDUM: I am running Vista, 64bit PrinzPH (talk) 01:22, 9 June 2011 (UTC)

Did you by any chance install an update recently for Vista, after which this problem occurred ? StuRat (talk) 07:26, 11 June 2011 (UTC)

Regarding archiving on CD/DVDs

Hello! I have a couple questions before backing up a bunch on files on CDs and DVDs.

1. In the past, I've used a Sharpie to label the disk, but after some Googling, some people say that's okay and others that it's a horrible idea. Do I need to buy a special marker for this purpose?
2. DVDs have a capacity of 4.7GB, but when I burn a lot of files through Windows Vista (with the "Master burn option," which yields a UDF-formatted disk), it caps me off at 4.37GB. I know that some data is reserved for filesystem information on the DVD, but over 300MB for 4.37GB of data? What accounts for this?
3. Using Windows Vista's default application for burning DVDs (whatever it calls when you right-click a selection of files and send them to your DVD-Recording drive), I've had several problems in the past and present. Just now I had it set up to burn about 4.36GB, then it told me the burn was successful but when I took the DVD out, the optical side looked only half-written, and Explorer freezes when it tries to read it. Can anyone suggest CD/DVD-writing software that is better suited for backups and more reliable?

Thanks for your help.--el Aprel (^facta-_facienda) 05:45, 9 June 2011 (UTC)

Optical media capacity is always quoted in decimal prefix whereas Windows always uses binary prefix so you aren't missing any capacity. Also with DVDs the metal layer is under a layer of polycarbonate so even in the unlikely event the solvents or ink can cause damage to the metal, it's going to have to bleed a long way before it can. In that vein, I would completely ignore the comments of anyone who doesn't make a distinction between CDs and DVDs Nil Einne (talk) 13:01, 9 June 2011 (UTC)

Infrarecorder is pretty good, I've certainly never had any problems with it. As for the Sharpie thing, I'd agree with Einee about the solvents, but again I've never had nay problems using Sharpies on CD/DVDs. Touch wood. Jackacon (talk) 14:26, 9 June 2011 (UTC) Forgot to include a link! — Preceding unsigned comment added by Jackacon (talk • contribs) 14:27, 9 June 2011 (UTC)

Of course a Sharpie will write on a disc, but the point is that you want to be able to read the disc back in the future. I've read that alcohol in standard markers is bad for discs, but I don't know if that is true. However, Sharpie and others make markers designed for that (e.g. the Sharpie CD/DVD Marker), so might as well use them. Remember - it is the top surface that you most don't want to damage. Bubba73 ^{You talkin' to me?} 16:05, 9 June 2011 (UTC)

No, as I pointed out this isn't true for DVDs. Damaging the bottom surface is in fact clearly worse. (Since if you do something that penetrates and damages the metal you're screwed either side, but if you do something which damages the polycarbonate enough to prevent the laser properly reading the pits this is only a problem on the bottom. Obviously we're not talking about double sided DVDs where top and bottom are meaningless.) Nil Einne (talk) 17:36, 9 June 2011 (UTC)

Y'mean "Damaging the top surface" there? ¦ Reisio (talk) 07:08, 10 June 2011 (UTC)

Imgburn AvrillirvA (talk) 16:36, 9 June 2011 (UTC)

Nil Einne, regarding your point about the Sharpie, you said DVDs have a metallic layer that would likely block any solvents from breaking through to the bottom layer, but is the same true for CDs? If not, what would you suggest for labeling them instead? (I'm guessing CDs don't have the same protection, because I recently bought a pack of Verbatims, wrote in permanent marker on the top, and could see through to the permanent marker when holding it upside-down up to the light. It still plays fine [for now], but it makes me a bit concerned.)--el Aprel (^facta-_facienda) 20:11, 9 June 2011 (UTC)

Also I suggest using high-quality discs. I use Taiyo Yuden/JVC and I think Mistui are also good. Bubba73 ^{You talkin' to me?} 20:20, 9 June 2011 (UTC)

DVDs have a polycarbonate layer on the bottom, then a reflective recording layer, then another polycarbonate layer, then the label on top. CDs have a polycarbonate layer on the bottom, then a reflective recording layer, then the label is directly on top of the recording layer. It's the lack of an upper polycarbonate layer that's a problem when writing on CD labels. --Carnildo (talk) 00:55, 10 June 2011 (UTC)

If you intend to archive valuable data on optical media, you may want to consider using scratch-resistant Imation Forcefield CD-Rs and Forcefield DVD±Rs. I've found them to be quite durable. Rocketshiporion♫ 00:04, 15 June 2011 (UTC)

Knowledge Management System of an organization using RDF in Semantic Web

Dear Sir/Mam,

How can I design a Knowledge Management System of an organization like an University using RDF in the aspect of Semantic Web. — Preceding unsigned comment added by Subhakoley (talk • contribs) 06:08, 9 June 2011 (UTC)

Why THAT password?

According to this article the passwords

seinfeld, password, winner, 123456, purple, sweeps, contest, princess, maggie, 9452, peanut, shadow, ginger, michael, buster, sunshine, tigger, cookie, george, summer, taylor, bosco, abc123, ashley, bailey are often used.

What ist the "reason" for choosing 9452 ??? Greetings from France! Grey Geezer 13:21, 9 June 2011 (UTC) — Preceding unsigned comment added by Grey Geezer (talk • contribs)

http://www.irs.gov/pub/irs-pdf/f9452.pdf ? everybody pays taxes... but that bearded "good" guy in human target... --Homer Landskirty (talk) 13:30, 9 June 2011 (UTC)

Thank's for the answer. Love "the bearded guy", too. Never take cheques either ... ;-) Grey Geezer 13:33, 9 June 2011 (UTC) — Preceding unsigned comment added by Grey Geezer (talk • contribs)

It could be related to whatever the people were signing up for. For example, the password "kasparov" might be common on a chess site, but not elsewhere. —Tommyjb Talk! (13:37, 9 June 2011)

The password's owner might have been born on 1994/5/2, have a child born then, or have some other reason for choosing that date. CS Miller (talk) 13:48, 9 June 2011 (UTC)

The original article here states that these were apparently user/pass values for a competition or sweepstakes of some kind. There must have been something on the signup form that caused the users to lean towards 9452. -- kainaw™ 13:54, 9 June 2011 (UTC)

Meaning they saw something (permanently on that page), so they could remember it easily later? Grey Geezer 14:10, 9 June 2011 (UTC) — Preceding unsigned comment added by Grey Geezer (talk • contribs)

how does NAT traversal work? (or busting through a firewall)

hi guys. how does NAT traversla work (or busting through a firewall). someone said that they thought hte reason skype really succeeded was because it worked from behind corporate firewalls (even though it was a peer to peer technology) -- how is that? what did it do to get through, or did all traffice from behind two firewalls have to go through skype's servers, after being initiated from the skype client? that seems like an awful lot of traffic to go through skype's servers... thanks for any responses or insight you might have on this point of curiosity. 87.194.221.239 (talk) 17:02, 9 June 2011 (UTC)

p2p networks work with NAT as long as a sufficient number of nodes in the network are not behind a NAT (or have port forwarding set up, so the NAT isn't blocking inbound connections to them). The networks use these supernodes which proxy connections from one NATted client to another. This post discusses Skype and supernodes, and speculates that, rather than relying on ordinary users running supernodes (as p2p file sharing networks do) Skype Inc. actually runs it's own supernodes as well. -- Finlay McWalter ☻ Talk 18:52, 9 June 2011 (UTC)

Most P2P clients including Skype using a variety of techniques including what FM has mentioned but also other things. Start with NAT traversal and read linked articles like UDP hole punching and TCP hole punching Nil Einne (talk) 21:25, 9 June 2011 (UTC)

If the NAT device supports Universal Plug and Play (UPnP), then Skype can ask for its standard to be forwarded to it. NB UPnP is not related to the similarly named plug and play for hardware detection in PCs. CS Miller (talk) 22:47, 9 June 2011 (UTC)

Generally, firewall busting exploits an assumption in stateful NAT devices, that when one device communicates to an external address from a particular port, any response received from that address, to the particular port on the NAT device is most likely intended for receipt by the original device. For example, consider a topology where there is a NAT device with the public address 4.4.4.4, and the private address 10.0.0.1, and everything behind the firewall is in the 10.0.0.1/24 network. When a device at, say, 10.0.0.16 transmits data from it's own port 500 through the NAT device to the external address 8.8.8.8, the NAT device will assume for some period of time thereafter that data received on the public interface addressed to port 500 from 8.8.8.8 is, in fact, intended for receipt by the internal device 10.0.0.16. You might also review STUN for more information. 24.177.120.138 (talk) 00:00, 12 June 2011 (UTC)

NSA and Cryptography

Reading a bit about modern cryptography, I am wondering that as far as we know, the popular cryptographic algorithms (like RSA with primes large enough or AES) are secure. They have been intensely scrutinized and studied. But we don't know much about the resources at NSA. Could it be the case that we (the public) think that an algorithm is secure but in reality the NSA might have broken it? How sure can we be that our messages are secure even from someone with (practically) infinite resources like the NSA (compared to an average user like me)? I mean they do have the best minds. What if they have some new cryptanalysis techniques we don't know about (like when differential cryptanalysis was discovered) or a massive massive computer that can bruteforce something "infeasible"? This is just an opinion question and just wondering what the computing community here thinks? -Looking for Wisdom and Insight! (talk) 17:30, 9 June 2011 (UTC)

1. publicly known approaches take very long time on normal computers... even if overclocked and MP... :-)
2. there might be better algorithms and/or better hardware (somewhen)... but: since research at universities is public and leading, it would be very funny, if the government would have a "secret university", that is so much better than regular universities (why would they pay for all those professors that r not as good as those at that "secret university")...
3. i think the government uses different approaches to prevent severe crimes (like undercover cops)... and really bad criminals (like spies) possibly use unbreakable cryptography (like one time pad)... --Homer Landskirty (talk) 18:28, 9 June 2011 (UTC) e. g.: in f. rep. germany the military police repeatedly questions their employees about their friends in school... --Homer Landskirty (talk) 18:32, 9 June 2011 (UTC)

For things like passwords on common home computers, the government certainly has the resources to crack any normal person's password in a matter of hours (if not minutes). As has been shown multiple times recently, anyone with about $1,000 in hardware and some open-source free software can crack an 8-character password in far less than a day. The key is a distributed brute-force attack. Assume that I tell you it will take 1,000 years to crack some encryption with a brute force attack. So, you distribute the attack among 1,000 computers and do it in a single year. Or, you distribute it among 10,000 computers and crack it in 1/10th of a year. So, consider the NSA. What if they have a warehouse with 1,000,000 computers set up to do a distributed brute force attack? Well, we wouldn't know about it. But, here is where it gets a bit scary... Do we know ANYONE who has over 1,000,000 computers waiting to do his or her dirty work? Yes. Every day, thousands of people ignorantly (and stupidly) install malware on their home computers. That malware puts the computer under the control of a criminal who has the ability to use the computer for anything he or she likes. It could be something simple like spamming the world with ads for fake viagra. It could be something like cracking Sony passwords (assuming they've found a database where Sony actually encrypted the passwords to begin with). The real threat to the common person is not the NSA, it is the criminals who are behind the malware programs that make so much money that they can blatantly run ads on television that tell people to install the malware so your computer will run faster. -- kainaw™ 18:37, 9 June 2011 (UTC)

i would guess a brute force attack against a 2048bit PGP key takes much more than 1000 years on normal hardware (even on a farm of 1 million boxes)... http://www.rsa.com/rsalabs/node.asp?id=2103 --Homer Landskirty (talk) 19:14, 9 June 2011 (UTC)

Notice that those contests are no longer active. I believe it is because they don't want people to work on using GPUs to crack them. When they quote "it takes a computer XXX years to crack this", they mean "a standard single-core single-processor Intel/AMD computer." By comparison, a modern single CPU can brute force about 10 million crack attempts per second. A modern single GPU can brute force about 3.3 billion crack attempts per second. Consider a computer with a 4-GPU graphics card. With simple distribution, you can achieve 12 billion crack attempts for second. So, the GPU clearly outperforms the CPU in this case. Further, claims of 1,000 years almost always ignore improvement in the hardware. Assume that we are using a GPU and the GPU doubles in speed of brute force attack every 10 years. So, after 10 years, we aren't looking at 990 years. We are looking at under 500 years. Another 10 years and we are looking at under 250 years. Another 10 years and we are looking at under 125 years. In all reality, that 1,000 years will take around 50 years. -- kainaw™ 19:27, 9 June 2011 (UTC)

What is a "crack attempt"? Attacks on RSA use the general number field sieve or other complex multistage factoring algorithms, not trial division. The idea that GPUs are enough faster than CPUs to make RSA suddenly insecure is ridiculous. Ditto AES or Triple DES or any other widely used cipher primitive. It is physically impossible for speeds to keep doubling as you imagine. I don't know why RSA Inc. decided to stop giving money away, but my guess is it was no longer generating enough positive PR to justify the expense. Stop spreading FUD on the reference desk, please. -- BenRG (talk) 21:45, 9 June 2011 (UTC)

Let's look at what we do know (before we get onto the fast amount of stuff we don't):

• We know that NSA (and their hand-in-glove cousins at GCHQ) are major employers of mathematicians (in this story at Government Executive magazine they're called "probably the largest employer of mathematicians in the United States"). We really don't know what they do, but they don't publish in ordinary mathematical journals. We know Cliff Cocks invented asymmetric cryptography ages before Diffie and Hellman, and the design of DES' S-boxes is strong circumstantial evidence that NSA was aware of differential cryptanalysis long before Biham and Shamir. We know they design government ciphers for military work and for non-secret applications (like Skipjack and Rambutan). Beyond all that we have no idea what all those people are doing.
• We know NSA has huge datacentres (e.g.). Some of that is plainly for bulk cryptanalysis, some clearly for HARVEST/ECHELON/ThinThread. But we don't know, and there's no way of finding out what, and in what proportion. We know they're very comfortable with custom ASICs as well as general-purpose computing.
• We know from the history of GCCS (Bletchley) and SIS (Friedman et al) that these guys revel in sneaky way of breaking into things; of exploiting non-random characteristics of things, human biases and statistical weaknesses. So, like any decent cryptanalyst, they're lazy - they'll do whatever they can to avoid or minimise brute force. Attacking RSA or AES head-on is the last thing they'll try.

So we don't know (and your guess is as good as anyones):

• ... if there are broad technical weaknesses in modern ciphers like AES which NSA and GCHQ are aware of and the public community isn't. The public community doesn't think so, but is probably outnumbered (and vastly outspent) by the secret community. And the "lots of clever people have studied it" idea shouldn't be that reassuring - lots of clever people tried to solve Fermat's last theorem, but their failure doesn't imply that it was impossible. One notable thing is that, as far as we can tell, the ciphers that NSA and GCHQ makes for their own government "customers" to use appear all to be LFSRs. Maybe they've got good evidence that LFSRs are more secure, or maybe they're institutionally a collection of reactionary Blimps. There's no way of knowing.
• ... what clever japes they have for avoiding having to do hard work. Maybe AES leaks key material in some subtle way. Maybe the PRNG that GPG uses has some massive statistical bias that no-one has noticed. Maybe the probabilistically-prime numbers RSA generates as keys aren't as probably prime as you'd hope.

Combining a modest weakness in an algorithm with a modest weakness in its implementation (from timing analysis to poor random numbers) with massive computing power and rooms full of custom ASICs, they may very well be able to drill into actual implementations of modern cryptosystems. Given the money they take up, they darn well should be able to do so. But we don't know, and can only begin to guess.

For governments (in general) interested in individuals, we know that strong crypto is probably one of the harder points for them to attack. It's much easier to:

• Sneak into your house and put a keylogger into your computer, or plant a virus there. Like Magic Lantern, for example.
• Plant some child porn on your laptop, or some heroin in your bag when you check it at the airport. Then they have plenty of leverage over you.
• Put a gun to your head.

Beyond that, we're into black-helicopter land. I rather doubt it, but maybe Big Brother can:

• read your keyboard's usb traffic using its Van Eck emissions with a SQUID on an NRO satellite in low orbit, or on a drone.
• activate the firmware built into the southbridge of your computer, to inject polymorphic surveillance code into your PC (there was much of a stooshie a few years ago about such "Chinese Spy Chips" allegedly existing)

So really we've no way of knowing what NSA is doing, or what it can do. Electronic communications security is absolutely their home ground and betting that anyone (bar another such agency, and even then) can best them there is unwise. Realising that is one thing that kept Osama alive as long as he lasted. But as Kainaw points out, the NSA really aren't your problem, and if they are, no amount of crypto is really going to help. -- Finlay McWalter ☻ Talk 20:01, 9 June 2011 (UTC)

To Finlay McWalter's list of bullet points, I can only contribute a link to this xkcd cartoon. Comet Tuttle (talk) 17:03, 10 June 2011 (UTC)

Just on your final point (in a very nice post), I would point out something similar: the problems with the NSA are not that they can potentially crack crypto. The problem is that they can monitor all electronic communication without (apparently) any legal checks.

On the "secret university" mentioned above, we do indeed have plenty of these in the US; they are called the national laboratory system. The US spends gobs of money on secret R&D every year. Much of it has evolved to be, essentially, a "secret university" system in many ways, including classified journals and other "black" versions of what are traditionally academic structures. --Mr.98 (talk) 20:51, 9 June 2011 (UTC)

Let me add some anti-fear here:

• Differential cryptanalysis was discovered independently by IBM and NSA researchers. The IBM researchers agreed to keep it secret in the national interest. I don't think that would happen today; the community now is much more open (against security by obscurity) and much more international. It's also much more active; there's no reason to think the NSA has the multi-year lead now that it did then.
• In addition to breaking foreign crypto, the NSA is supposed to approve unbreakable crypto for use by US agencies. They currently approve AES-128 for secret information and AES-192 and AES-256 for top secret. I have enough confidence in the system to believe that any NSA executive who would choose to do that, while knowing that the NSA can break AES, would be fired and replaced by someone more competent.
• Skipjack, developed by the NSA and later declassified, was successfully attacked (though not completely broken) by academic cryptographers, which suggests that the NSA's cryptographers aren't all that. Could the vulnerabilities have been deliberate? Sure, but that would have been foolish given that the NSA would have had access to the keys anyway. Vulnerabilities would only aid the enemy.
• If the NSA does have the ability to break AES or RSA, they won't use it against the likes of you. Unless you're the despotic ruler of a country with nuclear capability, nothing they might learn about you would be important enough to justify the risk of leaking a secret that valuable.

-- BenRG (talk) 01:44, 10 June 2011 (UTC)

.. And with all those people working for these organisations, there's probably a good chance that at least one of them might pass the information to Wikileaks or similar. (Cf Moon landing conspiracy theories - the thousands of people involved can't all be keeping the secret.) AndrewWTaylor (talk) 12:32, 10 June 2011 (UTC)

"Vulnerabilities would only aid the enemy." Not so — they're perfect for giving false information out and having it appear to be legitimate. There are all sorts of ways in which apparent vulnerabilities have been used in the past in order to play a double game.

And re: Wikileaks and conspiracies — there have been a few instances of whistleblowers (e.g. the Thomas Drake case, currently in progress), but if anyone is going to figure out who gave NSA information away, it's the NSA. (Do you think people at the NSA trust Wikileaks to keep them anonymous or use the data in the most efficient manner possible? I doubt it. I suspect they'd go to Congress, not Wikileaks.) As far as large scale "real" conspiracies go, the NSA so far pretty much takes the modern cake with things like ThinThread. Anyway, I wouldn't expect leaks unless there were people in the organization who knew what was being done on the very top levels (compartmentalization means that only a handful of your giant organization has to know the big picture) and that those people at the high level thought there was really compelling reason to destroy their own handiwork. It does exist (see Drake, again) but it's a relatively rare constellation of phenomena — the people who get into high levels of these kinds of organizations are people who have demonstrated their loyalty and their acceptance of the overall purpose of these organizations. The Drake cases are the exceptions, not the rule, because it's rare for high level people to torpedo their entire careers (and possibly get jail time) because of abstract values. The big State Dept. Wikileak was all stuff of relatively middling classification (secret but not top secret) that was accessible through a non-compartmentalized database. I think we can presume that the NSA has more sophisticated ways of divvying up its information, given that most of it is much higher classification.

Even with the moon landings, one could imagine a technical setup in which only a few dozen people out of the thousands knew it was a cockup. In that case it would be admittedly hard, since doing the cockup would require almost as much technical work as the legitimate one. There are definitely different ranges as to what can be kept a secret and what cannot. But it's worth remembering that the Manhattan Project was kept pretty secret (in most respects) despite having 130,000 employees. The reason was simple: of those 130,000 people, only a few hundred knew what they were building (most were entirely ignorant of the ultimate aims of their research), and those few hundred were generally screened for their loyalty, watched by intelligence agents, and even physically isolated from the rest of the world. Now, the Manhattan Project also demonstrates the problems of this: a few of those hundred definitely "leaked" the information, and got through the security nets one way or another. But they did so only to an enemy power — not, say, to the press. So even having a "leaker" doesn't mean the secrets become "public" — it just means that someone unauthorized knows about them. But that person might keep them as secret (or more secret) than you might. (Also, I don't think the moon landings were a conspiracy, obviously. There are lots of better reasons to doubt it was one than the fact that it would have been hard to keep secret.) --Mr.98 (talk) 14:13, 10 June 2011 (UTC)

As the original poster wrote, this is sort of an opinion question, or, more accurately, it's only answerable with guesses, because all the good sources we could research and cite are of course classified. As a counterweight to some of the above, I would caution against a faith-based belief that the NSA can crack or work around AES in some sneaky way — there is apparently a very widespread belief outside of the US that its CIA is close to omniscient and omnipotent, despite ample evidence to the contrary. Comet Tuttle (talk) 17:13, 10 June 2011 (UTC)

Though the ability to predict/affect world outcomes is different than having sneaky technical solutions. American intelligence has long been good at sneaky technical things. It's the sort of thing that throwing lots of money and R&D actually can affect, unlike many of the CIA's issues. If I were to believe in any agency having miraculous powers over their area of expertise, it would be the NSA. I have considerably less faith that the CIA, DOE, FBI, etc., have such miracles at hand. The NSA and the crypto establishment have long been good at keeping secrets, long been good at compromising seemingly uncompromisable systems, and so on. I'm not sure there's any good reason to believe that the situation today is so much different than, say, decades past. The only real variable that has changed is that the public crypto community has gotten bigger, which does change the dynamic a bit, but it is still dwarfed by the secret crypto community. --Mr.98 (talk) 18:31, 10 June 2011 (UTC)

Wouldnt it make sense for western intelligence to have some top-secret dedicated super-computers somewhere already set up to crack terrorists encoded communications, like the Colossus computers in the past? 2.97.219.191 (talk) 21:31, 10 June 2011 (UTC)

Of course! But the problem with these modern algorithms under discussion is that they are theoretically uncrackable even with dedicated supercomputers, at least within a reasonable timeframe. That's what the question is asking: could there a special way around that limitation, or might they have machines that defy this fact? The answer is: maybe there are vulnerabilities the public world doesn't know about (seems possible, but as others have pointed out, the very openness of these algorithms makes them easier to theoretically evaluate), or maybe (more speculative) there are machines in use whose mode of operation either is unknown to the public domain or is only a nascent technology in the public domain (i.e. quantum computers). The latter seems like a much more speculative thing than the former (but not totally out of consideration — the history of US secret R&D is littered with instances of the classified tech world being 10 years or so ahead of the public domain). The NSA certainly does have top-secret dedicated supercomputers for this sort of thing; the question is whether they would help you in cracking something that is made with modern, "military-grade" cryptography (which is quite easily available at this point in time). --Mr.98 (talk) 22:45, 10 June 2011 (UTC)

Keep in mind that AES and the rest of Suite B isn't approved for all classified information. There's Suite A for the most super-duper secret stuff. It's entirely plausible (IMO), that only Suite A contains the algorithms that even the NSA isn't able to crack. 24.177.120.138 (talk) 17:29, 11 June 2011 (UTC)

Those articles are very interesting (I hope you won't mind that I fixed your red link, although someone should probably add a redirect for "Suite A"). On the NSA's website referenced by those articles, NSA does approve AES for TOP SECRET information, even though it's outside of Suite A (presumably): "AES with 256-bit keys, Elliptic Curve Public Key Cryptography using the 384-bit prime modulus elliptic curve as specified in FIPS PUB 186-3 and SHA-384 are required to protect classified information at the TOP SECRET level."--el Aprel (^facta-_facienda) 19:24, 11 June 2011 (UTC)

Yeah, I got the redirect created now. I've always been confused about the purpose of Suite A, given that Suite B is good enough for all of the known levels of classification. My best guess is that it's a tell that there exists levels of classification beyond top secret, which are, themselves, classified. (Cue the black helicopters.) 24.177.120.138 (talk) 23:51, 11 June 2011 (UTC)

Actually, the cynic in me has another explanation for Suite A. Parts of ECC are patent encumbered, which effectively forces implementers of Suite B algorithms for TOP SECRET data to pay royalties or risk lawsuits. Suite A might be the algorithms not so encumbered, and the secrecy surrounding it might be just to line the pockets of Certicom. IOW, perhaps the explanation is more RICO than UFO. 24.177.120.138 (talk) 00:12, 12 June 2011 (UTC)

Pop-up pictures in Firefox

When you click on an image, sometimes websites provide a larger version which pops up with a white border. Is there any way to make them appear in a full Firefox window instead? For example is there an about:config setting that will do this? Thanks 92.24.185.180 (talk) 21:38, 9 June 2011 (UTC)

This is caused by javascript. You could disable javascript entirely in Firefoxs settings (Tools -> Options -> Content) or use an extension such as NoScript or Adblock Plus to selectively disable the javascript responsible for each individual site. You could probably also write a simple Greasemonkey script to force the images to load normally AvrillirvA (talk) 23:42, 9 June 2011 (UTC)

Also sometimes simply right clicking the image and selecting "open in new tab" will work. AvrillirvA (talk) 10:20, 10 June 2011 (UTC)

Wouldn't that be nice? Stupid lightbox nonsense. ¦ Reisio (talk) 06:59, 10 June 2011 (UTC)

The article linked to by Reisio had this http://www.huddletogether.com/projects/lightbox2/ as a reference, and that is what I was referring to. Is there any more specific advice about disabling it please? 2.97.219.191 (talk) 21:06, 10 June 2011 (UTC)

June 10

Back button

I seem to have lost the ability to go back to last screen previously viewed (alt-left arrow). How can I restore it please? Windows XP and firefox 4.0.1. Kittybrewster ☎ 10:13, 10 June 2011 (UTC)

We need some more information. Firstly, maybe there's a problem with the hardware - run Notepad or some other software that lets you check whether the left arrow key and the alt key each work. Secondly, try clicking the "Back" button in the browser instead of alt-left-arrow, and let us know if that works when the alt-left-arrow key does not. It may be a problem with some other software "hijacking" that particular key combo on your keyboard and not passing it along to Firefox. Thirdly, visit a couple of other websites and experiment with alt-left-arrow to see whether the problem behavior is all around the Web, or just on one website. Comet Tuttle (talk) 16:45, 10 June 2011 (UTC)

It seems that is not working / is greyed out on one one multi-window of firefox, while it is not on another. So it is not hardware. Kittybrewster ☎ 21:18, 10 June 2011 (UTC)

Sorry for the dumb question but are you sure there is anything to go back to in that Firefox window? Nil Einne (talk) 22:50, 10 June 2011 (UTC)

Yes, if it opened a new window or tab, then you can't go back, since there is nothing previous in that window or tab. I've thought that the back button, in such a case, should close the new window or tab and return you to the previous one, but the software coders don't seem to agree with me. StuRat (talk) 07:20, 11 June 2011 (UTC)

Yes I am sure. I can open a new window and go from URL to URL but never back. I can't refresh Ctrl-R either. Kittybrewster ☎

Use 'F5' to refresh. 24.177.120.138 (talk) 17:23, 11 June 2011 (UTC)

It doesn't work. Kittybrewster ☎ 20:13, 11 June 2011 (UTC)

Blogs

Exactly what is a blog and how is one set upPersephonelost (talk) 17:25, 10 June 2011 (UTC)

Blog should do you nicely--Jac16888 ^Talk 17:41, 10 June 2011 (UTC)

Fastest mainstream computer language to learn?

Which mainstream computer language would be the fastest for a beginner to learn? I mean languages which are used by programmers. Thanks 2.97.219.191 (talk) 21:13, 10 June 2011 (UTC)

I don't know programming very well, but from what I hear Python isn't hard. --T H F S W (T · C · E) 22:22, 10 June 2011 (UTC)

You either want something like Scheme, which is designed to facilitate general programming understanding; or Perl, Python, Ruby, or PHP, (& JavaScript), which are the most popular "scripting" languages; or shell scripts (Unix land) or batch files (Windows land), which are somewhat more simple scripting systems usually used only locally. ¦ Reisio (talk) 23:16, 10 June 2011 (UTC)

I would caution about picking Scheme. Although it's designed as a teaching language, it was to teach computer science, rather than just programming. So there are a number of pons asinorums (e.g. immutability, recursion) which may be difficult for a beginning programmer to wrap their head around (but once they do, will make them a better programmer). I would second the recommendation for Python. Although not written explicitly as a teaching language, it was created by Guido van Rossum after his experience working with ABC, which was. Guido took what he learned from ABC, but altered it to be better suited for actual production use, rather than just teaching. But "which programming language is better for X" ventures into religious war territory - your best bet is to find resources or a community (either online or in physical space) which is willing to help you (and I mean you, specifically, at your current level) learn, to answer your questions, and to work through your stumbling blocks, and then go with the language they're using. Once you get the basics down, it's relatively easy for a committed person to switch computer languages. -- 174.31.219.218 (talk) 16:39, 11 June 2011 (UTC)

Python is good. I wouldn't inflict Scheme, or any other Lisp-based language, on my worst enemy. All those fucking parentheses -- there's no excuse for that sort of thing in the modern world. Looie496 (talk) 17:17, 11 June 2011 (UTC)

I would. Once you're good at Lisp, learning any other language is trivial. 24.177.120.138 (talk) 17:22, 11 June 2011 (UTC)

I come from a school that teaches Racket, a dialect of Scheme, to first-year students. It's quite successful. In part, this is because the language is very direct. In Java, you have to use object-orientation to structure your code, and you're pretty much forced to use side-effects to do actual computation. In Scheme, you can learn to program first, and then pick up OO, side-effects, continuations, macros, etc., later. (I wouldn't say that learning any other language is trivial, since declarative languages like Prolog and SQL don't have a lot in common with Lisp or Scheme, but I agree with the folklore that it's faster to learn Scheme then Java than it is to learn Java.)

Incidentally, making a language that works like Scheme but doesn't have parenthesized syntax is a matter of ongoing research. (It's more subtle than it seems, and I will talk about it all day if given the opprotunity.) Whether this is useful or not is a matter of some debate, but I would argue that at least it would force people to find something more substantive to complain about ( : Paul (Stansifer) 18:05, 11 June 2011 (UTC)

In general I think any programmer of modern languages today recoils from these "teaching languages" because they are very alien from the syntax of any modern language. They look ancient, clunky, inflexible, and, frankly, not very powerful. To someone who knows, say, PHP or Java or C, they look hard. I don't know if this is really true — I know these languages have their supporters. But it seems like an awfully ass-backwards way of learning to program, unless you are really shooting for the kind of deep understanding that only CS majors care about (and most of those that I have met don't care much about that, anymore, for better or worse). I do think that it's fair to say that those languages were created for a very different era of computing, and if your goal is to quickly build useful applications, those languages are not going to be very satisfying. --Mr.98 (talk) 17:34, 11 June 2011 (UTC)

LISP, and Scheme in particular, are quite easy to learn (to a useful degree) if you do not have any preconceptions. Only if your mind is already locked into the mostly imperative but allegedly object-oriented framework, and into a C-like syntax, does Scheme look "hard" or "alien". And Scheme is a very useful language for anything that has some algorithmic complexity. That said, for beginners with some previous exposure, I tend recommend Python (which, BTW, has some unexpected similarities with LISP under the hood). --Stephan Schulz (talk) 19:42, 11 June 2011 (UTC)

With something like php you don't even really need to learn it to start using it. There are loads of guides available at php.net and other sites. By looking at example scripts to see how they work and trial and error I manage to write quite complex php scripts without having to learn everything about php language AvrillirvA (talk) 17:55, 11 June 2011 (UTC)

I personally can't stand Python, but you can make good software with it, and there are plenty of jobs for someone that knows it. ¦ Reisio (talk) 20:44, 11 June 2011 (UTC)

Are Rebol or Lua any good? 92.24.181.38 (talk) 23:12, 11 June 2011 (UTC)

Adjusting sound

How is one suppossed to adjust the sound for the music on this page: Happy hardcore? When you put your cursor on the loudspeaker symbol, a scale appears. But when you move your cursor towards it, it disapears. What is the secret? Thanks 2.97.219.191 (talk) 21:25, 10 June 2011 (UTC)

The music is just a typical Ogg Vorbis file using HTML5 audio. Therefore it will depend entirely on which player you use. If you use the Cortado (Java) player (should be selectable via the 'More...' link) there is no volume adjustment so it's a moot point. If you use the native browser support then it obviously depends on your browser. On Chrome 12.0.742.91 there is no problem adjusting volume. On Firefox 3.6.17 and 4.0.1 you can adjust volume when the audio is loading but once it has loaded the behaviour you described happens. I guess it's a bug, try reporting it to the Firefox developers. If you are using a different browser you will have to look for help for that browser. If you are using a plugin for your browser rather then builtin support for HTML5 Ogg Vorbis then you will need to look for help with that plugin. Nil Einne (talk) 22:36, 10 June 2011 (UTC)

Found the Firefox bug is already reported here [1]. If I understand the discussion correctly, it will be properly fixed in Firefox 6 (unless perhaps there is sufficient community demand). Meanwhile the volume control may be hidden sometime in the future. Nil Einne (talk) 22:42, 10 June 2011 (UTC)

Meanwhile, as a workaround, I suggest you use your PC volume control. StuRat (talk) 07:16, 11 June 2011 (UTC)

June 11

Using unmanaged .dll with managed code

I'm currently trying to use functions from an unmanaged .dll that I created from in a Windows Forms application. I spent hours on Google without luck, currently following [2]

[System::Runtime::InteropServicesDllImportAttribute("Harvester.dll")]
   extern "C" int testFunction(int num);

int Form1::test()  { num = testFunction(10);  }

but I get a "'System.EntryPointNotFoundException' Additional information: Unable to find an entry point named 'testFunction' in DLL 'Data_HarvesterDLL.dll'."

Any ideas, or am I barking up the wrong tree so to speak? What should I do at the .dll end to declare/export the functions so that .NET can find them?

I'm not new to C++ but haven't done any programming in some time and haven't done anything of this nature before. Benjamint 09:01, 11 June 2011 (UTC)

I haven't done this specific thing, but in general calling between C and C++ and other languages runs into issues of name mangling. It's particularly the case on Windows, which introduces calling convention into the mangling scheme. If you can call Windows API calls successfully, but not your own, you probably want to declare Form1::test to be _stdcall -- Finlay McWalter ☻ Talk 12:22, 11 June 2011 (UTC)

Oh, I didn't read your code properly. Your function isn't called test, it's called Form1::test, and the C++ compiler will have mangled it (as specified in the name mangling article). I can't immediately find a reference about how the Microsoft C# compiler references external C++ mangled names, but I'd first try using the syntax that C# would if Form1::test was a C# method, which I think is just Form1.test -- Finlay McWalter ☻ Talk 13:20, 11 June 2011 (UTC)

I can't seem to follow your link, so I'm not sure what you're trying to do here. You're trying to call native code from C#, or the other way around? You don't look to be trying to call a function named "test" in "Data_HarvesterDLL.dll," which makes that exception inexplicable. 24.177.120.138 (talk) 17:09, 11 June 2011 (UTC)

Not sure why the link wasn't working, I re-pasted it anyway. I'm trying to use native C++ functions in a managed C++ application, yeah. Spent all last night on it and still haven't got any further. It can't find the entry point 'testFunction' Benjamint 01:31, 12 June 2011 (UTC)

You've got a calling-convention issue, I think. Try specifying the __cdecl calling convention in the DllImport attribute. See also MSDN. 24.177.120.138 (talk) 03:17, 12 June 2011 (UTC)

cdecl didn't work.. and I've implemented __declspec( dllexport ) in the class which along with using "extern "C"" should, according to this solve any name mangling. would it be a problem that I'm implementing a class in the dll: i'm importing "testFunction()" should I be importing "ClassName::testFunction()"? I'm completely at sea on this one, must be missing something obvious. Benjamint 07:14, 12 June 2011 (UTC)

TV rebroadcast

My house has multiple floors and TVs/PCs with tuner cards, but over-the-air reception is poor on the lower floors. The traditional cure is to install a good roof or attic antenna and snake the antenna wires throughout the house to distribute the signal to each device. Aside from all the antenna wires, another problem with this setup is that a directional antenna needs to be rotated for each station, requiring a remote control to control this for each TV, and causing the the problem of different viewers on different TVs wanting different stations at the same time. So, my idea was to have a device near the antenna to boost and rebroadcast the signal right there, so no wires are needed. Some questions:

1) Does such a device exist ?

2) Would such a device be legal under US law, if the rebroadcast signal is low enough ?

3) Rebroadcasting on the same frequency might cause interference with the original signal, such as ghosting, due to a time delay. This might be addressed by rebroadcasting at a TV frequency not used in that area. Is that legal ?

4) It sounds like a separate tuner would be needed for each frequency to be rebroadcast in this way. Or, alternately, perhaps the same remote control device used to turn the antenna could also select the channel to rebroadcast. Does this sound practical ?

5) We'd still have the problem of only one channel being viewable in the household at a time, so I guess the only solution to this would be 2 or more antenna/rebroadcasting units. Does anyone see an alternative to this ?

6) If rebroadcasting in the TV band is illegal, how about at other frequencies ? (Of course, this would require a device at each TV to convert the signal back to the TV band, so not an ideal solution.)

StuRat (talk) 17:01, 11 June 2011 (UTC)

1) Video sender

3) They use a different frequency

-- Finlay McWalter ☻ Talk 17:18, 11 June 2011 (UTC)

1. AFAIK, in the US,not commercially, no.
2. It wouldn't be legal under Part 15, since it effectively would be causing intentional interference, and you'd need a license to use any device sold under a different section.
3. Yes and no. There are conceivably areas of the spectrum where unlicensed devices can broadcast with considerable power, but your TV wouldn't be able to tune them. You could also rebroadcast anywhere in the TV spectrum where you don't cause interference, but, since the DTV switch, most of the TV spectrum is used.
4. Sort-of a moot point.
5. Nope.
6. There are a number of unlicensed bands at the top of UHF into microwave. There are also MURS frequencies in VHF, but there's not enough bandwidth there to effectively rebroadcast. You could get an amateur radio license, which opens up a great deal more of the spectrum to play with.

Taking a different approach, can I suggest that you use something like a HDHomeRun, and then redistribute the signal over standard WiFi as MPEG streams? 24.177.120.138 (talk) 17:19, 11 June 2011 (UTC)

That sounds like something an old TV couldn't handle without lots of hardware for each TV to convert it back into a TV signal. StuRat (talk) 18:51, 11 June 2011 (UTC)

That's true, you effectively need a computer. But that's the case already with broadcast DTV and older TVs. 24.177.120.138 (talk) 23:45, 11 June 2011 (UTC)

Subtitling

I spent part of this morning downloading various subtitle editors until I settled on one (which was actually freeware - I had a few trial versions of others, and this one blew them away!). Anyway, I tested it out, and it seems to be fairly straightforward - I open the vid, do the subtitles for it, then save the subtitle file (in .srt format) in the same directory as the vid - they then play together nicely. However, what happens if I want to send the vid to someone? Must I include the .srt file with the video in order that my friend can view the subtitles? Then - and this is the more important part of my question - what happens with YouTube? How do I upload those subtitles? Or, must I 'burn' them to the vid somehow before uploading? If so, how? Cheers. --KägeTorä - (影虎) (TALK) 18:48, 11 June 2011 (UTC)

I suggest you burn them in, and send that version, along with the non-subtitled version, to provide a choice, regardless of the software they have at their end. (Netflix doesn't do this, so foreign language films viewed on streaming always have subtitles, which can be quite annoying, if you happen to speak that language.) StuRat (talk) 18:54, 11 June 2011 (UTC)

Actually, StuRat, many videos on Netflix now have optional subtitles. I just started noticing this last week. --Mr.98 (talk) 19:05, 11 June 2011 (UTC)

I believe you can upload subtitles for YouTube as SRT files. I've never done it, though, but if you Google "youtube SRT subtitle" you'll find a bunch of people who discuss doing it.

As for distribution, most people aren't accustomed to using subtitle files, and will be confused if you send them along separately. I would burn them in if you want to distribute to people, unless you are willing to do a lot of explaining (and possibly have them download different video viewers). --Mr.98 (talk) 19:07, 11 June 2011 (UTC)

Right - cheers for the very swift replies. How do I burn subtitles into a video, then? Let's say I have a .wmv and an .srt to go with it (I actually do - two test files from earlier). What I want to make is a new .wmv with the subtitles on it, and still have the original .wmv (and of course the .srt for a total of three files). How do I go about doing this? --KägeTorä - (影虎) (TALK) 19:12, 11 June 2011 (UTC)

EDIT - Following your advice, Mr 98, I googled 'youtube srt' and found a video of how to burn an .srt to .avi - thanks a lot. For reference, the video is here. --KägeTorä - (影虎) (TALK) 19:52, 11 June 2011 (UTC)

Referrer spam

My web logs show that I've been getting an increasing amount of referrer spam over the past few months. My logs aren't publicly accessible, and have NOFOLLOW/NOINDEX on them anyway. I've tried Googling for them (and other search engines) and haven't found them indexed anywhere. Yet the referrer spam has kept increasing.

Is there any likely explanation of this other than, "I'm on some referrer spam list and thus they do it all the time anyway, since it doesn't take any of their resources"? The only reason it even bugs me is that it makes shifting through my logs more irritating, since there is all of this obvious junk in them. --Mr.98 (talk) 22:14, 11 June 2011 (UTC)

how much faster would it be if instead of normal SSD, the hard-drive was a little computer behaving like a hard-drive but going off of RAM?

If instead of a normal solid-state hard-drive (flash memory), instead it was a little computer with its own flash memory, but many, many 2 GB ram sticks (like 32 of them) to create 64 gigabytes of ram, and then using it all to behave like an even faster hard-drive, then how much faster would that be than a 64 gb ssd (flash-based) by itself? In other words, does a flash-based media already saturate the hard-drive interfaces, and stuff like that, or would having a 100% ram sticks based buffer inside the hard drive enclosure actually speed it up even further? THanks. 188.28.157.156 (talk) 23:09, 11 June 2011 (UTC)

p.s. the reason I say "little computer" is because I imagine it would have to have its own secondary storage (the flash media), as well as a battery so that it has time to write out to the flash if it loses power suddenly, as well as a processor to direct all that, all attached to a motherboard, etc... basically, "logically" it should be just another hard-drive, whereas in fact, it's a little computer pretending to be a hard-drive while having it all in its ram, and just enough battery to write it all out to ssd when it loses power... Question is: would this, in fact, even be any faster than a flash-based ssd drive by itself? 188.28.157.156 (talk) 23:11, 11 June 2011 (UTC)

That basically sounds like a RAM disk, which the article says is "orders of magnitude faster than other forms of storage media". When it writes the data to the SSD it would obviously only be as fast as the SSD is, and if the entire thing was connected to the computer via Serial ATA then that would also limit the speed AvrillirvA (talk) 23:19, 11 June 2011 (UTC)

Right. If the RAM device is connected to a hard disk interface, though, there might not be much gain. According to our Serial ATA article, the fastest hard disk interface in widespread current use will handle a throughput of about 600 MBytes/sec, which is only moderately faster than the fastest currently available SSD devices. Looie496 (talk) 23:40, 11 June 2011 (UTC)

Note that there's transfer speed (or bandwidth) then there's random access times and latency. While these are generally related, particularly with a RAM (of any type) device, performance generally depends on both and the requirements of the app. SSDs using SATA often have 2-4x the transfer speed of magnetic HDs their biggest advantage tends to come from their very low random access times and latency compared to magnetic hard drives because they don't have to seek. A DRAM based SSD will potentially have better random access and latency although I doubt it will make any difference over a SATA interface for modern high performance SSDs (and you also have to consider the limits of the interface).

Even at the extreme high end (way beyond SATA) from what I've seen most (okay I only ever saw 2 or 3) companies who used to make DRAM based SSDs e.g. RamSan are moving to Flash which suggests we've reached the tipping point. Edit: Came across [3] who still seem to be concentrating on DRAM based SSDs. Edit2: [4] says demand for DRAM based SSDs is growing.

Note if you haven't realised yet, DRAM based SSDs are still usually called SSDs. Our article on SSDs mentions them and some of the devices at IOPS appear DRAM based. RAM disk as per our article usually refers to a software based disk using the system RAM (which is different since it's connected directly to the CPU).

Nil Einne (talk) 02:29, 12 June 2011 (UTC)

June 12

New PC config question

My 9-year-old desktop computer is on its last legs, and a replacement must likely be arranged in the next month or so. The biggest change in our requirements from last time is the addition of other portable computers to the household, thus a need to set up some kind of disk, folder, or file sharing system. At my office there is a network file server, so I can envision that model, but I have no idea if that's the best/right/only model for home systems.

I don't have a clue where to begin on that. Assuming the new box will also be a desktop system with a honking big disk, what are the very first questions I need to ask, and then learn how to answer? Some I've thought of are

• Should I share a whole spindle, or a whole file system, or just some folders someplace? Are there specific pros and cons to any of those options?
• How, electronically, do the remote systems (presumably laptops) connect? I have a DSL router/modem which supports my kid's computer's internet access via wireless; does (or can) disk sharing work through that too -- but even if it can, maybe it shouldn't? Fer'instance, am I exposing data that way?
• Presumably this will be a Windows 7 box that will be built to order, including appropriate software. Is there some product that I should specifically configure to accomplish this, or is this a pretty basic feature of any Win7 variant?
• (Anything else I should be asking up front?)

I'll prefer something simple and idiot-proof over something feature-richer that requires continuous handholding.

I could probably learn much of what I think I need if I knew better what to look for -- what the buzzwords are. And, I'll admit that dealing with this, now a decade-plus removed from the software development biz, really makes me feel reeeeely stooopid, so thanks for your patience.

Thanks! DaHorsesMouth (talk) 01:08, 12 June 2011 (UTC)

Whether wired or wireless, sharing does work across the usual home network setup. The sharing strategy might depend on whether this is just for yourself or for the whole family including partner, kids etc. Although a bad idea from the security point of view (ie. not spreading malware & viruses), sharing everything might be convenient for just yourself. However, others in your household, teenagers in particular, might prefer not having to overtly share everything with anyone else (of course, depending on how paranoid your are, you might want to check up on what your kids are doing online from time to time, or at least talk to them about what you expect). Astronaut (talk) 10:50, 12 June 2011 (UTC)

Streaming IMDB trailers

In checking out the videos linked in this Entertainment Desk question, I noticed the video control bar (play/pause, progress bar, full screen) has disappeared. Is this just a 'feature' of me using Win XP/IE8 today, rather then the Ubuntu/Firefox 3.6 combo I was using last week (which definitely had the control bar)? Astronaut (talk) 11:29, 12 June 2011 (UTC)

formatted hex printing in C++

I'm adapting a program of mine from C to C++. I'm not very familiar with the C++ standard library, and doing an unsigned short to hex string conversion is proving to be clumsier than I'd hoped.

In C I do:

 unsigned short val = 0xABCD;
 sprintf(str,"%04x",val);

So far in C++ I'm doing:

 unsigned short val = 0xABCD;
 stringstream ss;
 ss.width(4);
 ss.fill('0');
 ss << hex << val;
 // and then using ss.str()

That works, but it's verbose and it seems clumsy to construct (and then dispose of) a whole stringstream object just for this one trivial format operation. Is there a more straightforward way to achieve the same result? I know I could still call sprintf from cstring:: but I'd rather know how to do these things in C++ "properly". -- TinyLittleRobot (talk) 13:15, 12 June 2011 (UTC)