Jump to content

Wikipedia:Reference desk/Archives/Computing/2012 September 29

From Wikipedia, the free encyclopedia
Computing desk
< September 28 << Aug | September | Oct >> September 30 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


September 29

[edit]

CD and DVD player laptop not working and cause

[edit]

What is the cause of CD/DVD player not working in the laptop with the disc in it and how can I fix it? — Preceding unsigned comment added by 65.92.151.92 (talk) 00:31, 29 September 2012 (UTC)[reply]

There could be many causes, like the CD not lying flat in the tray and getting jammed in, but the first order of business is to get the CD out. Have you tried rebooting ? Some drives also have a tiny hole in the front where you can insert a pin to force it to open. Try that. StuRat (talk) 00:48, 29 September 2012 (UTC)[reply]
How to reboot it?--65.95.105.26 (talk) 14:22, 29 September 2012 (UTC)[reply]
"Reboot" means to turn it off, wait a minute, then turn it back on. You may need to hold down the power button (on/off button) to turn it off. If even that doesn't work, unplug it and pull the battery out. StuRat (talk) 17:33, 29 September 2012 (UTC)[reply]

Camera lens not fully open

[edit]
See bits of the lens cover in the top right and bottom left

What causes issues such as this? Like this camera, mine often won't retract the lens cover fully when I turn it on, although it happens frequently enough that I know to touch the cover with my finger, and it immediately retracts the rest of the way. Is it perhaps a problem with a motor? The camera that took this picture is a Canon Powershot, as is mine. I've had the camera for about five years (bought it used from a friend), and it's been doing this for perhaps the majority of the time. Nyttend (talk) 00:39, 29 September 2012 (UTC)[reply]

I once had a camera that behaved similarly due to sand getting inside the lens cover, causing it to jam when it tried to open. Looie496 (talk) 02:41, 29 September 2012 (UTC)[reply]

Where are the math symbols ?

[edit]

When editing Wikipedia pages, I used to get a list of special characters to choose from below the edit window. I no longer get that list (actually, it seems to pop up only to be covered immediately by a new white bar at the bottom). I do have a list of special characters above the edit window, but that lacks the math symbols. How do I get them now ? StuRat (talk) 08:15, 29 September 2012 (UTC)[reply]

For me, in the "new" vector skin (which IIRC actually went public some time ago), they seem to be accessible from the top bar, Special characters > Symbols, etc.. I'm sure we could figure out how to keep them from disappearing below (which is indeed odd :p), etc., if it bothers you that much. As for a decent explanation for why this happened at all, good luck. ¦ Reisio (talk) 15:46, 29 September 2012 (UTC)[reply]
Thanks. StuRat (talk) 17:30, 4 October 2012 (UTC)[reply]

Compiling GFortran external subroutines

[edit]

I would like to have a GFortran program call a GFortran external subroutine contained in a separate source code file, and compiled at a later date, than the calling program. Is this possible ? If so, how do I do it ? (My goal is to compile the external subroutine while the main program is running, then call it.) StuRat (talk) 08:19, 29 September 2012 (UTC)[reply]

You need to build, and then link to, a shared object: howto. -- Finlay McWalterTalk 08:45, 29 September 2012 (UTC)[reply]
To achieve your goal of compiling the .so after the calling program, you need to do something like this:
  1. Decide upon the API of the shared object, and create and compile a shared object which presents that API. By compiling it you create a .so file, and you fix the ABI.
  2. Compile your calling program against that .so. The compiler uses the API/ABI information in the .so to generate the calling code.
  3. Later you can change the source for the .so and recompile it. Only as long as the API and ABI both remain the same (the latter should be unchanged as long as you're on the same platform and you don't change the compiler options) the calling code should be able to call the new shared object.
If you can't do step #1 for some (hopefully unlikely reason) you'd have to use dlopen and dlsym, and then you'd be responsible for managing the call stack manually (which is hard). -- Finlay McWalterTalk 09:53, 29 September 2012 (UTC)[reply]
Thanks. Have any sample GFortran compile statements to go along with your 2nd post ? StuRat (talk) 10:00, 29 September 2012 (UTC)[reply]
It's just the same as the first one. All it means is you mustn't (and hopefully can't; I don't know gfortran's name mangling) change the names or signatures of the calls exported by the .so. Just don't do the dlopen thing. -- Finlay McWalterTalk 10:05, 29 September 2012 (UTC)[reply]
Thanks. StuRat (talk) 17:30, 4 October 2012 (UTC)[reply]

Big fish in a small pond

[edit]

Is there such a thing as a table showing which Wikipedia editors (excluding bots) have made the most edits? And is it possible to refine that data so that it only reflects edits made in article space, not "community"-style interactions on Talk Pages, the Reference Desk, etc? I'm interested to know which editors contribute the most to this encyclopedia, and which of the prolific editors prefer to use it as a place to hang out. 87.112.50.156 (talk) 10:04, 29 September 2012 (UTC)[reply]

WP:List of Wikipedians by number of edits. Mitch Ames (talk) 11:01, 29 September 2012 (UTC)[reply]
And is it possible to refine that data so that it only reflects edits made in article space, not "community"-style interactions on Talk Pages, the Reference Desk, etc? 87.112.50.156 (talk) 11:54, 29 September 2012 (UTC)[reply]
If you could grab those names programmatically, and feed them through this edit counter, it wouldn't be too hard to get that data. But I don't know if anyone's done that already. --Mr.98 (talk) 15:16, 29 September 2012 (UTC)[reply]
You would find that some of the editors with the highest edit counts (even in article space) are not necessarily the most useful. ¦ Reisio (talk) 15:49, 29 September 2012 (UTC)[reply]
Right. Creating one well-researched 1000-words article is one edit. Pushing your opinion here and there can means hundreds of edits. OsmanRF34 (talk) 22:14, 29 September 2012 (UTC)[reply]
Policing vandalism generates lots of edits, without a single new word added to an article. HiLo48 (talk) 00:08, 30 September 2012 (UTC)[reply]

Multiple PCs on one home hub

[edit]

If I have several PCs connected to the same wireless hub, simultaneously using the web, presumably they are all using the same IP address assigned by the hub, so how do responses from websites get directed to the correct PC (or come to that, to the correct window or tab on a particular PC)?--rossb (talk) 08:03, 30 September 2012 (UTC)[reply]

If it's IPv4 (which it almost certainly is) then the box (which is much more a router than a mere hub) does Network address translation. -- Finlay McWalterTalk 12:35, 29 September 2012 (UTC)[reply]
The answer to the second half of your question is that different connections to the same machine are distinguished by a port number. See simple:network port (or the incomprehensible en:port (computer networking)). Since the network address translation article is also incomprehensible, I'll summarize: NAT maps IP address+port combinations on the Internet side to different IP address+port combinations on the private side. Although there's only one Internet IP address, there are 65,535 port numbers so this trick can handle quite a lot of simultaneous connections. -- BenRG (talk) 16:36, 29 September 2012 (UTC)[reply]
Many thanks for the explanation (and apologies for not having originally signed my question). Yes I certainly found en:port (computer networking) pretty difficult to follow. As a further refinement of this, am I right in thinking that a browser will connect to a well-known port (maybe 80) on a server, but the ensuing dialogue will different and more specific port numbers? This whole thing could do with being explained better in our articles.--rossb (talk) 08:03, 30 September 2012 (UTC)[reply]
The unique identifier of a TCP connection is the local and remote address and port. So when you connect to a web server on port 80, it's distinguished from other connections to that server only by your address+port. In principle a NAT router could reuse the same local port for connections to different remote address+port combos, allowing up to 65,535 simultaneous connections to each remote address+port instead of 65,535 simultaneous connections total. I don't know how many actually do that. -- BenRG (talk) 15:38, 1 October 2012 (UTC)[reply]

redirect not appropriate observation

[edit]

How/why I got to this page.? On the edit talk for the loudness page, a user was also looking for a page for volume sound for a different reason. quote "I went to Volume (sound) to learn about the entomology of the word "volume" in regards to its acoustics-related definition. It redirected me to Loudness."

And is advised to go to the reference desk.

= So may I ask/challenge why volume sound should redirect to loudness.

http://en.wikipedia.org/wiki/Volume_%28sound%29

= Volume in the context of sound as undefined in wikipedia. I would submit that volume sound should have its own page, with links to loudness. Would myself start to describe volume as a local relativistic indication of scale/phenomenon, and confined to the system your in.

I was looking myself for a description of the contrast between using gain vs/or volume in applied audio engineering. I had elsewhere read that "Gain commands are used to control the volume of selected audio data, in decibels" which seemed wrong, so that was the reason.

= The loudness article is all over the place and lacks tightness and focus. For example initially defining loudness as a perceptual phenomenon, then hearing loss (physical phenomenon) then something completely out of place about loudness compensation.

You should put your comments on the talk page of the relevant article(s). Or you can try to improve articles yourself. -- Finlay McWalterTalk 13:36, 29 September 2012 (UTC)[reply]
You might be interested in the etymology of "volume": The usage was first recorded (though with a rather wider meaning) in Busby's Dictionary of Music in 1801 "Volume, a word applied to the compass of a voice from grave to acute; also to its tone, or power: as when we say, ‘such a performer possesses an extensive or rich volume of voice’. By the late 1800s the word was being used chiefly to mean the power (decibel) output of a voice. It was then a simple step to transfer the meaning to acoustics, with "volume control" being first used in 1927. The original etymology of the size sense dates back to 1530, though it meant the size of a book originally, and Wycliffe used the words "gret volyms of newe lawes" (great volumes of new laws) around 1380, though he meant books of course. Dbfirs 16:17, 29 September 2012 (UTC)[reply]

Files and directories: could it be different

[edit]

Could an OS be organized without a tree-life directory system? It could be a matrix. Or without files? You could put files as entries in a DB? OsmanRF34 (talk) 15:16, 29 September 2012 (UTC)[reply]

Filesystems are already internally organised in ways quite unlike how the file tree abstraction makes them appear, with the abstraction built on top of those various underpinnings. Once you get to networked, replicated filesystems, the internals are very dissimilar from file/tree. People have build several database file systems, which work much as you describe, exposing a DB-like API. But often what ends up happening is a POSIX layer gets built on top of that, and the DB underpinnings get rather neglected (as was the case for the Be File System). It seems people either like a file/dir hierarchy, or that its use is so ingrained in them, and their software, that changing to some other paradigm is more difficult than just writing a different FS API. -- Finlay McWalterTalk 15:48, 29 September 2012 (UTC)[reply]
But, is there an FS API out there (obviously not mainstream) that doesn't organize files in this way? it's easy to imagine different logical and practical different ways of doing it. OsmanRF34 (talk) 15:56, 29 September 2012 (UTC)[reply]
Those described at File system#Database file systems, as I linked above; BeFS, Amazon S3, Hadoop, GFS, etc. It seems when it stops natively having a POSIX-FS API, people stop calling it a filesystem and start calling it a "persistent hash table" or "network storage engine". At the risk of tautology, it seems the world considers "file system" and "file/dir tree" to be synonyms. -- Finlay McWalterTalk 16:03, 29 September 2012 (UTC)[reply]
But are those for a single user on a PC? Can I navigate my own files through their metadata on my PC? I'll simply put all into one big directory (or whatever that be). OsmanRF34 (talk) 16:13, 29 September 2012 (UTC)[reply]
It should be noted that while the Apple iPad has a "file system" (in that it is able to store files, and as such must use some system to do so) it doesn't use a classic directory tree structure to do so, or at least doesn't do so from a user perspective. (I don't know anything about iPad programming, so there may be a tree-based directory structure in the API, but apps typically don't expose that to users directly.) Instead, files are stored on a per-application basis, and the application presents them to the user in whatever ad hoc user interface the app creator wishes. (Anything from a simple flat list, to a tag searchable database, to a non-exclusive hierachical layout, to a context sensitive mind-map.) -- 71.35.101.136 (talk) 04:35, 30 September 2012 (UTC)[reply]
  • I just want to add that in my view the idea of moving away from directories is really misguided. The human mind finds it easy and natural to organize information spatially, so it is helpful for us to be able to think of a file as located at a place; and there is no better structure than hierarchical directories for making that happen. Looie496 (talk) 17:13, 29 September 2012 (UTC)[reply]
  • Agreed. Consider installing a software application. Doesn't it make sense to put most of the files in a folder and it's sub-folders, to make uninstall simpler, among other things ? StuRat (talk) 17:38, 29 September 2012 (UTC)[reply]
In principal agree, but different OS organize files differently. In Linux the files of one application are not all in one directory/folder or sub-directory/foldre of it. bamse (talk) 18:39, 29 September 2012 (UTC)[reply]
If this is really just about tags - the same things used to find blog posts, search music collections, and organize emails in gmail - then they can be used exactly like folders. The user can choose to avoid using multiple tags per file; if the files come with pre-existing tags, the user can ignore those. Each file will then appear to be in only one folder. The resulting bland experience is perhaps the problem that Finlay was referring to; you can lead a horse to a database, but you can't make it adopt an efficient workflow. Except for when the horse is browsing through its music collection, for some reason. Then the horse likes to have Slow Ride simultaneously in the Foghat folder, the Fool for the City folder, the rock folder and the 70s folder.
- I guess you can't nest these "folders" to create a tree structure, but that feature sounds simple to add: it just needs a kind of file which, when double-clicked, searches for a particular tag. That would be equivalent to a folder icon.  Card Zero  (talk) 18:11, 29 September 2012 (UTC)[reply]
That is done in Unix/Linux systems: there is a tree of directories, but files are linked to directories rather than placed in them. As a result the same file can be found at the same time in multiple places, possibly under different names. --CiaPan (talk) 21:00, 29 September 2012 (UTC)[reply]
Or taking it to the extreme, it might not be a tree at all: http://xkcd.com/981/ --Trovatore (talk) 21:25, 29 September 2012 (UTC)[reply]
See CP/M#File system for an example of a very simple non-tree file system. My experience of using such a file system on a multiuser machine in the mid-80s was that executables were put in User 0, where they could be run no matter which user area you were logged into. Each user of the machine was given a user number to store their personal files. User 15 was used for games (and occasionally wiped by management).-gadfium 01:38, 30 September 2012 (UTC)[reply]