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

March 3

Keyboard shortcut in Internet Explorer

Something odd happened a few minutes ago. I went to type the four tildes as my signature and caught some other key (possibly SHIFT-ESC?). It deleted much of what I had typed and I couldn't recover it with CTRL-Z. Why on earth would I want a shortcut to remove paragraphs of what I had written? SHIFT-ESC isn't one of the shortcuts listed in the help. What happened and can I control the keyboard shortcuts in Internet Explorer 9? Astronaut (talk) 10:44, 3 March 2012 (UTC)

Yes Shift-Esc does what you said... funny one that, I didn't know of it. However Ctrl-Z works for me, and so does Ctrl-Y to redo. I'm usually paranoid after typing a lot in a wiki edit and tend to save the text in notepad before posting, as I've had wikipedia hanging before and it's not a nice feeling to lose your thought patterns. There's probably a way of trapping it with some funky keyboard interrupt utility, but I don't know of a way in Windows. Sandman30s (talk) 18:45, 6 March 2012 (UTC)

How to read natural earth's maps?

Hello, I've tried to download the database of natural earth and I got some text files... Does anyone have any software to read those text files? — Preceding unsigned comment added by 87.68.222.18 (talk) 10:51, 3 March 2012 (UTC)

Maybe the site's own data creation guidelines give a hint on how to use the data. For example, the one I looked at (coral reefs, because it was a small file) contained data that suggests it is a ESRI Shapefile - ie. it a .zip file containing a .prj, .shp, .shx, and .dbf files. Astronaut (talk) 11:35, 3 March 2012 (UTC)

Kindle Fire...not e-paper

It was almost a shock to read suddenly on Wiki that Kindle Fire has not got e-paper screen. I was under impression that Kindle ebook necessarily means e-ink as a matter of rule. Well, what it has it then ? The usual LCD like laptops ? Or IPS (whatever that might mean) is something great that makes LCD soft on eyes (like e-paper is) not harsh on eyes, that could not be read without light and could be read in broad daylight (as opposed to typical LCD), or is Fire just a tablet...?  Jon Ascton  (talk) 14:33, 3 March 2012 (UTC)

Yes, as the article says, the Fire is a tablet computer, not a dedicated eReader. It's meant to compete with the iPad (at least, it's meant to limit the iPad's encroachment onto Kindle territory). It uses the same in-plane switching LCD display technology as the iPad, which is faster and more colourful than eInk, at the expense, as you say, of contrast, daylight-visibility, and battery life. They say IPS and eIPS are much friendly on the battery than older LCD technologies, but it's still not the month or so you supposedly get from a "real" Kindle. -- Finlay McWalterჷTalk 15:39, 3 March 2012 (UTC)

Note also that Amazon continues to sell epaper Kindle readers too; they've segmented their product offering, so people to travel a lot, where regular recharging is difficult, can still get the epaper one, and they sell their own tablet for people who mostly use it at home. The also have a Kindle app for iPhone, iPad, Android, Mac, and Windows PC. And they offer an online reader for any connected HTML5 device. -- Finlay McWalterჷTalk 15:47, 3 March 2012 (UTC)

The basic Kindles still use e-ink. The Kindle Fire is meant to be an iPad competitor and uses an LCD screen. They happen to be manufactured by the same company, but they're really quite different devices. --Mr.98 (talk) 15:50, 3 March 2012 (UTC)

Note that eReader and eBooks ≠ eInk. eBooks are simply are really just a couple o different kinds of file types (.pdf and .epub, for example). You can read eBooks on pretty much any device, from smartphones to desktop computers, not just ereader devices. Mingmingla (talk) 21:49, 3 March 2012 (UTC)

I think Barnes & Noble's Nook Simple Touch runs Android and it uses electrophoretic ink. But I'm sure Android's capabilities are locked down on the Simple Touch. I know that both the bootloader and the boot ROM are locked up, so you can't install a custom kernel or bypass this by patching the bootloader. I know that the Fire's bootloader is not signature checked (search Google for FIREFIREFIRE) and the new Kindles are easier to hack. --Melab±1 ☎ 01:52, 4 March 2012 (UTC)

Wikispaces user rights

Hello. I'm used to editing Wikipedia, but am new to the site Wikispaces. I tried to ask this question on that site's Discussion Forum, but was ignored. Here is my question:

I'm working with a project to create a wiki/forum. I've created it, and have the job of setting it up and organizing it. As such, I have the "user rights" of organizer and creator. The "creator" user right gives me the ability to make other members "organizers" (a higher-level user right) and to demote organizers (the latter is something only creators can do). As part of a mutual agreement, I'm making a small group of other people "organizers," and I'm dropping all of my user rights. Will I still be listed as the creator? If I am eventually made an organizer again, will I get my creator rights back? dci | TALK 19:25, 3 March 2012 (UTC)

I'm not entirely sure how to answer this question; I did look at the FAQ, but I didn't see anything obvious. I suspect, however, that if you contact them, the Wikispaces support people may be of more help than their forum. But if anyone here knows the answer, obviously that would be more immediately helpful than that link. --NYKevin @967, i.e. 22:12, 3 March 2012 (UTC)

Bash problem with single quotes

Hello! I've written a script in GIMP that creates a playing card for a specifed rank and suit. Here's how it can be succesffuly executed from the terminal:

gimp -i -b '(script-fu-playing-card "2" "Sans" 1 0 "/home/me/Pictures/Cards-png/1-2.png")' -b '(gimp-quit 0)'

This produces the 2 of hearts; that is, the rank is a string (the first arg after script-fu-playing-card) and the suit is an integer {1..4} (3rd arg). The problem is that now I'd like to embed this command inside two for loops that iterate over suit and rank, but putting $suit and $rank variables inside that command doesn't return their values because of the single quotes. Removing the single quotes and escaping the double quotes and parentheses returns weird errors like GIMP-Error: Opening '/home/me/"Sans"' failed: No such file or directory. Haven't had luck with escaping everything, even the single quotes. So basically, how can I substitute values for the 2 and 1 in the command above with values in vars $suit and $rank while preserving the single and double quotes necessary for the command's execution (in a terminal in Ubuntu Linux). Thank you very much (this has been a big headache).--el Aprel (^facta-_facienda) 21:23, 3 March 2012 (UTC)

Sorry, but it is completely impossible to insert anything other than literal text into single-quotes at the command line. It's even impossible to embed a literal single quote into single quotes. The whole point of single quotes is to prevent the shell from doing any expansion or interpretation of the contents, so that you can be sure the contents you entered are used as-is. You might try using $(echo "'(script-fu-playing-card" "\"$rank\"" '"Sans"' $suit "0 \"/home/.../$suit-$rank.png\")'") instead of '(script-fu-playing-card ...1-2.png")'. Note that the echo command will insert spaces between arguments automatically. And make sure your script is running under bash (check the shebang). --NYKevin @957, i.e. 21:57, 3 March 2012 (UTC)

I hit similar problems all the time and, while (in my experience) it is usually possible to solve them using a bash script, sometimes it's more trouble than it's worth. I end up creating a text file using emacs and just source that. Emacs is great for this sort of thing. In this case I would start with

  echo {1,2,3,4}{1,2,...,13} > f

and then manipulate f using keyboard macros ('f' is my name for a nonce file). Then ". ./f" sources file f. Works every time, but it is perhaps not an elegant solution. HTH, Robinh (talk) 00:17, 4 March 2012 (UTC)

If you want to "interpolate" variables into a single-quoted string you can just end the single-quoted string immediately before and restart it after, e.g. '(script-fu-playing-card "'$rank'" "Sans" '$suit' 0 "/home/me/Pictures/Cards-png/'$suit'-'$rank'.png")'. The escaping approach would also have worked except that you forgot to escape the spaces (which are not shell argument separators but rather part of the third argument you're passing to gimp). You need to write \(script-fu-playing-card\ \"$rank\"\ \"Sans\"\ $suit\ 0\ \"/home/me/Pictures/Cards-png/$suit-$rank-2.png\"\), or something like that.

In situations any more complicated than this I would certainly advise using a better programming language than bash, which pretty much sucks for anything complicated enough to need debugging or maintenance. These days you can usually rely on Python or at least Perl being present. -- BenRG (talk) 04:14, 4 March 2012 (UTC)

(expanded quoting lesson, mostly written before BenRG's reply) If you want to do this without guessing, here's the key: First, understand the command syntax as a sequence of words. Second, decide what you want to put in each word. Last - and I really mean LAST - choose the quoting mechanism(s) to build the words.

The original command consists of 6 words. The first 3 are gimp, -i, and -b and they are simple enough that they don't need quoting. The other -b is also simple.

The last word is (gimp-quit 0) and it needs to be quoted because it contains a space and parentheses. If the space isn't quoted, then this one word would be split into 2 words. If the parentheses aren't quoted, they'll be a syntax error. This word is quoted by adding apostrophes to the beginning and end.

Notice that the quoting apostrophes are not part of the word. They're a piece of shell syntax, not part of the syntax of the gimp command. The word is passed to gimp as a string that is 13 characters long, beginning with the left parenthesis and ending with the right parenthesis.

Other quoting methods would work equally well in this case. You could change the apostrophes to quotation marks, like this

gimp -i -b '...' -b "(gimp-quit 0)"

or remove the apostrophes and quote each special character individually with a backslash, like this

gimp -i -b '...' -b \(gimp-quit\ 0\)

and gimp couldn't possibly tell the difference because the word that results from the quoting is identical.

Now we come to the hard part. The longest word in the original command, also quoted with apostrophes, contains not just spaces and parentheses but also quotation marks, and you want to add variable expansions to it. The biggest obstacle to doing this correctly is trying to somehow fit your answer into the pre-existing pair of apostrophes that surround the word, as if they were a non-negotiable piece of syntax for gimp itself. Instead look at them instead as part of shell syntax, which they actually are (gimp never sees them), and understand that the different styles of quoting are different ways of expressing the same thing, and solutions will suggest themselves.

The simplest (and ugliest) is to do the backslash thing. Drop the apostrophes and put a backslash before every special character in the word, then you can throw in a variable expansion anywhere you want. Like this

rank=2 ; suit=1 ; gimp -i -b \(script-fu-playing-card\ \"$rank\"\ \"Sans\"\ $suit\ 0\ \"/home/me/Pictures/Cards-png/1-2.png\"\) -b '(gimp-quit 0)'

That's unpleasant, so you'll be looking for something better. Replacing the apostrophes with quotation marks is also possible, and it will enable variable expansion. The side effect is that you'll have to put a backslash before each quotation mark in the word itself. There are 6 of them, which is still a lot of backslashes, but less than before:

rank=2 ; suit=1 ; gimp -i -b "(script-fu-playing-card \"$rank\" \"Sans\" $suit 0 \"/home/me/Pictures/Cards-png/1-2.png\")" -b '(gimp-quit 0)'

That's tolerable, but there's an even better way which will get rid of all the backslashes. You only need to realize that different kinds of quoted and unquoted strings can be combined to form a single word. For example, these commands have the same effect:

echo '((foo))'
echo '('"("foo\)')'

In both cases the characters that need to be quoted are quoted. In the first case, the letters are also quoted. In the second case they're not. That makes no difference because letters aren't special so quoting them has no effect. The result is the 7-character word ((foo)). This sort of mixed quoting can be applied to produce a word that is mostly inside apostrophes but has a few unquoted pieces with variable expansions in them, like this

rank=2 ; suit=1 ; gimp -i -b '(script-fu-playing-card "'$rank'" "Sans" '$suit' 0 "/home/me/Pictures/Cards-png/1-2.png")' -b '(gimp-quit 0)'

The shell builds the big word from these pieces:

1. '(script-fu-playing-card "'
2. $rank
3. '" "Sans" '
4. $suit
5. ' 0 "/home/me/Pictures/Cards-png/1-2.png")'

All the spaces and quotation marks are inside the apostrophe-quoted pieces, so they survive to become part of the word, and the variable expansions are unquoted so they work. After variable expansion, the pieces are put back together. One last improvement is to enclose the variable expansions themselves in quotation marks. If you don't, then when the variable's value contains a space, the word gets split. It won't happen here since rank and suit don't contain spaces, but it's a good habit. Doing that, and adding the loops, I get this:

for rank in A 2 3 4 5 6 7 8 9 10 J Q K; do
  for suit in 1 2 3 4; do
    gimp -i -b '(script-fu-playing-card "'"$rank"'" "Sans" '"$suit"' 0 "/home/me/Pictures/Cards-png/1-2.png")' -b '(gimp-quit 0)'
  done
done

68.60.252.82 (talk) 04:55, 4 March 2012 (UTC)

I am Stephan Schulz, and I approve this answer! Very well explained! --Stephan Schulz (talk) 09:04, 4 March 2012 (UTC)

Thank you all for your excellent responses, especially from 68.60.252.82. Once I understood how bash processes the words, as explained above, it made a lot more sense. This will be an indispensible reference for whenever I run into another bash problem, as well.--el Aprel (^facta-_facienda) 20:02, 4 March 2012 (UTC)

.m2ts

The .m2ts file, I'm told is the highest quality of Codec that a video can currently be on a computer. Is this correct? I know it depends on the bitrate/second and the highest the .m2ts has is in the 30,000's. Can anyone else explain this properly to me?Zaltaire (talk) 23:09, 3 March 2012 (UTC)

Well, it would help to define 'quality', but the answer is mu. .m2ts is a container format, not a codec. Additionally, bitrate is just a heuristic. Cleverer encodings can achieve better fidelity (again, appropriately defined as e.g. perceptual PSNR) reproductions of source data at smaller filesizes, but potentially at the cost of greatly increased complexity of encoder and/or decoder. However, the reason it is a heuristic is that higher birate means having more 'space' in which to store your approximation to the source material. All else being equal, this should mean your encoding can be more faithful to the source. As for the 'highest quality' of codec, if we take this to pertain only to fidelity of reproduction, then this would have to be some lossless codec - however impractical it might be for general use. 131.111.255.9 (talk) —Preceding undated comment added 02:12, 4 March 2012 (UTC).

I don't know about this codec specifically, but I doubt if any codec is universally better quality for all types of video. Cartoons might have one best, slide shows another, etc. StuRat (talk) 01:56, 4 March 2012 (UTC)

The highest quality would be a "raw" video stream with no compression. No compression artifacts would be introduced whatsoever. (Quibble: "codec" stands for "compressor/decompressor", so technically a codec with no compression is a misnomer ... but a "raw" codec would be called a "raw codec" anyway, so I'll let it stand!) Comet Tuttle (talk) 02:06, 4 March 2012 (UTC)

Here (full paper, summary) is a comparison between different lossless video codecs, which concludes:

• In Video Capture and Video Editing Area the overall clear winner is Lagarith.
• In Maximum Compression area the overall winner is YULS.
• The most balanced and flexible codec is FFV1: relatively good speed and high compression for various presets.

--NorwegianBlue ^talk 07:58, 4 March 2012 (UTC)

Before video data reaches the first codec, information has already been lost. Video data undergoes color space conversion, introducing rounding-error and clipping. Next, it undergoes chroma subsampling, because most codec algorithms (including lossless codecs) expect data in this format. So, "lossless" really only refers to data preservation by the codec, but does not mean "bit-identical" to the source data. A modern digital camera, or video software system, has a sophisticated media graph. If you would like to learn about how codecs fit into this picture, I recommend investigating GStreamer, a free software media and video pipeline. It will be a bit of a technical deep-dive to get into its source, available on GStreamer's website; but it will help orient you to the many stages of the software and hardware pipelines that take video frames from "somewhere" and put them on your screen. Codecs are only a small part of the story! On modern systems, high-quality video playback is one of the most demanding tasks your computer is asked to do - so you really should compare end-to-end quality, rather than codec bit-accuracy. If a codec stalls the pipeline to improve quality, you may lose audio synchronization, drop frames, or corrupt the image elsewhere. So, beware of claims by codec engineers - especially claims of bit accuracy - if they are uncaveated by engineering realities! Nimur (talk) 17:38, 4 March 2012 (UTC)

This surely depends on your intended purpose. If you are storing the video for immediate playback then you do have to worry about dropped frames. If you are storing the video for archival purposes or as an intermediate during editing, then dropped frames should not be something you have to worry about unless the system is incredibly flawed since it should not require real time performance and therefore should not be dropping frames. (Performance does matter, if it takes you 1 minute to decode a frame editing is likely to become unfeasible.) Of course in the archival case, it probably doesn't make sense to convert to a lossless format if your source data is already compressed and you are not doing any editing unless perhaps your source data is in a format you consider unsuitable for archival, e.g. a proprietary closed source format so you do need to convert to some format.

BTW, perhaps I'm mistaken but I was under the impression many lossless codecs do infact support Y'UV444 (i.e. no chroma subsampling is required). I believe some also support RGB, so colour space conversion may not be needed. (Remember this surely depends on your source data. If your source data is YCbCr, then you may not want to store it as RGB. Even if it was RGB sometime early on, if you don't have access to that data, it's a moot point to you.) Of course you may eventually need chroma subsampling and colour space conversion, precisely when in the process you should do that is something you need to consider.

Also from my experience, many lossless codec developers/engineers do understand these sort of things, however they do expect their users to already understand them as well so don't bother to offer explainations. I don't think it's really the fault of engineers if people use their codecs without understanding what they are doing. (Although realisticly, I expect most people considering lossless video compression who don't know what they are doing will quickly change their minds when they realise the bitrate needed. The bigger problem is people encoding video with lossy compression with perhaps excessively high bitrate and other such things.)

Nil Einne (talk) 04:09, 6 March 2012 (UTC)

March 4

Anvil Studio: Scrolling the piano-roll

I have a problem which should have an obvious and easy solution, but doesn't. I'm using Anvil Studio to compose MIDI, and can't for the life of me figure out how to get it to show me notes higher than C 5 or higher than the B above that on the Piano Roll view. How can I do that? 97.125.17.72 (talk) 08:15, 4 March 2012 (UTC)

Most MIDI instruments, synthesizers, and software have a limited 7-bit range for notes. This means that most syntesizers and software only permit a total range of 127 notes - because of a technical detail of how MIDI works. See, for example, the MIDI Note On- and Off- message format. Some synthesizers can expand the range by presetting a range with a custom Channel Control/Mode message (compare to the "octava" 8va and 8vb in conventional sheet music notation). However, there's no guarantee that a specific MIDI instrument or hardware sound card would understand such a special control/mode message. I'm not specifically familiar with Anvil, but it sounds like it is range-limiting you on purpose to remain compatible with the lowest-common denominator of the standard MIDI synthesizer output. Nimur (talk) 17:08, 4 March 2012 (UTC)

Even though 7 bits may sound limited, 128 notes (0 to 127) is very adequate. Refer to the chart at Range (music) which uses middle C = C4. Here's the MIDI range, depending on what you call middle C.

Middle C	MIDI range
C3	C-2 to G8
C4	C-1 to G9
C5	C0 to G10

In any case, C5 to B5 are not near the upper limit of the MIDI range. --Bavi H (talk) 02:57, 5 March 2012 (UTC)

I just downloaded the current version of Anvil Studio. The Piano Roll defaults to showing one octave from Middle C (which Anvil Studio calls C5) to the B above it (12 notes). The bottom half of the screen is devoted to help information. According to the help text, "To change which notes and octaves get displayed on the grid, press the Add Sounds button." (?!) Thankfully, in the Add Sounds dialog box you can choose both the lowest note and the number of notes to show. For example, if you want the Piano Roll to show the entire MIDI note range with a scroll bar, you could choose Lowest Note = "C 0 (midi 0)" and Numer of notes to show = 128. --Bavi H (talk) 03:00, 5 March 2012 (UTC)

Thanks! That was exactly what I was looking for. 97.125.17.72 (talk) 03:19, 5 March 2012 (UTC)

If you are comfortable installing programs from zip files, may I also suggest Sekaiju as another possible MIDI sequencer. Note that you have to change the language from Japanese to English: Alt+S, L, English, restart. (Or before starting Sekaiju, open Sekaiju.ini and change Language=Japanese to Language=English.) --Bavi H (talk) 03:47, 5 March 2012 (UTC)

google docs image problem

Hi. I have a large (2.5MB) jpeg image which I would like to include in a google document. Using 'insert -> image' does not work, giving an error "The image you selected is too large. Only files smaller than 2000 by 2000 pixels are allowed." OK, so I open the image in gimp and save at 4% quality, which reduces the size to about 40K. But this does not solve the problem: google docs gives the same error. This is presumably because the resolution (which gimp reports as 3072x1728 for both the original image and the reduced quality copy) is still too big. Can anyone advise (kubuntu/windows 7; free software only) please? Robinh (talk) 08:34, 4 March 2012 (UTC)

You don't need to save it at a lower quality level, but at a lower resolution. Good old xv can do it. Also, ImageMagick can do everything. Try convert -resize 1536x864 old.jpg new.jpg. ImageMagick is, as far as I know, available for every OS on the planet, and usually installed by default on most Linux systems. --Stephan Schulz (talk) 09:19, 4 March 2012 (UTC)

Since you're already using GIMP, the feature is right there - select Image > Scale Image, then change the width to 2000 or less. AJCham 09:24, 4 March 2012 (UTC)

thanks guys. Two perfectly good solutions! Best wishes, Robinh (talk) 09:46, 4 March 2012 (UTC)

Resolved

From vector to coordinates, from coordinates to vector

If you have a svg image, can you convert it back to its mathematical coordinates or if you have some mathematical coordinates (imagine circle, r = 30 pixel, position (50, 50) in a imagine 100 x 100 pixel), how can you convert it into an svg file? XPPaul (talk) 14:08, 4 March 2012 (UTC)

Start by reading the W3C Standard describing the SVG Rendering Model, and learn the primitive data types. Then, decide on a case-by-case basis. For example, SVG has a primitive to render a circle; or, you can approximate a circle by rendering as many line segments as you like. So, you have to decide which model to use when you draw a circle for your own project. In general, the SVG Coordinate object is the data primitive that defines position in a rendering context. A coordinate can be used as part of another element, like a spline, line-segment, or circle. If you want to convert a general SVG document back to cartesian coordinates, by definition you must use an SVG renderer (or, you can sloppily re-implement one of your own). If your SVG document has a lot of line-art, you may be able to extract those coordinates and draw them using a dramatically simplified subset of the full SVG specification. Nimur (talk) 17:21, 4 March 2012 (UTC)

If you're looking for a way to render mathematical coordinates as SVG graphs, gnuplot is a common way to do it. Going backwards from SVG to a gnuplot-like language is probably not possible (or doesn't make any sense, anyway — SVG is already a structured vector format; so converting the SVG circle specification to another one seems like a rather arbitrary conversion scheme). --Mr.98 (talk) 20:48, 4 March 2012 (UTC)

math question on program

I posted a question on the math reference desk that is actually a script, and perhaps this is a better place to ask if others get the same results. The mathematicians can tell me why that happens, if so. Thanks. --80.99.254.208 (talk) 19:42, 4 March 2012 (UTC)

This is the question: Wikipedia:Reference_desk/Mathematics#weird_result RudolfRed (talk) 21:38, 4 March 2012 (UTC)

I think they're answering your question pretty well. You should consider condensing your explanation a bit. And you're also leaving out the most important details: how you are working your random numbers to get a dice throw; is there something going on with your math here (seems to be your expected values might be wrong). It seems either you're getting artifacts due to the underlying RNG you're using, or how you're handling it. Use the same method and see if your distribution is uniform.

It also appears everything may just be working fine. Why don't you ask the math desk this simple question: For n rolls of a fair dice, what is the largest expected number of runs of consecutive 6s. That would provide you the longest expected number of busts for your "enormous number" of runs. Shadowjams (talk) 19:46, 5 March 2012 (UTC)

As a follow up, here's a little perl program that will count for you the number of runs of each duration for a given n. It does this by rolling a dice, seeing if it's a consecutive run, and then recording the length of the runs. It should provide you a reasonable approximation of what you'd expect.

#!/usr/bin/perl

my $last_roll = 0;my $in_a_row = 0;my $x = 1;my %count;
while (1) {
    my $roll = (int rand(6)) + 1;
    if ($roll == $last_roll) {$in_a_row++} else {$count{$in_a_row}++ if $in_a_row > 2; $in_a_row = 0}
    $last_roll = $roll;
    unless ($x % 1000000) {print "\nIn a row\t#\t($x runs)\n"; foreach (sort {$a <=> $b} keys %count) {print "$_\t\t$count{$_}\n"}}
    $x++;
}

I get this distribution after 100 million runs:

In a row        #       (100,000,000 runs)
3               321785
4               53392
5               8887
6               1488
7               271
8               40
9               5
10              3
11              1

Does this match your data at all? Shadowjams (talk) 21:45, 5 March 2012 (UTC)

It's not what the script was. The script was originally this: http://codepad.org/bU68pFDc - as you can see, it busts! Keep running it and it keeps busting. --80.99.254.208 (talk) 10:30, 6 March 2012 (UTC)

sorry, it has a mistake (it shouldn't exit but just exit the loop), but you get the idea...

As I said on the Math desk, some language implementations (including Perl) use libc's defective random number generation. Try again using Math::Random::MT. Paul (Stansifer) 18:35, 6 March 2012 (UTC)

Yes, I know it's not the same program... I'm not rewriting the OP's program for him/her... I was however trying to help the OP wrap their mind around random walk distances. As for perl's RNG... I suspected it wasn't very robust, although that link is interesting. I actually looked at some distributions of numbers that came out of it and they were less uniform than I'd expect, although I didn't get any more elaborate in checking than looking at a few runs, so that's just an anecdote. However, I'm not sure they change my results significantly, although I may try one of the modules and see what happens. Shadowjams (talk) 20:11, 6 March 2012 (UTC)

Ran it again using MT. No statistical difference in results. Shadowjams (talk) 21:23, 6 March 2012 (UTC)

Skype: incredibly loud artificial background noise

If I use Skype on my latest laptop, more often than not I get artificial background noise (the other end not hearing it at all, not like feedback or anything). It works up and down in loudness in waves, can get really quite loud and can easily disrupt conversation. The sound definitely comes from Skype, and is inseparable in the Windows mixer from the actual Skype conversation. I don't know if this is related, by my Skype-computer combination also deducts system sounds from the microphone input (the other end can't hear music I'm playing, for example). Does anyone have any ideas on how I can rid myself of this din? Thanks! - Jarry1250 ^{[Deliberation needed]} 20:33, 4 March 2012 (UTC)

Does it happen when you don't have it deduct system sounds from the microphone input? That seems like the obvious thing to rule out first. And I wonder if you can't describe the noise a little bit better. Is it white noise or something else? --Mr.98 (talk) 20:51, 4 March 2012 (UTC)

It's not white noise, it's more like someone with one of those handheld electric fans passing it by your ear every few seconds. I've never found a way to stop it deducting my output from the input, but it's probably possible. - Jarry1250 ^{[Deliberation needed]} 21:10, 4 March 2012 (UTC)

One thing that might suggest it's not that is that the sound is independent of whether or not I have my microphone enabled or disabled (another sign it's not feedback). - Jarry1250 ^{[Deliberation needed]} 21:20, 4 March 2012 (UTC)

Maybe it is the other end's processor fan or disk drive? I find that a significant problem with built-in microphones in laptops. Ask the person on the other end to use an external mic, or a headset. Astronaut (talk) 16:54, 5 March 2012 (UTC)

Sounds an interesting idea, will have to try it. - Jarry1250 ^{[Deliberation needed]} 16:59, 5 March 2012 (UTC)

March 5

Migrating all SVN metadata into root directory of working copy

I edit some text files using PlainText on my iPod touch and my iPad. They are synced through Dropbox, which then syncs with my computer. I use subversion to keep track of these files and to do this I copy over the files from /Users/User/Dropbox/Plaintext/[SpecificFiles] to the working directory after clearing out the visible files. Now I have added folders to these files and this procedure has become annoying. You know how each directory of an SVN working copy has its own .svn/ folder? I want to migrate the information in each .svn/ folder into the .svn/ folder in the root of the working directory. How do I do this? --Melab±1 ☎ 04:39, 5 March 2012 (UTC)

One thing you might try is updating to the latest version of subversion. It uses one .svn file for the whole tree instead of one per directory, so that might make your task easier. RudolfRed (talk) 04:45, 5 March 2012 (UTC)

How will I update the current working directory then? --Melab±1 ☎ 20:22, 5 March 2012 (UTC)

I use Mac OS X. Is there a way that I can update to the latest SVN without compiling and building? It's all a hassle. --Melab±1 ☎ 20:32, 5 March 2012 (UTC)

Unable to play audio file

I recently downloaded a song from a file sharin program called Soulseek. When I clicked to play the audio file, a pop up appeared that said, “The file you are attempting to play has an extension (.mp3) that does not match the file format….” After I pressed ok, another pop up appeared and it said, “Windows Media Player cannot play the file. The Player might not support the file type or might not support the codec that was used to compress the file.” The audio file is an mp3 according to the properties. It is 8.9 MB in size. The name of the file is displayed. However, the length of the audio file (I do know that it is 3:54 long), the bit-rate, and all the other properties are not displayed for some reason. I’ve been researching the Internet to see what I could do to make it play. I went to a website where I read someone mentioning K-Lite Codec Pack, so I downloaded its full version and its updates just to try, but nothing worked, nor did I completly understand how to use it, but I tried everything there possible though, so I uninstalled it. I also tried uploading the audio file to 4shared to see if it would play there, but it could not play there. However, for some reason, 4shared mistakenly read it as just a 32 kbps, forty-something minute long song. So, my question is, with all this in mind, what do I need to do to play the audio file? By the way, I was able to contact the person, who is apparently a DJ because it is included in his username, yesterday on Soulseek by searching for the file again and then clicking the option of “Send this user a message,” and he replied back in Portuguese though saying that he has no problems playing the file. I don’t speak Portuguese. However, I was able to understand what he said because I speak Spanish fluently, which is similar to Portuguese. I added the extra info that he replied back to me in Portuguese if his location sheds light or not on the problem. Willminator (talk) 05:03, 5 March 2012 (UTC)

My go-to program for looking at audio files is the freeware foobar2000 (while the name might not inspire confidence, the program is nevertheless great.) For checking file types in general (e.g. in case the file is really a zip archive of the mp3) there is file (command). Googling for file command for windows finds Windows versions. Please be mindful of viruses if you install one of those, only download from trustworthy sites. 88.112.59.31 (talk) 12:03, 5 March 2012 (UTC)

This morning I did some research and I tried downloading a special media player that supposedly plays everything called, VLC Media Player, but when I tried to play it there it said that it doesn't support some weird extension that I don't remember and that there was nothing I could do about it. Then, I downloaded GPSpot and it wasn't able to read the song's audio codec. I downloaded Foobar2000 like you told me and it was unable to play the audio file I've been trying to play, but it did work for my other songs. 88.112.59.31, or anyone here, would you mind if I ask you, if it is permissible, if I can send to you the audio file so that you can be able find out what the codec is, if it is fixable, or if there is nothing one can do? The person I got the song from at Soulseek, like I said previously, told me that he didn't know what I was talking about because he can play it. Willminator (talk) 16:01, 5 March 2012 (UTC)

How about getting a look at the start of the file in a hex editor and comparing the first few bytes to those listed in List of file signatures or on this site. It might indicate what the file actually is. Astronaut (talk) 16:50, 5 March 2012 (UTC)

I downloaded Freware Hex Editor XVB2 and the hex editor said, 00 00 00 00 00 00 00.... I opened another song, and there were pairs of numbers. I don't know how that guy is able to play the song and he was confused when I contacted him saying that the song didn't work. I guess I'll have to move on, but I'll try to contact him again and I'll tell him that I've tried everything I could to play the song. Willminator (talk) 18:42, 5 March 2012 (UTC)

movement of mouse (computer)

How is the movement of the mouse is measured? Is it by micky? Thank you.175.157.248.35 (talk) 09:15, 5 March 2012 (UTC)

Start by reading this section of our article. The "mickey" is a sort of a joke terminology; most technical mouse specifications usually use "count per inch" or "dot per inch" as the basic unit of mouse motion. For example, USB mouses follow the HID device class specification. The actual measurement of mouse motion varies; originally, a spinner wheel with a rotary encoder converted mouse motion into digital counts. Modern optical mouse systems use a complex optical sensor and onboard miroprocessor to track motion. Touch-screens and trackpads usually use capacitive sensors and software processing to track movements. Nimur (talk) 09:56, 5 March 2012 (UTC)

The Win32 API, at least, officially uses mickeys ([1]). The name is obviously a joke, but there is a need for some term to describe the raw motion deltas reported by the mouse. -- BenRG (talk) 18:42, 5 March 2012 (UTC)

Password Folder

Is there any way to protect a certain folder in Windows 7 with a password, or at least make it hard to access to anyone but me? I'm using a computer with a single account on it, so this isn't a matter of administrator privileges. Interchangeable|talk to me 18:16, 5 March 2012 (UTC)

Are you concerned about others accessing the folder via the Internet, or by logging directly onto your computer ? StuRat (talk) 20:02, 5 March 2012 (UTC)

Others coming along and using my computer while I'm logged in; the computer can't access the Internet and I'm using another one right now (so downloads such as TrueCrypt below are out of the question). I want to make the folder such that if anyone tries to open it, they will receive an "enter password" prompt or something similar. Interchangeable|talk to me 21:12, 5 March 2012 (UTC)

And do you want to allow them to use your computer, or is it OK to prevent that, as well ? If so, setting your computer to screen lock whenever you leave for a minute would do it. The problem with just protecting one folder is that if you open any of the files inside it, you will likely have temporary copies created elsewhere. StuRat (talk) 21:26, 5 March 2012 (UTC)

I still want to allow them to use the computer; I just want to protect that one folder. And what do you mean by that? Will other copies of the file automatically be distributed throughout the computer? Interchangeable|talk to me 21:52, 5 March 2012 (UTC)

I know other copies appear to be in other folders, but they might just be links back to the original. In this case, hopefully they would only be able to see the original file name, but not it's contents. StuRat (talk) 23:23, 5 March 2012 (UTC)

TrueCrypt 82.45.62.107 (talk) 20:18, 5 March 2012 (UTC)

This advice is for Windows XP, not Windows 7, but hopefully it still applies there:

1) Set up a "secure" user name that only you can use. You could just use the Admin user name for this, if you wish. Leave the general user name available to everyone else.

2) Log in under the secure account, right click on a folder, pick the "Sharing" tab, then check the "Make this folder private" box. The folder will have to be owned by the new user name to do this. You might want to just move the contents of the old, unsecure folder into this new folder.

3) Set the secure user name to screen lock when not in use, in case you walk away and somebody else walks up before you have a chance to log out. StuRat (talk) 23:30, 5 March 2012 (UTC)

That works; thank you very much! Interchangeable|talk to me 00:30, 6 March 2012 (UTC)

Actually, if I click Properties → General → Advanced there's an option to "Encrypt this Folder", but it's not clickable and I don't see any way to make it that way. Interchangeable|talk to me 01:12, 6 March 2012 (UTC)

Apparently folder encryption is available only on business/premium editions of some versions of Windows; perhaps you have a home edition. 88.112.59.31 (talk) 09:02, 6 March 2012 (UTC)

You could encrypt it.. like compress them and 7zip them... when they won't be available as you have to decompress the file you want to use.. but it's more secure... 190.60.93.218 (talk) 12:54, 6 March 2012 (UTC)

I advise against using NTFS encryption in any case. It has its uses, but if you don't understand how it works you can easily have problems. For example, if the files are accessible to an account that has a weak password, or no password, or that is set to log in automatically, then they aren't secure even though they're shown as encrypted. If you ever reinstall Windows then any NTFS-encrypted files will probably be lost forever, even if they're on an external drive that wasn't plugged in when you did the reinstall. Truecrypt is better because its security model is simple: a volume is mounted or unmounted; to mount it you enter a volume-specific passphrase; when it's unmounted it's secure. -- BenRG (talk) 18:26, 6 March 2012 (UTC)

Can't search using Google in Firefox

In Firefox, when I try to search Google, it will not search until I log in to my Google account. Works OK in Chrome & IE but not Firefox. Has anyone any idea what is wrong? — Preceding unsigned comment added by Gurumaister (talk • contribs) 19:59, 5 March 2012 (UTC)

It seems not. Can you help us out with a fuller description of what's wrong? When you say it 'will not search' - do you get an error message? If so, what? Or does your computer simply 'hang' or crash? Or maybe you are taken to a different website. All these things are possible problems, and each has its own solution. If we know more about the problem we might be able to find the answer. - Cucumber Mike (talk) 16:10, 6 March 2012 (UTC)

Of course, silly of me. Whenever I enter something into the search box and press search, I get the login dialog box for my google account and I can't get past it without logging in. Hope you can help me. Gurumaister (talk) 18:46, 6 March 2012 (UTC)

It shouldn't be a problem as long as you have a Google account. Does this happen every time you try to search, or after you log in does the problem go away for the rest of the browsing session? Interchangeable|talk to me 19:34, 6 March 2012 (UTC)

Well, no it isn't a major problem but I don't want to have to log in just because I want to use Google. Once I'm logged in I can search as often as I like. But no one else I know is forced to sign in first - is there a setting I can change that will get rid of this nuisance? I am grateful for any suggestions. 82.71.20.194 (talk) 19:52, 6 March 2012 (UTC)

Hmm. That's a little weird. I've been unable to reproduce this, and I haven't really found anyone else on teh interwebz who seems to have the same problem. Can you tell me exactly (i.e. using copy/paste) which page you are searching from? That is, the page where you enter your search query before being forced to log in. That might be http://www.google.com or http://www.google.co.uk/ig?hl=en or something else. I'm not really sure what it'll prove, but it might be worth a try. - Cucumber Mike (talk) 12:27, 7 March 2012 (UTC)

gluing images side by side

Given three images, each 4096 pixels wide, I want to glue them side-by-side to create a single image 12288 pixels wide. What's a good way to do that with, say, GIMP? You'll understand that I don't want to rely on the accuracy of my mouse-dragging; and I plan to do this several times, so a script would be especially helpful. —Tamfang (talk) 22:07, 5 March 2012 (UTC)

This isn't exactly a script, but if you set the snap size high it will make this task much easier to do by hand. Shadowjams (talk) 22:09, 5 March 2012 (UTC)

Basically you want to make a panorama. If you google for "panorama gimp" you'll get a number of relevant pages, including a video tutorial, but this script might be the most useful. Looie496 (talk) 23:05, 5 March 2012 (UTC)

We also have a comparison of photo stitching applications. Such programs can do more than simply placing the images side-by-side; they can "stitch" and blend images together, taking account of features that are common to each frame and providing a smooth blend and geometry corrections. I recommend Hugin as it is very feature-rich, but requires a good deal of technical knowledge. Nimur (talk) 23:18, 5 March 2012 (UTC)

If you just want to glue them together, ImageMagick can, of course, do it. Try convert +append in1.jpg in2.jpg in3.jpg out.jpg. --Stephan Schulz (talk) 00:51, 6 March 2012 (UTC)

Thank you! —Tamfang (talk) 01:23, 8 March 2012 (UTC)

March 6

Ethernet Ports' IP Addresses

Good Afternoon!

  In a computer with multiple Ethernet ports, does each network interface (i.e. Ethernet port) have its own IP address, or does the computer have a single IP address which is shared by all of the network interfaces?

Thank you, RefDeskers!

07:15, 6 March 2012 (UTC)

Normally each port has its own IP address.

In a corner case the addresses might theoretically be the same. E.g. a router between separate subnetworks, and the router has the same private IP address on each. Such a configuration might cause problems, though, such as an ambiguity with the bind(2) system call. 88.112.59.31 (talk) 08:58, 6 March 2012 (UTC)

Oh goodness, where to begin. Every ethernet capable device has a 48 bit MAC address, each of which should be unique (like a serial number; although you can fake them, so they're no good for authentication). Those devices may connect to an IP enabled device, in which case they could possibly be assigned an IP address.

What you're talking about works on layers of abstraction... ethernet is below TCP which is below IP which is below HTTP, etc. There's a lot of those layers. And your question is mixing up about 3 of them. So while my answer is correct, I worry you may be wondering about something else... Shadowjams (talk) 09:00, 6 March 2012 (UTC)

As a supplement to the links that Shadowjams provided, the article OSI model gives an explanation about the different layers of abstraction that Shadowjams is referring to. --NorwegianBlue ^talk 10:09, 6 March 2012 (UTC)

Thank you for that link. I couldn't think of it at the time... I wrote my answer... OSI model is the perfect article to start with. Shadowjams (talk) 20:01, 6 March 2012 (UTC)

Vector animation

Does there exist some kind of software that would let someone define animation similar to how a vector image is described and then export the animation using a variety of parameters like length of the video, frames per second, and pixel density? --Melab±1 ☎ 04:08, 6 March 2012 (UTC)

Yes, that's how Flash works, and there are other vector animation formats as well. Looie496 (talk) 06:05, 6 March 2012 (UTC)

That is not entirely what Flash was made in mind for. --Melab±1 ☎ 03:02, 7 March 2012 (UTC)

Why do you say that? It's perfect for animation and Adobe Flash Professional is actually used to make animated films and television shows shown on networks like Nickelodeon. You can actually define the movie entirely in code, if you prefer. Another good tool is Toon Boom Studio. If you need 3D vectors then you're better off using a tool like 3Ds Max or Maya.—Best Dog Ever (talk) 03:24, 7 March 2012 (UTC)

You can always generate an animated SVG, then screen capture it (or convert? I don't know of any converters though). - Jarry1250 ^{[Deliberation needed]} 12:43, 6 March 2012 (UTC)

There is SVG animation, you need to know some xml, there are lot of tutorial out there... IDK about external links but.. this might help 190.60.93.218 (talk) 12:47, 6 March 2012 (UTC)

I don't think it is capable of complex animation. --Melab±1 ☎ 03:15, 7 March 2012 (UTC)

Screen capture would reduce the quality and I don't trust all the software to work together properly for something like that. --Melab±1 ☎ 03:03, 7 March 2012 (UTC)

Programming Languages

What programming language are programming languages written in? For example, what is BASIC written in? And what is that language written in? And so on? Thanks. Eiad77 (talk) 05:45, 6 March 2012 (UTC)

Eventually you get down to machine language. A step above that is assembly language. StuRat (talk) 06:00, 6 March 2012 (UTC)

What you are really asking is what language the compiler or interpreter for each language is written in. The answer is that it varies -- I think C is probably the most common choice, though. In fact, modern C compilers are themselves written in C. In the very beginning, compilers or interpreters for languages such as BASIC or FORTRAN were written in assembly language. Looie496 (talk) 06:01, 6 March 2012 (UTC)

So is an instruction set the lowest level of a computer language? Eiad77 (talk) 06:20, 6 March 2012 (UTC)

For all practical purposes, yes. Microcode is a level still lower, but there's no longer any connection with a programming language as it's usually understood.-gadfium 08:18, 6 March 2012 (UTC)

The language itself is defined using a programming language specification, which might be a mathematical description, a description in a natural language, or a reference implementation (e.g., a compiler that other compilers are expected to conform to). Implementations of the language (compilers or interpreters, for example) are initially written in a language that has already been implemented for the platform - but once you have a working implementation, you can write a compiler/interpreter in the language itself - this is called bootstrapping. This is popular as it means you only need knowledge of the one language to understand and help develop the implementation. 130.88.73.65 (talk) 11:21, 6 March 2012 (UTC)

Historically many were written in a combination of C and yacc/lex. The Glasgow Haskell Compiler is written in Haskell. CPython (the default Python implementation), Perl 5[2] and the earliest Java compiler are written in C. --Colapeninsula (talk)

I don't really see the difference between a self-hosting compiler than a not self hosting compiler... 190.60.93.218 (talk) 16:32, 6 March 2012 (UTC)

For most purposes there isn't any. It's only when you get into questions like "What programming language do I write a compiler in?" does the property matter. If you have a self-hosting compiler system, you don't need any other external systems in order to make changes to and re-write the compiler. - Also, as no one has mentioned it yet, the essay Reflections on Trusting Trust (direct link) is an interesting read on the topic. Although the main point on trust is ancillary to the current topic, there's a good overview of what goes into bootstrapping a compiler - that is, how you introduce/encode a new programming concept into a compiler. -- 71.35.120.88 (talk) 16:56, 6 March 2012 (UTC)

Bootstrapping (self-hosting) is often done out of an attitude of self-reliance, or as a form of dogfooding. It requires extra effort (since you need at least a basic implementation in some other language to get started), but it does mean that the team building the compiler only has to think about one language. Paul (Stansifer) 19:47, 6 March 2012 (UTC)

Thanks for all the answers everyone, it's much appreciated. Eiad77 (talk) 20:20, 6 March 2012 (UTC)

Clocks inside my computer, changing time

Hello, Wikipedians

Some software can have its use restricted time-wise. Is there any 'clock' inside my computer which such software synchs up to? I'd like to change any such clock, ie change the date, but have no idea how to do this on my own. If you know of any techniques I would be very thankful!

Thank you in advance. 83.108.140.82 (talk) 17:14, 6 March 2012 (UTC)

Changing the system time usually requires admin privileges. In all probability the person (boss? teacher? parent?) who has placed time-restrictions on your computing would not have given you such privileges, as it would make it ridiculously easy to circumvent those controls (but you knew this). Even if you changed the time, most modern OSes are set to update it automatically via NTP. AJSham 17:44, 6 March 2012 (UTC)

Furthermore, using an invalid system time will break network and security services. For example, an SSL certificate or a Kerberos ticket is only valid during a period of time ... which must be mutually agreed on by both sides! Now that modern operating systems are highly network-enabled and highly secure, these issues may impede basic system functionality; so it's sometimes critical to have your system clock set to a correct time. You can read anecdotal stories of some of the ways a system, e.g., Windows, will break: here's an MSDN blog called SSL and System Time written by a Program Manager at Microsoft. I can attest to seeing similar "weird unexplainable system errors" when I have tried to muck with my system clock on other platforms, including Ubuntu and Red Hat Linux, and Solaris; and of course, I rely on Kerberos for authenticating my Mac to my home network. So, such problems due to "faked system time" can occur on almost any networked operating system. Nimur (talk) 19:16, 6 March 2012 (UTC)

Keeping in mind the above warnings, in Windows you normally double click on the time in the lower, right corner to bring up a full clock and calendar. If you have the proper privileges, this is where you change the time or date. StuRat (talk) 19:31, 6 March 2012 (UTC)

Isn't this another case to use Linux, which lets you do with your computer that (you own) rather than what Microsoft allows you to do? On linux, one simply syncs and its remains in perfect time until the seas boil dry/ or that astroid hits/ or what ever..--Aspro (talk) 19:41, 6 March 2012 (UTC)

That's not really a valid point. On Linux, if you are not a computer administrator (Superuser, root, zeus, on the wheel, or whatever)..., you may not change the system time. See, for example, date. This is entirely a matter of "are you the system administrator," not "is your system's kernel open-source." And, as I have pointed out above, even if you are the system administrator, incorrectly setting your system-time will break other services, especially network and security services.

Along the same lines, your hardware's real-time clock accuracy is independent of your operating system; whether you run Linux, Windows, or NetBSD, your hardware clock will drift at whatever rate its hardware tolerances and the laws of physics dictate. This is the motivation for network time protocol, or for using a secondary clock source. Running Linux is not sufficient to keep your clock "in perfect sync," which is why most Linux distributors pre-install an NTP sync service, ntpd. In fact, on Ubuntu, the system is pre-installed with ntpd openntpd_3.9p1+debian-1_i386, a Debian-provided port of OpenNTPD, which is not part of Linux at all; it is pulled from the OpenBSD project. Nimur (talk) 20:46, 6 March 2012 (UTC)

OK. So, I should have said running a Linux distribution with ntpd installed. But one can instal that oneself. Linux doesn't stop you. I've just checked my clock time against a radio clock and is spot on still and nothing seems broken, so please don't try to confuse me with computer science- I just want an Operating System that works, instead of an OS that works - just. 'Linux' is open source, so yes, I can make it do what I want (-when I suss out how that is), It does not deny me those rights and sudo is often all I need. And as for having to edit the Windows registry...@#✻ !!!– forget it!--Aspro (talk) 21:43, 6 March 2012 (UTC)

What is your point? Various protocols, independent of operating system choice, require synchronized clocks. Windows 7, at least, has NTP synchronization enabled by default, and the ability to do so has been around for much longer. Any wider discussion of the relative merits of the two operating systems is irrelevant to the discussion at hand. 131.111.255.9 (talk) 06:15, 7 March 2012 (UTC)

There was a time when software was offered free (often on a 5.25" floppy glued to the front of a magazine) for a limited period of time, after which you had to pay for a code which you could enter for further access. (See Shareware and Crippleware for a little more on this topic.) Failure to enter the code would prevent the software from working, unless you worked out that you could just re-set the date on the computer and 'fool' the software that you were still in the grace period. Unfortunately, I believe that software manufacturers have got wise to this since the days when men were men and disks were floppy, and this no longer works. I'm sure there is probably some other way around these limits, but I'm not going to find it for you. - Cucumber Mike (talk) 21:28, 6 March 2012 (UTC)

Free Web Based/Hosted Bug Tracker

I'm about to release a fairly detailed tool for a game (Dwarf Fortress if you care), and I'd like to have a system setup where people could submit bugs or feature requests. All of the bug tracker software I found required that I download it and host it, which isn't an option for me at the moment. Is there something out there where I could register, create a new application, and then invite others to come and issue bugs/feature requests? Thanks! Chris M. (talk) 21:55, 6 March 2012 (UTC)

If it's open source you could check out github, or google code, both have issue/bug trackers. However I don't think they have free host options for non open source projects. Vespine (talk) 22:00, 6 March 2012 (UTC)

It is going to be open source, do you happen to know if I can restrict a project on github so that I'm the only one who can update the source? I don't mind if others view or use my source for their own purposes, but I'd like to be able to control the code completely if possible. Thanks! Chris M. (talk) 22:28, 6 March 2012 (UTC)

By default, you are the only one with write permissions to your repo. Of course others can clone it, and they can submit requests for you to pull their changes back into yours, but you're not going to find your copy changing behind your back. 131.111.255.9 (talk) 06:06, 7 March 2012 (UTC)

You can always ask them just to email you bug reports. Not as good as a proper bug tracking system, of course, but it is free. StuRat (talk) 22:07, 6 March 2012 (UTC)

I'm going to be releasing it on a forum so there will be the option of just posting messages to let me know of bugs. It's partially something for myself, or any dedicated users who want to help test extensively.

Out of curiousity, what does it do? It doesn't happen to be a version of Dwarf Therapist that works on Linux, does it? Actually, wait, if it is that, lie and say that it's something else. Paul (Stansifer) 23:00, 6 March 2012 (UTC)

It's a complete rewrite World Viewer that works with the newest version along with some cool new features. Chris M. (talk) 02:14, 7 March 2012 (UTC)

March 7

Factory settings

I purchased a new laptop and want to donate my old one to a charity. How do I set it back to the original settings (I no longer have any disks (of course)). 99.250.103.117 (talk) 01:32, 7 March 2012 (UTC)

Do you mean the computer or the monitor (screen) ? For the monitor, there's probably a choice on the menu you get to by pushing buttons on the monitor. As for the computer, those defaults come from the operating system, so formatting (blanking out) the hard disk(s) and reinstalling that is the way to go, which will also remove any personal data. However, if you don't have the disks, you would need the validation codes to download it. If not, you're pretty much stuck with formatting the disk(s) and donating it with no O/S or leaving it as-is, which might expose your private data. StuRat (talk) 01:44, 7 March 2012 (UTC)

A lot of recent laptops have a recovery partition on the hard drive instead of separate recovery disks. There may be a factory reset option that you can select at boot time (before the operating system starts to load). -- BenRG (talk) 02:21, 7 March 2012 (UTC)

Thanks. I have already removed all personal data. 99.250.103.117 (talk) 04:21, 7 March 2012 (UTC)

I don't think it's as easy as you think, since personal data is stored all over the place. StuRat (talk) 04:43, 7 March 2012 (UTC)

Assuming that you really don't want random strangers reading your personal data, and you really have deleted all content you don't want them to see, you will need to scrub the areas of the disk which once held them to prevent someone from running an undelete program or forensic program to view unused sectors of the disk. There may be a solution which allows you to securely wipe all unused areas of the disk, which I am not aware of. If such a program doesn't exist or is too expensive for you, you could backup all existing data using a bare-metal restore program, securely wipe the disk, and then restore the data again so the next user has at least an operating system.-gadfium 05:07, 7 March 2012 (UTC)

Read this, this, this and this. Removing *all* personal information is not an easy process (depending on just how sure you want to be that it actually has been deleted) but these sites run you through it. As for restoring default factory settings, unless you have a recovery disk from the manufacturer, the easiest way is just to reinstall Windows. - Cucumber Mike (talk) 08:11, 7 March 2012 (UTC)

Downloading jar file from webpage

Is there any way to download this jar file?? The browser doesn't display the source, which is very interesting... --190.158.184.192 (talk) 04:54, 7 March 2012 (UTC)

If you use one of the common web browsers, you can display the source using a menu alternative or a keyboard shortcut. Websites can not circumvent this as far as I know. I tried to view the source for that page you linked, and it worked fine. I will not track the jar file for you, but looking at the source for the page in the iframe, I found part of the filename for the jar file, so that's a start. 46.239.104.78 (talk) 23:36, 8 March 2012 (UTC)

Allow program to make changes, no queastions asked

In Windows 7, I often get the message "do you want to allow this program to make changes to your computer?" Generally this is good, but is there a way to turn it off for specific programs? Bubba73 ^{You talkin' to me?} 06:34, 7 March 2012 (UTC)

Yes. You want to allow the program to 'Run as Administrator'. - Cucumber Mike (talk) 08:14, 7 March 2012 (UTC)

That doesn't sound right. Running as Admin would allow you to run programs which you don't have rights to run as a normal user, but that's not the question here, which is about programs the normal user does have the rights to run, but which ask for confirmations before taking certain actions. StuRat (talk) 08:43, 7 March 2012 (UTC)

This is a feature of Windows' User Account Control. Quoting from here: The primary difference between a standard user and an administrator is the level of access that the user has over core, protected areas of the computer. Administrators can change the system state, turn off the firewall, configure security policies, install a service or a driver that affects every user on the computer, and install software for the entire computer. Standard users cannot perform these tasks, and they can only install per-user software.

Unlike earlier versions of Windows, when an administrator logs on to a computer running Windows 7 or Windows Vista, the user’s full administrator access token is split into two access tokens: a full administrator access token and a standard user access token. During the logon process, authorization and access control components that identify an administrator are removed, resulting in a standard user access token. The standard user access token is then used to start the desktop, the Explorer.exe process. Because all applications inherit their access control data from the initial launch of the desktop, they all run as a standard user.

After an administrator logs on, the full administrator access token is not invoked until the user attempts to perform an administrative task. When a standard user logs on, only a standard user access token is created. This standard user access token is then used to start the desktop.

Normally, even when logged on as an administrator, you open programs using your standard user token. If the program requires your administrator token at some point (say, to edit the Registry), it will ask for it as described by Bubba. However, if you open the program as an administrator in the first place, it already has the token, hence no need to ask for it again. QED. :-) - Cucumber Mike (talk) 09:07, 7 March 2012 (UTC)

I have a program that I start many times per day on my Windows 7 computer. I run it from a desktop icon. One type of problem I've been having with it someone suggested running it in Win XP SP3 compatibility mode. Since I made that change, it is asking me that question about allowing it to make changes to the computer each time. How can I keep it from doing that (once should be enough)? Bubba73 ^{You talkin' to me?} 15:56, 7 March 2012 (UTC)

mini screens

So, if I do get one of these raspberry pi things, can anyone reccomend a decent but reasonably cheap screen I can run it on, I'm hoping for something around 6-9" that I can easily carry around with me and plug in when I need it, nothing too fancy or expensive. Also, a tiny keyboard if possible. I know someone with one that rolls up to fit in his pocket but it often doesn't work too well.

148.197.81.179 (talk) 18:36, 7 March 2012 (UTC)

Why? It's basically a Microcontroller for installing your own toy robot or whatever. Keyboard or screen would not help it do this. Jim.henderson (talk) 19:03, 7 March 2012 (UTC)

The Raspberry Pi is quite capable of working with a keyboard and screen, and will probably make a reasonable tool for web browsing - and indeed Wikipedia editing. It may not have the capabilities of a modern desktop PC, but the ARM CPU is a similar model to that used in iPads and the like.

In answer to the original question, I've been looking around for the same thing, but without much luck so far - at least, on a budget. AndyTheGrump (talk) 19:17, 7 March 2012 (UTC)

Ah. I didn't realize. So, when you add the keyboard, screen, Wi-Fi radio, browser and interface software, you'll end up with a tablet similar to the low-end seven-inch Coby Electronics Corporation Kyros I bought new last month for $90 USD, but larger, slower, more fragile, more expensive and less versatile. Very good. Jim.henderson (talk) 03:48, 8 March 2012 (UTC)

Keep taking the tablets. ;-) AndyTheGrump (talk) 06:28, 8 March 2012 (UTC)

One typically low-cost option is to get a rear-view monitor intended for use in a car. These can be very cheap, and typically have an RCA connector to hook up to the Raspberry Pi. The resolution might be marginal, though. The other output is HDMI, but I think it's pretty unlikely to get a small, low-cost monitor supporting that as an input. Buddy431 (talk) 06:21, 8 March 2012 (UTC)

Why get a monitor at all? Get a HDMI cable and plug it into any handy TV screen (that supports HDMI obviously). After all, the objective of the raspberry Pi is "...to stimulate the teaching of basic computer science in schools" in the same way as the BBC Micro/Commodore 64/ZX Spectrum did back in the 1980s and back then we didn't rush out (at least not straight away :-) to buy a new TV just for the computer. Astronaut (talk) 14:52, 8 March 2012 (UTC)

Yes - that's the whole point of the R Pi. For around thirty quid (inc. VAT and delivery), I get a nice little toy I can plug into stuff I've already got. I'll probably end up spending another £20 or so on bits and bobs (power supply, USB hub, SD card etc), but I'll have something I can play around with to my hearts content. I've already got a ridiculously-overpowered desktop gaming PC, and a cheapo netbook, so I don't need another computer. This is a toy. It may also be useful, though in what ways I don't know yet. AndyTheGrump (talk) 16:41, 8 March 2012 (UTC)

I just wanted something that I could use whilst out somewhere or sitting on my bed, rather than needing the huge heavy screen from my main computer. 148.197.81.179 (talk) 23:47, 8 March 2012 (UTC)

How does a call to a cell phone find its destination?

Routing in the PSTN doesn't say. Does the carrier keep a table of the current locations of all phones? I assume the call is first routed to the central office owning the phone's number, and from there it's forwarded to the appropriate tower, but how the "appropriate tower" is known and the mechanism of forwarding aren't clear to me. Is it different for international calls? -- ke4roh (talk) 19:17, 7 March 2012 (UTC)

When a GSM phone moves into a new cell, it "registers" with that cell (strictly, I think a given phone can be registered to several adjacent cells, to handle cases where it's moving and where the strength of signals from the different towers varies with terrain). That tower signals home to its owner, which in turn looks up the phone's home carrier and it signals to that company. So if you're abroad, your come carrier knows which country you're in (and which network you're temporarily registered with there). And that network knows which cell(s) you're in. 91.125.148.199 (talk) 19:41, 7 March 2012 (UTC)

Most mobile telephones have a globally unique identifier - either a Subscriber Identity Module, a Mobile Equipment Identifier, or an International Mobile Equipment Identity (in practice, often a combination of two or more such ID numbers). This is very different to the internet; while most ethernet cards do have a globally unique MAC address, the MAC address is explicitly not used by the internet protocol. In fact, the internet protocol trusts its routing tables, while a mobile telephony network protocol explicitly does not. This is one reason why you can't get free subscriptions to a telephone network by "faking" your phone number! The network relies on secure certification to prove that your phone is authorized and routable.

As a result of the uniqueness of a mobile telephone's identity, routing to its address can be managed with a much faster, more scalable protocol. This is what enables the "cell" part of "cellular" - a mobile device can switch towers - it can even switch network providers - all without losing network connectivity. For example, the GSM network is designed to use a global, authoritative giant centralized database. Contrast this with internet protocol, which does not rely on any centralized routing.

At least a few resources online refer to SS7 packets, which actually convey the data that, for example, updates the Home Location Registration database. This would be similar to the ARP portion of the OSI stack; but beyond this level of detail, my familiarity with GSM is exhausted. There are some resources at our external links section that look very promising. Nimur (talk) 22:28, 7 March 2012 (UTC)

Strictly, the Home Location Register (HLR) isn't global - there is one per mobile operator. When a user moves out of their own network operator's "reach" and roams onto another network, the new network notes their location in their Visitor Location Register (VLR) and informs the owning network, who enter the VLR details in their own HLR. So a call to them when roaming first goes to their own network provider, who then refer to the roamed network provider, who determine the relevant tower from their VLR. --Phil Holmes (talk) 17:16, 8 March 2012 (UTC)

Digital Photo Frame

How does it actually work ? Is it LCD or E-ink (article is not much helpful) 124.253.51.170 (talk) 23:25, 7 March 2012 (UTC)

I've seen an LCD version. However, that doesn't exclude there being others using different technology. StuRat (talk) 03:28, 8 March 2012 (UTC)

E-ink is primarily a monochrome technology - they have a product with 4096 colors, but that is a fraction of the what most LCD screens offer. --LarryMac | Talk 13:03, 8 March 2012 (UTC)

(EC) As hinted at in Electronic paper and a simple search will confirm, commercial colour e paper displays are still very rare, even with e-readers. In the case of e readers they only really became available over the last year using Qualcomm Mirasol screens which are Interferometric modulator display, and E Ink Triton screens which are electrophoretic [3]. (The Mirasol is only entering in to mass production later this year [4].) They are also very expensive. So while it's possible someone produced a colour eink digital photo frame, it's unlikely and it will be very rare. Also the colour depth of such displays seems to be still fairly low, e.g. E Ink Triton only does 4096 colours. I'm not sure of Mirasol, but it doesn't see much higer. So photo quality likely won't be that great anyway.

Pixel Qi is planning to introduce displays which can keep colour even when the backlight is off [5] [6] (their previous screens lost colour when the backlight was off) although I'm not sure what sort of colour quality it will achieve in that mode. I'm also not sure whether the reflective mode in Pixel QI displays is really as good as e paper displays. And in terms of power consumption, I believe even without the backlight (in reflective mode) power usage for the display is still fairly high compared to e paper displays. The E Ink Triton is I believe zero power for static display. Despite some confusing contradictory information including in our article, I believe the Mirasol at least isn't quite zero power, see the comments [7] and [8]. However the power usage is apparently around 1 mW [9] [10] [11]. This [12] gives 0.1W for a Pixel QI type display without backlight. (Due to the way LCD works, I don't believe the power usage will vary that significantly if the image is static.)

In any case, it's a moot point as the Pixel QI product able to retain colour without the backlight still isn't available [13] so even if someone wanted to make a digital photo frame with one, you won't be finding it yet. (Also 10" would be a fairly large photo frame.)

Nil Einne (talk) 15:08, 8 March 2012 (UTC)

March 8

1=

[14]

I have technical issue there with the template where including something like "this" will "break" the template. I thinking making it more userfriendly would be an improvement. Any help or discussion would be greatly appreciated.Curb Chain (talk) 01:42, 8 March 2012 (UTC)

The problem was to use an equal sign (as part of a url) in an unnamed template parameter. This is a general problem. See Help:Template#Usage hints and workarounds. For your situation, the documentation at Wikipedia:Dispute resolution noticeboard/Header could be changed from "please put {{DRN archive top|<reason for closure>}} at the top" to "please put {{DRN archive top|1=<reason for closure>}} at the top". Did you copy the code from there so this change would have fixed it? PrimeHunter (talk) 02:09, 8 March 2012 (UTC)

Question about Sales Receipt Templates in Quickbooks 2010

I am currently using Quickbooks 2010 to record my daily sales. I have created a template, and I have added items for Sales and for the types of payments that we accept.

The problem that I am having is that every day, I have to re-enter each Item into the receipt. Since I use the same items every day, I would like quickbooks to just have these items listed on the receipts so that each day I can enter the totals.

How can I do this? I've searched google for "using the same items on a quickbooks sales receipt" but that's been no help.

Thanks. John173.243.168.189 (talk) 05:22, 8 March 2012 (UTC)

These comments aren't about Quickbooks specifically, but hopefully will still apply:

Do you sell each item separately ? If so, perhaps you need a receipt template for each item, and one without the item specified. Then just bring up the proper receipt template for each item, or, for unusual items, bring up the blank receipt template.

If you sell multiple items together, you could have a receipt template which lists all the items you sell, and you'd just indicate the quantity of each. However, this approach only works if you have a small selection of items you sell.

If you sell thousands of items, with dozens of different items sold each time, then you're out of luck, and will have to continue to do as you are currently doing. StuRat (talk) 05:40, 8 March 2012 (UTC)

"Delayed" page numbering in MS Word 2010

I have a multi-page document in which I need to number the pages from the third "physical" page. The first page is a title page, the second is a table of contents, the "actual" document only starts on the third page which must be numbered "Page 1" and onwards. For the purposes of page numbering the fist two pages must be ignored. Roger (talk) 08:08, 8 March 2012 (UTC)

It's traditional to ask a question on the Reference Desk, but my psychic powers tell me that you might find your answer here: [15]. Henry^Flower 08:47, 8 March 2012 (UTC)

Thanks, but that supresses page numbering for the first page only. I need it to ignore the first two pages. Roger (talk) 09:23, 8 March 2012 (UTC)

Yes, I was surprised to find that Word doesn't allow numbering to start at a negative page number. Dbfirs 09:45, 8 March 2012 (UTC)

You could put a section break after the table of contents. Turn "link to previous section" off in the footer formatting for the second section, so that you can control section footers separately. Add page numbers in the fotter for the second section, and start page numbers at 1. Gandalf61 (talk) 09:35, 8 March 2012 (UTC)

That works! Thanks, you're a wizard! Roger (talk) 10:29, 8 March 2012 (UTC)

What was the internet like...

I started using the internet in the late nineties, and I wonder what it was like before then. What content could be found on the internet? What did people use the internet for? How did people browse without indexing services? How did it differ from our current Darknet? Could anyone tell me, or link me to a place where it gets explained in more detail? Thanks.212.123.1.140 (talk) 08:43, 8 March 2012 (UTC)

There were lots of text-only pages, and we used Gopher (protocol) browsers to view them. There really weren't search engines, so somebody had to give you an address (which you had to type in flawlessly) or you had to follow a link to find new things. I used to look up David Letterman's Top Ten list each day, for example (just the text, not a video of them being read, like today). Electronic bulletin boards were popular then, kind of a bare-bones version of a wiki. StuRat (talk) 08:50, 8 March 2012 (UTC)

Usenet, which still exists, was also big back then. Roger (talk) 09:35, 8 March 2012 (UTC)

Honestly, I don't recall gopher having a significant presence until a year or two before the WWW exploded. In the Eighties it was mostly e-mail and Usenet, and sometimes ftp. --Trovatore (talk) 22:36, 8 March 2012 (UTC)

And of course if you wanted to know the opening times and price of something, you didn't use the internet. You phoned directory enquiries, got a phone number and waited to the next office hours before talking to a human. -- SGBailey (talk) 11:08, 8 March 2012 (UTC)

Here's a Slate article on the web in 1996. If you google Internet in 1995 etc, you'll get lots of links.

In the late 90s there were starting to be search engines like Altavista (launched at the end of 1995), but if you wanted to find info your first stop was usually Yahoo! (formerly Jerry and David's Guide to the World Wide Web) which started out as a categorised directory of websites, all generated manually. There were a lot of text-heavy pages set up by people with access via their work or educational institution (on bands, TV programs, and everything else that interests geeks), but unless you had that sort of connection, few people had personal webpages, and also there were far fewer companies with web pages. Amazon.com launched in 1995, and eBay was another early site, though initially in the San Francisco area, and they didn't have the domination they have now.

Pre 1995, it was very different, as already mentioned - there were text pages available with Gopher (protocol), and Usenet, but not much in the way of pictures, and essentially no e-commerce; there were also "walled gardens" like Compuserve and AOL which offered email and syndicated content to subscribers, but no access to the public/academic internet until the mid 90s.

See History of the Internet. --Colapeninsula (talk) 16:23, 8 March 2012 (UTC)

My first ever connection with the Internet wasn't with WWW, Usenet, or even Telnet. It was with FTP. I looked at what my classmates were doing in computer class after the teaching had ended, and they told me they were connecting to Aminet. This was like a treasure trove to me. All that non-commercially-published software which I previously had had to order through mail-order was right there, for the taking, and for free. The only things I had to do was to type ftp some.site.somewhere at a DOS prompt (the computers we had there then only had Windows 3.1), then type cd pub/aminet, find some interesting file, and type get InterestingFile.lha. After about ten to fifteen minutes of displaying row after row after row of "#" marks on the screen, the file was there on the PC's hard drive. Then I'd just insert a floppy disk into the PC, copy the file to it, and when I got back home from school, copy the file from the disk to my Amiga (Amigas can read PC disks, but not the other way around) and start using the software. Soon afterwards I learned Amigas could connect to the Internet too, so I could cut out the middle man if I wanted to, but back then, modems were pretty expensive, and the Internet connection at school was faster and more reliable than the dial-up connection I had at home. JIP | Talk 20:50, 8 March 2012 (UTC)

And what's more, Aminet also gave me the opportunity to upload files, allowing me to publish software I had written myself. I would never have even imagined being able to do that with a mail-order company. But Aminet just accepted whatever I sent to it. And I remember that some people out there actually liked my software. I was much pleased when I got my first ever registration request for a Shareware game I had written, I think it was from an American man who had been stationed in Guam. I think I got something like four to six registrations in total. JIP | Talk 20:55, 8 March 2012 (UTC)

I got on in 1994. Most addresses were IP addresses, IIRC. You got files through FTP. I didn't have a graphical way to access the internet until after I got Windows 95 in 1996. Bubba73 ^{You talkin' to me?} 22:18, 8 March 2012 (UTC)

In those days you could do USENET messages, email, download some files, and play text-based games. Then in late 94 or early 95 my wife took me down to her work and showed me the first graphics I aw on the internet. It showed the logo of the organization and you could click to get various information (although it wasn't working at the time.) Bubba73 ^{You talkin' to me?} 22:33, 8 March 2012 (UTC)

Best practices for application protocol and library API design

I'm looking for best practices for application protocol and library API design. I sometimes deal with systems that are structured as a collection of services interacting with one another using custom application layer protocols. I'm looking for guidelines and best practices for the design of these application layer protocols. I'm also interested in guidelines and best practices for the design of software library APIs. The libraries here are for some rather specialized purposes but are intended to allow multiple applications to be built using them across several device platforms. I can think of some good practices but would like to benefit from the experience of others. Thanks in advance. — Preceding unsigned comment added by 173.49.9.204 (talk) 11:26, 8 March 2012 (UTC)

Some rules I've found helpful:

1) Try to categorize your APIs. For example, those dealing with graphics might go into one library, those dealing with database access in another, etc. Avoid any APIs that fit into multiple categories, by breaking them up. In your case, you might want libraries for each platform and a library of APIs which work on all platforms (and a "That platform not supported." error would also be nice).

2) Clearly state which arguments are input only, input/output, and output only. I also suggest always keeping them in that order, separated by an extra space between types, to make it obvious which is which without having to bring up the API documentation and read through the comments ("inp1,inp2, io1,io2, out1,out2" but with more useful argument names, of course). Argument names should also include the variable type, and, where applicable, the length. So, "FilenameC32" instead of just "Filename". I can't stand trying to use an API which isn't properly documented.

3) Return useful error messages. So, "The string length parameter can't be zero" is more useful than "Division by zero error".

4) Clearly state whether each API can have multiple instances running at once, and, if not, the second API instance should ideally return an error saying that an instance of the API is already in use.

5) Allow a debug flag to be passed in, perhaps with multiple levels, to specify the detail of the output traces produced.

6) API name should fully describe what it does. So, names like "OpenDbTbl", not "MyAPI38". StuRat (talk) 16:28, 8 March 2012 (UTC)

My own suggestions for API stuff:

1. Never assume there is just one of anything (just one screen, one network adapter, one user, one mouse). Almost always it's sensible to have the API allow for multiples, even if your implementation doesn't. The Java AWT implementation is (was) hamstrung by having a singleton Toolkit object, which was bound to a single graphics output device. Want to open windows on two different screens? Want to do font-rendering in an offscreen context, without a round-trip to the X server? No dice. They fixed that later, but had to complicate the API to do so. If they're just had an explicit CreateToolkit(device foo) call this wouldn't happen.
2. No implicit global state. This is almost a corollary of my first point. API users get state objects (which are opaque to them) and pass them to other calls. Your implementation may (will, surely) have global state, but this shouldn't be overt in the API. This way to can add and change stuff later, and you don't have to guess which thing the caller is talking about.
3. Use abstract types, even when you think it's obvious what type something will always be. X11 is very good about this. Even though many of their types resolve to uint32_t, the API doesn't admit to that. That way, ten years from now, if you decide to add vastly more support for something, no-one can have legitimately made assumptions about how many things are supported. So the API should be full of connection_t, user_t, pid_t, even if they're just typedefs for int. An example: a while ago I worked on adapting an existing networking library. Its author hadn't obeyed my first rule: he'd assumed there was only one network adapter. My job was to make it work with 16 adapters. But the author had been good about using abstract types - so he used connection_id_t for each connection, which was just a synonym for an int. I redefined this to a struct which had the adapter_id_t and a local connection_t, and I only had to change the places where a connection_id_t was created, destroyed, read, or written. In 95% of the code it was just passed around (per point #2) and that code didn't change at all.
4. Worry about synchronisation and reentrancy now (and when you do you'll see again why #1/#2 is such good sense). Retrofitting reentrancy is hard if you didn't think of it at the beginning. But remember that synchronisation can also be expensive; some APIs deliberately chose not to be reentrant (Java Swing, QNX-Photon); Swing makes very good provision where multi-threaded applications can use Swing effectively, even though it's single threaded (Photon doesn't, and is a pain). In many languages, accessors (getters/setters) are easier to synchronise than direct access (e.g. in Java one can implicitly synchronise foo.setHeight(2) but not (really) foo.height=2
5. Worry about resource deallocation (in both normal and error/exception circumstances). The abstract objects your API presents may correspond with concrete OS objects like memory allocations, sockets, threads, file descriptors, etc. If a caller hangs onto a bunch of those, they may innocently consume system resources and cause subsequent calls to fail. So you need to understand where those resources are going, and be able to tidy up and reclaim them, perhaps aggressively. And you may choose not to allocate such "real" things until they're actually needed. For example the Java JVM allows a program to have code synchronised on potentially thousands of different locii (which logically confirms to thousands of OS mutex objects) - but the JVM doesn't actually allocate an OS mutex until one is actually needed (and some implementations aggressively cull unused ones).
6. Design for virtualisation. You might be coding with real users talking to real disk drives on real connections, but design for all that stuff to be pluggable so a user can be a script, a disk drive can be remoted over a network, etc. This makes design-for-test much easier, because you can plug in a test script in place of a user, and another for remote parties, disks, etc.

87.114.9.76 (talk) 17:25, 8 March 2012 (UTC)

Another one: combine rough- and fine-grained errors. Fine grained is vital for actually pinning down problems: imagine you get a report from the field, and it simply says that some user is getting a "network_error" - that's obviously not much for you do go on. If the error is "network_error(dns_lookup_failed)" you're much further along. But wrapping that with rough grain is essential (because most users of errors just care that it's a network_error, not which specific type of network_error). Python's habit of having an error as as a tuple, which contains (perhaps nested) payloads is a good one; unix/C's errno isn't a good example to follow, if you have the choice. 87.114.9.76 (talk) 17:36, 8 March 2012 (UTC)

A corollary of #5 is that an implementation of your API, and a user of it, is quite entitled to ask for a detailed accounting of what's been used. As best as you can, design the API so an implementation can track who owns what resource. That way to can have accounting APIs that will list all the scarce resources, to allow someone to track down which apps or subsystems are hanging on to stuff they could be releasing. And a test harness can deliberately push the system into the red (where there aren't any more sockets or IPC queues or memory left for apps to get) to verify the system, and the apps, behave as best as possible when system resources get exhausted. 87.114.9.76 (talk) 17:45, 8 March 2012 (UTC)

Google ads

Did we start having google ads in Wikipedia? Why?--188.4.233.216 (talk) 13:07, 8 March 2012 (UTC)

I don't see any ads. --Mr.98 (talk) 13:12, 8 March 2012 (UTC)

I use Google Chrome and now in all pages but the main, there is a Google Ad. I tried Explorer too. What can I do?--188.4.233.216 (talk) 13:15, 8 March 2012 (UTC)

There are no ads. Are you sure you are viewing http://en.wikipedia.org/ or just some other site replicating Wikipedia content? Astronaut (talk) 14:46, 8 March 2012 (UTC)

Absolutely! Even in this page, there is an ad in the top. I really don't make fun. I talk seriously.--188.4.233.216 (talk) 14:53, 8 March 2012 (UTC)

I suspect you have a virus/malware or something that is inserting ads into the pages. There are no ads on Wikipedia. --Mr.98 (talk) 15:12, 8 March 2012 (UTC)

Can I do anything? I must say that yesterday night, there were ads only in the GREEK Wikipedia. This didn't have.--188.4.233.216 (talk) 15:49, 8 March 2012 (UTC)

Try running Malwarebytes' Anti-Malware and see if it finds anything. --Mr.98 (talk) 17:13, 8 March 2012 (UTC)

Well, I think it's not a virus. From the... symptoms I mean!--188.4.233.216 (talk) 17:33, 8 March 2012 (UTC)

We all think it is, since if the site is not sending out the google ads, but you are receiving them, then something is interfering with your browsing pleasure. Either your PC is inserting the ads, or your internet service provider is doing so. If we can exclude the latter - can we? - the the former looks the likely perp. And if you did not tell the PC to do this, then someone else did. We tend to call that a virus. Why exactly do you think the symptoms point away from a virus? What do you think are the symptoms of a virus? --Tagishsimon (talk) 18:12, 8 March 2012 (UTC)

Well, I haven't searched viruses that good, so I think you're right. However, these three points:

1. it does not happen in the Main Page.
2. It happened first at the minor Wikipedias and then at the central (english).
3. It doesn't happen at any other part of the Wikimedia Project or any other site.

make me think it's not a virus. Am I wrong? Anyway are here any Chrome users to tell us--188.4.233.216 (talk) 19:23, 8 March 2012 (UTC)?

It's almost certainly some form of extension you've installed in Chrome - perhaps unintentionally, as this sort of thing often masquerades as a harmless download manager, etc. Try looking at your extensions (which I believe are under the preferences menu), turning some off, and seeing what happens. Shimgray | talk | 19:38, 8 March 2012 (UTC)

The fact that it apparently shows up in IE makes me think it is not a Chrome extension but something else. (And I can confirm that Chrome isn't, by default, showing ads on Wikipedia.) Perhaps if the OP took a screenshot and uploaded it, we might get a better idea of what's going on. --Mr.98 (talk) 19:44, 8 March 2012 (UTC)

Well, I just installed AdBlock and the're gone...--188.4.233.216 (talk) 19:59, 8 March 2012 (UTC)

I uninstalled it, because it seems that with it installed, I cannot have a spages sequence. However, I just remebered that when I first installed internet, with Firefox there were Google Ads at Wikipedia for some weeks and then diappeared. I hope that it is such a situation.--188.4.233.216 (talk) 20:19, 8 March 2012 (UTC)

Why don't you try running Malwarebytes as suggested? It is free. It will help rule out it being a virus. It is really not that much work. --Mr.98 (talk) 21:32, 8 March 2012 (UTC)

User contribution pages

I checked through the Village Pump but couldn't find anyone complaining of the same problem. Using Firefox on Ubuntu with several tabs open. All the pages, both Wikipedia and non-Wikipedia, look fine except for user contribution pages, and my watchlist, which are in a larger font. I can reduce the user contribution size but that also reduces the size in the other tabs. I checked and the problem is not there in Chrome. Any suggestions? CambridgeBayWeather (talk) 21:42, 8 March 2012 (UTC)

MSN

I'm having a couple of problems with MSN messenger. I used to have it pop up on my desktop every time I turned my computer on and found this annoying because I never used it, and eventually found some way of getting rid of it. Now, though, I do want to use it and can't find out how to get it back. Any ideas, because otherwise I have to have my emails open on another tab for anyone to contact me. Second problem, I can't seem get send anyone messages on it, I have to wait for them to send me one. I've tried every page on the site I can find, nothing seems to work, even when I followed the instructions given to me by someone on another site, it just opened the profile page over and over.

148.197.81.179 (talk) 23:53, 8 March 2012 (UTC)