Wikipedia:Reference desk/Archives/Computing/2011 August 23

= August 23 =

Firefox: Is this a problem for anybody else?
Anyone who has upgraded to FireFox 6: please go to the BBC website, and scroll down to the horizontal-scrolling sections delimited by the big vertical gray bars: Health, Technology, etc.

On my system today, only the initial (or default) panel appears. Clicking any of the other three sections, whatever they are, gets you a blank panel -- although the cursor DOES change from the arrow to the pointing finger. . .?

My problem, or FF problem? If I knew what this feature was properly called I could search mozilla myself, but didn't get far with my current state of ignorance. --DaHorsesMouth (talk) 00:53, 23 August 2011 (UTC)
 * I can repeat this too. BBC call it "drawers".  The classes for the style sheets here are called contentBlocks cbg0, contentBlocks cbg1, contentBlocks cbg2 and contentBlocks cbg3.  It is something to do with their style sheet, if you do a Frefox: view, page style, no style, all the content will reappear. Graeme Bartlett (talk) 02:51, 23 August 2011 (UTC)
 * When my Firefox upgraded itself I looked at the list of known issues, and I believe this was mentioned, with a wording that implied the problem was due to the BBC site using incorrect code and that they had been notified about it. Apologies for fuzziness, I'm going on memory of something I didn't pay attention to. Looie496 (talk) 03:45, 23 August 2011 (UTC)

Looking up Chrome's history file
I'm trying to perform an SQL search in my —still quite small— Google Chrome history file History with the program "SQLite Database Browser 2.0 b1". I'd like to see all my recent Wikipedia look-ups so my query is: SELECT * FROM urls WHERE url = 'http://en.wikipedia.org/*' but it's extremely slow. I've also tried by using the REGEXP operator but I didn't get any results. Finally, there are these files called "History Index 2011-08" and History Index 2011-08-journal" whose names suggest faster look-ups, but I don't know how to use them. How can I do that in "finite" time? --Belchman (talk) 01:46, 23 August 2011 (UTC)

Okay, modified the query to SELECT * FROM urls WHERE url LIKE 'http://en.wikipedia.org/%' and now it's working. --Belchman (talk) 01:54, 23 August 2011 (UTC)

Learning Print Conversion to Web
Hello, I am employed as an artworker, which means that I use Adobe's Creative Suite to create documents for print. I use mainly InDesign, Photoshop and Illustrator. I have noticed an increasing trend for documents not to go to print but to be published directly to the web, and to be broken down for tablets and ereaders. Given that this is the way the wind is blowing, I see my current knowledge of papers, binding methods, cutters, inks etc is going to become a little outmoded and I would like to begin to reskill to stay abreast of this curve.

My question is, what should I learn? I know that CS5 and Quark already have tools within them for publishing to the web and they would be the obvious and ideal place to begin, but if I wanted a broader knowledge of this and the web publishing world in general, how would I approach the learning? Would XML be the place to begin? Or HTML? Or Flash?

Ideally I suppose I need someone to sketch out the learning path that would take a print-based artworker to fully-fledged web designer, then I would get the broadest view.

Thanks 195.60.20.81 (talk) 09:16, 23 August 2011 (UTC)


 * I would hazard to guess that HTML5 and CSS are probably your safest bets. You can do quite a lot with those two things by themselves. If you want to get into coding, that's an entirely different world with different considerations. Personally I would skip Flash for now — most sites are not done in Flash these days, and its usefulness for portable devices may be limited if iOS doesn't ever support it. At this point most of the Flash one sees anyway is based in heavy coding, not graphics design. --Mr.98 (talk) 15:21, 23 August 2011 (UTC)


 * If I were you, I'd start with Adobe Dreamweaver for a week and see how that suits you, since you're already quite familiar with Adobe's other software. As you gain familiarity with it, you'll start to be able to ask questions that are more specific, like "How can I copy the functionality ABC that I see in website XYZ?  I can't see how it can be done with Dreamweaver", and that will steer your education a bit as you either learn to do it with Dreamweaver, or learn the technologies that let you learn how to execute functionality ABC.  Comet Tuttle (talk) 15:33, 23 August 2011 (UTC)

Firefox keyboard shortcut
Is there a keyboard shortcut in Firefox for moving the text cursor backwards in a row of fields? For example, on the Yahoo mail login page, moving back from the password field to the login field.

109.74.50.52 (talk) 09:43, 23 August 2011 (UTC)


 * Shift+tab should do this. AndrewWTaylor (talk) 09:56, 23 August 2011 (UTC)


 * And note that this is a general shortcut, not one restricted to Firefox. StuRat (talk) 02:38, 24 August 2011 (UTC)

Why www and http?
Why not simply w. or h://? That would make for addresses like h:/w.cnn.com. Or even more simple: w cnn com. I suppose it has historical reasons, but why not change it? Quest09 (talk) 14:08, 23 August 2011 (UTC)


 * I can't answer "why", but most modern browsers do not require http or www. Just type 'cnn' and hit ctrl-enter (may be different on Mac), and bob's your uncle.  Some browsers, such as Firefox, offer additional shortcuts, such as shift+enter to get .net instead of .com.  I don't think there's much payback in shortening the existing prefixes, since the need to type them is nearly obsolete.  -- LarryMac  | Talk  14:30, 23 August 2011 (UTC)
 * I don't think the need to type them has gone away in that many cases. The shift-over would be great; billions of web address links that would either be wrong or would need redirecting. Easier to convince the big browsers to allow h:/ as a shortcut, perhaps with hs:/ for https:// (secure), I think. That may happen, but as Larry says they can normally make it most unnecessary anyway - like with the awesome bar or similar. Grandiose (me, talk, contribs) 14:43, 23 August 2011 (UTC)
 * (ec) (WP:OR alert) I once went to a lecture given by Tim Berners-Lee in which he said (at least partly in jest) that the length and complexity of "http://" was one of his biggest regrets about the development of the WWW. (And of course WWW itself is unusual in being an abbreviation that takes longer to say than what it stands for.) AndrewWTaylor (talk) 14:45, 23 August 2011 (UTC)


 * (ec) Larry is describing browser shortcuts that append the string that is sent out, and that works fine for many applications. However, there is a server side issue here as well. In that respect, it is worth mentioning that many people share your opinion that www is deprecated, and many sites no longer use it as a sub-domain (see e.g. many US state universities). Other people so love www that they advocate spurious usage, e.g. www.www.extra-www.org SemanticMantis (talk) 14:49, 23 August 2011 (UTC)


 * Why? Because computer geeks built the internet and the web, and they thought things like "hypertext transfer protocol" rolled right off the tongue. Watch who you put in charge of infrastructure development.
 * But it should be noted that the www. prefix and the http:// prefix are two totally different types of things. The www. prefix is changeable at each and every server level — hence you see a lot less of it these days than you might have in the past, and usually you can omit it without difficulty. A lot of that is because the servers used today are usually dedicated for the web and are happy to ignore or use the www. subdomain at their leisure (or use completely different ones -- like this very one you are reading). Http://, by contrast, is a command telling your browser what kind of site it is connecting to (as opposed to ftp://, which says, "this is for file transfer"). All modern browsers I have seen assume http:// unless specified otherwise, because that's what web browsers mostly deal with. --Mr.98 (talk) 15:15, 23 August 2011 (UTC)


 * As to the scheme part (the part of a URL left of the colon), remember that the Internet existed for decades before the web, and Tim's 1994 proposal for a unifying name scheme had to allow for the many existing protocol schemes that were already widespread. Before HTTP existed, and in the decade or so it took to sweep away a lot of other things, you'd frequently end up directly using wais, ftp, uucp, nntp and sometimes telnet and rlogin, and all kinds of weird stuff that's either defunct or never seen by real users. So Tim used the simple name for the protocol; to try to abbreviate it would just have led to more confusion and to clashes over which protocol got which letter. As to AndrewWTaylor's thing about Tim regretting :// (I've heard it said his only regret is the 2nd slash), but in this interview he says "Well, the double slash could have been omitted if I had reserved the ":" for that use only ... but people in some computer systems understood double slash".  That's the problem for designing such new protocols - you have to do so in the context of all the other stuff, and often when that other stuff dies you're left with a standard that's unnecessarily complex but unfixably standardised. So it is with railway gauges (and supposedly Space Shuttle SRB diameters) and 80 column displays. -- Finlay McWalter ☻ Talk 15:39, 23 August 2011 (UTC)


 * URLs weren't intended to be friendly names, and weren't intended to be the primary, user-visible identifier of a web site. They were supposed to tell the browser where to find a document—the Internet server to contact, the protocol to use when speaking to it, and extra data to pass using the protocol. So your question is like asking "why do I have to type 'C:\Program Files\' before the name of the program I want to run? Why not 'C:\P\' or 'P:\'?". You shouldn't have to type any of those; there are better ways of running programs. There were supposed to be better ways of going to web sites also. The current situation with URLs is the result of companies that used third-party web hosting wanting their company name in the server part of the URL instead of the name of the actual machine that served their web site. So the hosting companies registered multiple DNS names for the same server. But that wasn't good enough, because browsers only told the server the part of the URL that was intended for that purpose, i.e., the "path" part in protocol://server/path. That means the server didn't know which company's server it was supposed to pretend to be. So either all of companyname's URLs had to look like "http://[www.]companyname.com/companyname/...", or else the server had to have a different IP address for each name. The first approach was unacceptable and there aren't enough IP addresses for the second. The solution was to modify HTTP to send the server its own name in addition to the path, thus enshrining the misuse of DNS in an official W3C standard. Add a decade of stupid decisions by ICANN, and that's where we stand today. It's not too late to introduce an officially sanctioned friendly-name service that would let ads say "type 'go:cnn' to visit the CNN web site", but the trend seems to be in the opposite direction: unreadable shortened URLs and even more unreadable 2D bar codes. At least unreadable strings of random characters don't lead to trademark disputes. -- BenRG (talk) 19:36, 23 August 2011 (UTC)

Changes in BootCamp 4.0
In version 4.0 of BootCamp (the version shipped with Mac OS X Lion), are there any changes from 3.2 other than dropping support for Windows XP and Vista? Companioncube31 (talk) 15:54, 23 August 2011 (UTC)


 * Apple's official website provides some answers: What is new in Boot Camp 4.0 for OS X Lion??
 * Electronic Software Distribution
 * Support for the Windows 7 ISO installer
 * Upgrade or "clean" install Windows without using the Boot Camp Assistant
 * You can find more information at the official Apple Support website, or by asking in person at an Apple retail store. Nimur (talk) 02:29, 24 August 2011 (UTC)


 * Thank you very much! Companioncube31 (talk) 13:59, 24 August 2011 (UTC)

Any NLE that can natively edit MKV files?
I know MKV is a container, not a format. But are there any non-linear editors that can import and edit MKV files natively? I usually have to convert them to Quicktime .MOV, or h.264 .MP4/.M4V before importing. --24.249.59.89 (talk) 16:03, 23 August 2011 (UTC)


 * You specify a non-linear editor, but then state that you are converting. I would convert with mencoder. It can convert mkv to any other video format you have a codec for. It can also convert the audio if necessary. So, what are you doing other than converting that requires a non-linear editor? -- k a i n a w &trade; 16:10, 23 August 2011 (UTC)


 * I use a plugin to open them in Adobe Premiere using Avisynth: . Also, since MKV is a container, you can extract the video and audio streams as separate files and then import the file into your NLE of choice:.


 * I don't think he wants to convert first because doing so can harm video quality and because encoding can take a very long time.&mdash;Best Dog Ever (talk) 04:57, 24 August 2011 (UTC)

QR codes for Wikipedians
Does Wikipedia have instructions for Wikipedians to make their own QR codes, such as http://en.wikipedia.org/wiki/File:QRcode_User_TomMorris.png? —Wavelength (talk) 19:38, 23 August 2011 (UTC)
 * GIYF. This is pretty simple. &mdash;Akrabbimtalk 19:52, 23 August 2011 (UTC)


 * On Linux QRcodes can be encoded with libqrencode; this is used by the command line program qrencode.  Decoding QRcodes and other types of barcodes can be done with libzbar; this is used by command line programs zbarimg which reads a barcode from an image (like a photo or a scan) and zbarcam which searches a live video stream, such as one from a webcam, for QRcodes and other barcodes. -- Finlay McWalter ☻ Talk 20:07, 23 August 2011 (UTC)


 * One advantage to making it yourself is you can control the degree of error-redundant coding used. The image to the right is encoded with the highest level; I've abused this by then drawing stuff on the code (in a ways that's fairly sympathetic to how it's encoded) producing a QRCode that's still valid. -- Finlay McWalter ☻ Talk 20:53, 23 August 2011 (UTC)


 * Possibly interesting trivia: The 6 bits immediately to the right of the lower-left bull's-eye (reading bottom to top) are the same as the 6 bits below the upper-left bull's-eye (reading left to right). If the first two are both light, the highest level of error correction is being used. If they're both dark, it's the lowest level. Your code has the maximum number of correctable errors in two of the four Reed-Solomon blocks; it must have taken some fiddling to figure out just how big you could make the logo. 69.234.120.192 (talk) 13:28, 24 August 2011 (UTC)


 * That is interesting, thanks. What did you use to show the RS error levels?  I didn't really do much to get the logo to work properly; mostly I winged it.  But your question has piqued my curiosity, so I wrote a script that automates the overlaying to enumerate all the possibilities. That shows that about 44% of logo positions (where the logo is fully within the QR margin) produce a valid code - here is an animation showing them all.  The effect varies greatly by the row used (mostly due to the logo damaging the reference bullseye marks): the 4th row down is the worst (with only 16% of positions being successful) and the 12th row down is the best (with 88% being successful) - here is an animation of that row. -- Finlay McWalter ☻ Talk 17:07, 24 August 2011 (UTC)


 * Hey, that's nifty. I have my own decoding software (which isn't really fit for public consumption). You might find that you get a stronger effect from the column position if your logo is an even number of blocks wide. 69.234.122.31 (talk) 03:34, 25 August 2011 (UTC)


 * Is there similar programs (analogous to qrencode and zbar) for datamatrix codes? -Yyy (talk) 15:52, 24 August 2011 (UTC)


 * libdmtx, apparently, but I've not tried it. -- Finlay McWalter ☻ Talk 16:04, 24 August 2011 (UTC)


 * I've now built and installed libdmtx, and it works very nicely, much like the QR stuff above. The only little problem is that it doesn't come with a built-in program to poll a video camera, so you can't play at shops at home without writing some wrapper logic. -- Finlay McWalter ☻ Talk 20:29, 24 August 2011 (UTC)