Wikipedia:Reference desk/Archives/Computing/2013 January 25

= January 25 =

if i upload a pdf/docx online, will any of my info be in the metadata?
e.g. say its an assignment for school, and i have not written anything in the pdf. could it still be traced back to me somehow? e.g. will i be lised in the file's author? Haddaya9 (talk) 00:08, 25 January 2013 (UTC)


 * Docx can definitely store author information and I think PDF can. (Well, it certainly can, I'm just not sure there's a standard way of doing it.) You can see this information by viewing the document properties in an appropriate viewer. Even if you don't see your name there, there might be some forensic means of tracing the file to your computer, but probably only if the investigator had access to your computer and could produce other documents for comparison. -- BenRG (talk) 04:28, 25 January 2013 (UTC)


 * I agree with the above, there are ways to try and remove at least some of the information though - see for example  nonsense  ferret  16:34, 25 January 2013 (UTC)

Facebook telling me way too much about any email address
I forgot my password so I clicked the "forgot password" option on Facebook. Like any other site, FB asks me: what is you email address. After entering that, instead of saying: "we may or we may have not sent you an email with a link to reset your password, you might want to check your spam folder" (which is still the "industry standard" as far as I know) it says: "Ah, welcome back John! You must be John Einstein, living in New York, and this is your picture. Right?". Actually FB also tells me "We don't even think you are John", probably based on cookies or browser signature. I can't see any reason how that information might help a user who lost his password. I do see a lot of opportunities for spammers to check if an email address belongs to a real person, to personalize the spam or even blackmail a john@bla.com who appears to be Mary on FB. FB doesn't even show a captcha before telling the personal details it knows related to the address just entered. FB must have tens of programmers, lawyers, and otherwise smart people thinking about the problem that users sometimes forget their password, and they come up with this. So my question is: why? Joepnl (talk) 02:18, 25 January 2013 (UTC)
 * The point of the lost password link is that not everyone remembers the email address they have associated with their account so they provide other options to identify your account. Since some of these will identify accounts that may not belong to you particularly if you make a mistake or forget details, they want to make sure you have the right account. Arguably this isn't necessary if you enter an email address but since they also want to provide a method for you to recover your account if you no longer have access to the email address, there's no real harm in using the same thing since ultimately they are going to tell you who the account associated with the email address belongs to. Also does Facebook even require email addressess anymore? I was under the impression they were trying to get people to use Facebook for their email so potentially if you have a mobile phone number associated with the account, or some other suitable method of ID you don't need an email address. It's worth remembering that for some people their Facebook account is far more important to them then their email address and people will compromise far more by compromising their Facebook account then their email address and so they will protect their Facebook account far better then their email address. And some people will use throw away email addresses. Nil Einne (talk) 03:38, 25 January 2013 (UTC)
 * I agree if it works in this way it is a nasty bit of information leakage - however, it has always been possible to do this with a facebook search - if you search for an email address it will show you an account associated with the email. In fact i suspect that searching for it will then register a 'connection' between you and the person searched for and probably start recommending you as a friend to them.  nonsense  ferret  05:02, 25 January 2013 (UTC)
 * I think I'll just stop using FB. I am still a bit surprised, though. Joepnl (talk) 13:51, 27 January 2013 (UTC)
 * You can check Facebook's privacy settings to see if they still allow you to minimumize your 'public info' -- that is, what anyone can see who is not a friend, or friend of friend, or in your networks, or.. etc. I haven't checked recently, as over the past couple of years they've really stretched any decent sense of privacy standards, especially using a general default of opt-in, when it should be opt-out. I recall a story from 2011 (or earlier?) where they actually imported contact info from users' smartphones. After I heard that I cut down on whatever little personal info I was sharing. If the photo revelation bothers you, just switch the main profile pic to something non-descriptive. And write them a letter of complaint, though it probably won't matter much to them ya never know. El duderino (abides) 14:39, 27 January 2013 (UTC)
 * On second thought, don't bother with the letter. Just google-search "Facebook privacy lawsuit" and check out how much they care about your privacy. E.g. I imagine their rationale is something like, "well, our service is free and you agreed to all this stuff upfront... Oh, you didn't read the fine print?" El duderino (abides) 14:58, 27 January 2013 (UTC)
 * Well I've used a picture from Where the Wild Things Are for instance in the past when I needed an image to represent me ;-) It's called an avatar (computing). Dmcq (talk) 14:50, 27 January 2013 (UTC)


 * On the back end of the issue, so to speak, you may want to regularly empty your browser's cache before and after using Facebook. They've been known to data mine your websurfing history. And I'm sure they're not the only ones. El duderino (abides) 15:31, 27 January 2013 (UTC)
 * They would be using Like buttons for that (for which your browser tells FB which page the button was on), I'd be even more suprised if they somehow had access to your browser history otherwise. Anyway, I don't think it will still be cool to use Facebook when people find out it knows more about you than your parents did when you were 4 years old. BTW, even the Reuters article El duderino refers to has a like button, which hopefully at some time will be considered offensive. I wonder how big the profile is that FB has on people that are not even logged on. Joepnl (talk) 20:56, 27 January 2013 (UTC)
 * You can somewhat limit the damage by turning off "third-party cookies". In Firefox, go to Edit->Preferences->Privacy.  In the "history" section, choose "Use custom settings for history", and then un-check "Accept third-party cookies". --Trovatore (talk) 21:02, 27 January 2013 (UTC)
 * Thanks! I just did that. Meanwhile I've looked who else apparantly needed to know I just read an article on Reuters.com. I stopped typing after this: ensighten.com, outbrain.com, industrybrains.com, google.com, betrad.com, bkrtx.com, bizographics.com, npartio-inc.net, gscontxt.net, moatads.com, crowdscience.com, teads.tv, linksmart.com, cloudfrond.net, webtrends.com, adnxs.com, adsrvr.org, truste.com, twitter.com, licdn.com, invitemedia.com, revsci.com, webtrendslive.com, imrworldwide.com, bluekai.com, yieldmanager.com, krxd.com, wallst.com, globaltravelerusa.com, zendesk.com, thomsomreuters.com, domainnamewitercom, yourjewishnews.com, celebrations.com. Joepnl (talk) 21:15, 27 January 2013 (UTC)
 * Actually yes I was referring to them accessing your browser history, as well as other means. I don't have a link handy to what I've read about the former, but for the latter you could check the Beacon (Facebook) story for more detail. And given their current and recent legal battles, I'm guessing that they're still doing that kind of crap & just got better at hiding it. El duderino (abides) 10:33, 29 January 2013 (UTC)

understanding processor power
When I was younger, it was my understanding, correct or not, that the clock speed of a processor directly inversely with the time needed to complete a task. At some point it became apparent that clock speed is only one factor. It is no longer trivial to compare processors, assuming it ever was. What are the relevant factors in determining how quickly a processor will complete a task? Is there conceivably a simple system that could be employed when naming or describing a processor which would give a user the ability to compare processors without having to up benchmarks? — Preceding unsigned comment added by 78.150.17.65 (talk) 06:30, 25 January 2013 (UTC)
 * really depends on the task. Some tasks are more inherently paralizable (parallelizable if you prefer) and therefore advantage processors with a larger number of cores, other tasks may be memory intensive and advantage processors with large amounts of on-board cache - there are many differences like this which means that you can't really say one processor is faster than another, only that is performs a particular task faster.  ''nonsense'ser talk:nonsenseferret| ferret  06:34, 25 January 2013 (UTC)

Okay,so then what would you say are all the important metrics to know when comparing processors?


 * Clock speed (assuming adequate memory) was probably the main factor up to about 2004 (in fact I was expecting a "Moore's Law" speed of way over 20 GHz long before now), but the technical problems with ever-increasing speed, combined with the cheaper option of multi-core and multi-threading processor design meant that the manufacturers stopped pursuing the speed option. As explained above, there is no one benchmark.  This table compares processors by giving the total time to perform a wide range of tasks, but it might not be appropriate for your purposes.  Intel's "i3/5/7" is really just a marketing device and provides only a very crude comparison of the processors used.  If you are running Windows, then Windows Experience Index gives another comparison and shows the weak points of your configuration.  As explained above, it really depends on what you are using the processor for.    D b f i r s   12:49, 25 January 2013 (UTC)
 * Tom's Hardware charts are what I would normally look to as a starting point to see how different tasks and benchmarks produce different results for a given processor comparison - it is pretty comprehensive -  nonsense  ferret  16:29, 25 January 2013 (UTC)
 * Surprisingly no one has mentioned it yet, but see Megahertz myth. Clock speed is only part of the performance of a processor in a given task even ignoring the importance of the rest of the system. This gets most extreme when talking about quite distinct architectures (and note that to some extent Sandy Bridge and Ivy Bridge can be said to be distinct architectures). Intel was basically the last major hold out strongly marketing megahertz, but after they ended the disastrous NetBurst and also with the rise of multi core processors, even they abandoned it. To give an example of the problem about referring to 'important metrics', consider a specialised functionality like AES acceleration present on a number of modern CPUs include some recent Intel ones and most recent AMD ones. If AES is important to whatever task you're doing and a major chokepoint, the presence and usage of AES acceleration will probably make a very big difference to the performance in that task. If you're not doing AES, it's of course irrelevant. Nil Einne (talk) 18:32, 25 January 2013 (UTC)

It will help if you review (and fully understand) our article instructions per cycle. In "trivially simple" computer architectures, one digital clock tick corresponds to one instruction execution. In any real computer, that is not accurate - and hasn't been for at least half a century! (Read IBM System/360 architecture, for example). More to the point: very few useful software programs are designed or tuned at the machine-instruction layer - so "hardware instruction" is becoming a less and less useful term for describing software performance. Almost all modern software is compiled code that resides above an operating system. Specialized software libraries are translated to hardware instructions, often in very complex ways that interact with other software and other hardware. So, to fully describe a processor's performance and execution time, we now depend more than ever on a full description of the complete system - all its hardware peripherals, and all its software components, and how they all interplay. Only a few very specialized programs are so simple that we can meaningfully talk about execution time in terms of total-number-of-instructions, instructions-per-clock-cycle, and clock-cycles-per-second. Nimur (talk) 19:50, 25 January 2013 (UTC)
 * I don't know anything about the System/360, but I'll point out that the 65816, which has completely predictable (though variable) instruction cycle counts, is still in production. It was notably used in the Super Nintendo, which only ceased production in 2003. It was also used in the Apple IIGS, where I know people wrote cycle-precise code—for example the FTA's Split Demo altered the screen border color at precise moments to draw images in the border. The MIPS architecture originally executed every instruction in one cycle, and MIPS chips are still being made, but I don't know if they still use one cycle per instruction. Some newer processors have also moved back to in-order execution because of the enormous complexity of the bookkeeping in superscalar processors. Examples are the Intel Atom and the CPUs of the PlayStation 3 and Xbox 360 (no relation to the System/360, though I was hoping they'd call its successor the Xbox 370). I think it's still pretty hard to count cycles on these CPUs, though, because of the complexity of the memory hierarchy if nothing else. -- BenRG (talk) 01:25, 29 January 2013 (UTC)
 * Good points, BenRG. To some extent, the points you bring up indicate that the long-standing debate between CISC and RISC style designs has not yet been definitively decided.  At the very least, we can say that in certain domains - especially embedded computers, simple pipelines and memory systems are still considered more efficient.  And, you bring up "variable but deterministic" execution times; I think this is also true for even the most pathologically complex out-of-order-execution processor, but is a prohibitively difficult task.  Nimur (talk) 16:05, 29 January 2013 (UTC)

Downloaded Japanese files display random symbols instead of japanese characters
Is there any possible ways to get around this problem? I am using Windows 7, I never had this problem with Windows XP..

This is an exerpt from a .txt file, entitled ‚è[‚Ç‚Ý[ ; its folder name is y”wŒiz‹_“¹ê‚Á‚Û‚¢”wŒiPART1‚Æ•›ŽY•¨@

ˆÈã

‰ü•ÏA‰ü—Ç“™‚ÉŠÖ‚µ‚Ü‚µ‚Ä‚ÍAŽ©—R‚É‚µ‚Ä‚­‚¾‚³‚¢B

‚Æ‚¢‚¤‚©Žv‚¢‚Â‚«‚Æ¨‚¢‚Åì‚Á‚½‚©‚ç‚©A

Žg‚¢“¹‚ª‚‚Ü‚è–³‚³‚»‚¤‚ÈŠ´‚¶‚È‚Ì‚Å‘¢Œ`‚ª‚¢‚ë‚¢‚ë‚Æ‚¢‚¢‰ÁŒ¸‚È‚Ì‚Å

A‰ü—Ç“™‚µ‚Ä‚¢‚½‚¾‚¯‚é‚Æ•‚©‚è‚Ü‚·B

Õ“Ë‚µ‚½ê‡‚ÍA‚¨Žè”‚Å‚·‚ªƒŠƒl[ƒ€‚ð‚¨Šè‚¢‚µ‚Ü‚·B

‚È‚¢‚Æ‚¨‚à‚¢‚Ü‚·‚ª”O‚Ì‚½‚ßA¤‹Æ—˜—p‚Í‹ÖŽ~‚Å‚¨‚Ë‚ª‚¢‚µ‚Ü‚·B

ÅŒã‚ÉŽg—p‚³‚¹‚Ä’¸‚¢‚½Œ³MOD»ìŽÒ—lAŠeŽíMOD‚Ì»ìŽÒ—lAŠeŽíƒc[ƒ‹‚ÌŠJ”­ŽÒ—lA

ƒeƒbƒNƒA[ƒc—lA‚RDƒJƒXƒ^ƒ€—‚ÉŠÖ‚í‚é‚·‚×‚Ä‚Ì•û‚ÉŠ´ŽÓ‚¢‚½‚µ‚Ü‚·B

These are all supposed to be Japanese characters, but for some reason the computer is unable to decipher them as such, and spits out these random symbols instead. The same happens for any Japanese executibles. I am assuming that Windows 7 already comes with Japanese locale, so I don't know what is causing the problem. — Preceding unsigned comment added by 72.235.221.120 (talk) 09:06, 25 January 2013 (UTC)
 * Mojibake may help here. Nil Einne (talk) 12:15, 25 January 2013 (UTC)

moving blogs
I used to post a lot on wordpress, but I had a habit of getting bored or distracted after a few months and letting my blog fade away, only to start again some months later. The last time I did this, I thought it best to start again on a new account rather than try to catch up on all that had happened in that time, and to get a chance to reorganise everything a bit different. However, some months after I stopped posting to that blog, I feel the need to try again, except now I regret that earlier decision and want to get all my old posts back, merging the two blogs into one and continuing from there, preferably without having to repost each of them one by one. However, now wordpress is messing me around in awkward ways and I have lost interest in the new redesigned site, and wish to migrate over to somewhere better. (I was looking at this 'blogger' site, but if anyone knows somewhere more suitable to what I want, I can check them out too) Of course, this would mean taking everything I have posted onto wordpress and copying and pasting it to the new site piece by piece, and having it all turn up on one day, rather than spread out over a couple of years as things are now. Is there any way of avoiding this? Some way of moving both wordpress blogs over to another site as quickly and easily as possible, and keeping all the dates they were posted the same, rather than setting them all to today?

many thanks,

86.15.83.223 (talk) 16:57, 25 January 2013 (UTC)


 * Wordpress as an import/export mechanism. Go to one blog, export it (to an XML file on your computer). Same for the second.  Then on the final one, import both (one at a time). -- Finlay McWalterჷTalk 17:04, 25 January 2013 (UTC)


 * ok, so if I stick with wordpress, any way of fixing this bug where when I try to reset my password it says there is no such account, but when I try to create a new account under that name, it says it's already in use? Or if not, can I import to another site from there? 86.15.83.223 (talk) 17:19, 25 January 2013 (UTC)


 * Other sites like MoveableType, Squarespace, and TypePad, support import with the Wordpress XML format. -- Finlay McWalterჷTalk 17:40, 25 January 2013 (UTC)


 * If you think Wordpress has a bug, or has messed up your accounts somehow, that's something you should take up on their support forum. -- Finlay McWalterჷTalk 17:43, 25 January 2013 (UTC)

website privacy
So, I am in the process of setting up a website within which I can organise various personal projects, each hosted on a different part of the site, and for my university course, I need to upload my work to a website for some reason, so I thought why not put it on my own website rather than using something simple and easy like google. However, some of the items on my site are... well things I'd rather not have my tutors see if at all possible, and thinking of it, the same goes for people there for each of the different projects, they don't need to know what is going on elsewhere on the site, unless it also involves them directly. so, would it be possible to set up the site such that people linked to one section cannot access the others without a separate link, or at least knowing the address of those other sections? something like [project1].[mysite].co.uk, [project2].[mysite].co.uk and so on, or mysite.co.uk/project1, mysite.co.uk/project2... or something along those lines? Where unless people know the name 'project2' they would not be able to access those pages, or even know they existed?

86.15.83.223 (talk) 18:24, 25 January 2013 (UTC)
 * Sure if you set up your site properly they won't be able to access another section from hyperlinks within your site. Make sure you don't have any sort of site tree. But of course people could still easily access other parts of your site via other ways. If you're using the subdomain method, make sure that you've set up your DNS properly (e.g. don't allow zone transfers). Then you'd need to make sure you don't allow search engines to index the pages, most will respect your properly set up request but not all will. More importantly, how do you know someone else isn't going to link to the pages? Also while it's just an example, remember if your name is as easy to guess as project2 no one needs to go to that level of complexity. And depending on how well your site or hoster security is, someone could try to bruteforce and find hidden pages or subdomains. Nil Einne (talk) 18:56, 25 will be inaccessibleJanuary 2013 (UTC)


 * If you buy some webspace one of the facilities normally supported is the ability to password protect subdirectories so the content is inaccessible on the web without the password. That helps with for instance family photos which you only want your own family to be able to see.. You still need subdomains though if you want other people to have full control of their own stuff and yet not be able to look into each others directories. Dmcq (talk) 00:19, 26 January 2013 (UTC)