Wikipedia:Reference desk/Archives/Computing/2014 July 8

= July 8 =

Why is Google finding an old Wikipedia article?
Maybe it's different for you, but when I Google (no quotes) "HM patient" or "HM disorder", I get the current Henry Molaison page first, with the full name highlighted as a synonym, along with the other words. When I try "HM memory", I get HM, which is currently a disambig page, but used to be Molaison's article title, and is still treated as such. The word "memory" isn't highlighted in the snippet, but "H.M." is.

Does the term "memory" have a special back-in-time function on Google? Or what else might be happening here? InedibleHulk (talk) 03:04, 8 July 2014 (UTC)

"HM 1926" also finds the old article, with "1926" highlighted, while "HM 1926 memory" finds the new one, with all terms highlighted. So I guess it isn't a special "memory" keyword doing this. InedibleHulk (talk) 03:10, 8 July 2014 (UTC)


 * Remember that Google locates pages not just by their content but also by the terms people use when linking to them. These are likely to be similar for both locations where the article has been, so you get both locations in the search result.  In addition, different pages may stay different lengths of time in their caches, producing this sort of anomaly.  --50.100.189.160 (talk) 05:24, 8 July 2014 (UTC)
 * Google is weird like that. I don't use my real name on my user page, but searching for it in Google brings back User:Crisco 1492 as the first link. — Crisco 1492 (talk) 05:56, 8 July 2014 (UTC)


 * That's a bit scary. A search for "bun" finds me blood urea nitrogen first and bun...somewhere beyond the first seven pages, probably. No Bun (hairstyle), either. I'll accept "Google is weird" as reason enough, I think. InedibleHulk (talk) 06:19, 8 July 2014 (UTC)


 * Google uses locale for ordering pages. If people in your IP locale do medical searches often, your results will lean towards medical terms, such as a BUN lab. — Preceding unsigned comment added by 209.149.115.166 (talk) 11:31, 8 July 2014 (UTC)
 * Probably explains why my Wikipedia profile comes up, as I do a lot of Wikipedia-related searches. — Crisco 1492 (talk) 11:34, 8 July 2014 (UTC)
 * Yes, but the question remains how it put your WP searches together with your real name. It's not, from what I see, on your user page.  Dismas |(talk) 14:21, 8 July 2014 (UTC)
 * There's several sites (of course, I can't think of any right now) that prowl the web and aggregate accounts with the same username and data - for example, for me, if I had a blog that listed my Phoenixia1177 email on it, it might pool that with my wikipedia username. Once all that is aggregated, it then offers up usernames, site memberships, addresses, phone numbers, etc. indexed for searching for paying members. Essentially, they stalk people online for you - but, some of this data is given for free, like what real names correspond to Twitter/FaceBook/etc. Perhaps Google does something related; or makes the association from such sites when it scans them. --I have no sources, but I've stumbled across my own info before by searching various user names I use; it's actually a little scary (sadly, not shocking) that there are automated systems out there doing all that right now...Phoenixia1177 (talk) 17:10, 8 July 2014 (UTC)
 * Right, which is why some security-conscious internet users use different pseudonyms for each site. I myself was considering using my WP moniker on a different site. I decided against it. Then I wondered if even "franticmantis" would be close enough for cross-contamination. But now that handle is linked to this one, so I can't use that either o.O SemanticMantis (talk) 17:15, 8 July 2014 (UTC)
 * Oh, I know exactly how my real name is linked to my user name in Google's databases. I take credit for my photography and digital restorations under my real name, and that has appeared on the MP. I'm just somewhat surprised that Google puts my Wikipedia page before other instances of my name (my Flickr feed, for instance, isn't even on the first three pages) — Crisco 1492 (talk) 01:04, 9 July 2014 (UTC)
 * Google simply loves Wikipedia, from A* search algorithm to Z (1969 film). Maybe for the best. When I search "potato" and exclude Wikipedia, my mind is blown. Potatoes shouldn't do that. InedibleHulk (talk) 01:23, 9 July 2014 (UTC)
 * The suggestion by 209 can't be correct. I doubt that InedibleHulk and I are in the same IP locale yet a search for bun gets the same results as him. For the HM searches I get the same results except that the first link for "HM memory" says it is HM on Wikipedia but when clicked on goes directly to Henry Molaison. It's easy enough to put user names with real life names if, like me, you use the same user name on different sites. My real life name is in the first hit on the second page of Google results (it's on my user page sort of). CBWeather, Talk, Seal meat for supper? 00:49, 9 July 2014 (UTC)
 * The HM thing is the same for me, too. Goes to Molaison with no redirection. Hovering over the link shows the URL is right, just the title is wrong. InedibleHulk (talk) 01:03, 9 July 2014 (UTC)

Running a different OS (Linux? Ubuntu?) on a 10 year old laptop
I have a laptop that is about 10 years old and is currently running Windows XP. It has nothing of value saved on the hard disk. I find that XP is very glitchy and would like to completely reinstall it on the laptop. However, it seems pointless installing XP on a laptop in this day and age since it is so outdated. I also do not want to buy a new version of Windows. I would quite like to just install an OpenSource operating system - the two I have heard of are Linux and Ubuntu, but I am sure there are many others available. Firstly, I have only ever used Windows. How challenging is it for someone (who is not very computer adept) to learn how to run a new OS on a computer? Secondly, will I encounter any hardware problems running a different OS on a computer that ran XP? Thirdly, will I have any problems getting software to run on the new OS? All I will really need is OpenOffice, something to play music and film, and the Open Source version of Photoshop (Gimp?). Also, do you have any other advice for me as to how I go about doing this? Hella New Thing (talk) 14:05, 8 July 2014 (UTC)
 * Ubuntu is one flavour of Linux. It's the one I would normally recommend to beginners, but newer versions don't seem to universally play well with old and slow graphic cards. You can try e.g. CentOS or Fedora instead - or try them all (they boot from USB sticks for tryout) and stick with the one you like best. All will have OpenOffice (or LibreOffice), and support for playing and manipulating most modern media. As for the difficulty: From my controlled study of 70 year old non-computer-users, the initial barrier automatically wears down in a week or two;-).   --Stephan Schulz (talk) 14:18, 8 July 2014 (UTC)


 * Ubuntu is probably too much for that machine, particularly for the graphical reasons you mention. Ubuntu variant Xubuntu has lower requirements, and Lubuntu is lighter still. I'd be pretty confident of Lubuntu working well on any machine beefy enough to run XP, and Xubuntu may work okay, depending on the machine's specifications. I doubt the machine will be so puny that it can't run these; if it were, there's a lower-tier of Linux distributions with a fanatical concentration on running on old, underpowered machines - things like Puppy Linux. -- Finlay McWalterᚠTalk 21:02, 8 July 2014 (UTC)


 * VLC_media_player is your go-to media player, and I might recommend Inkscape in addition to Gimp. Whatever *nix you install, make sure you get a feel for the package manager, e.g. Ubuntu uses apt-get. This will make it much easier to install new software. Many of the best open source software packages are hosted on Sourceforge, you can browse through there to find all kinds of fun free stuff. As for ease of learning, I'm not sure if e.g. apt-get has a GUI front end on the Ubuntu distributions, but a little time invested in learning a command line interface like the Bash shell will probably pay off, though this is probably not strictly necessary for normal email/web/media use. SemanticMantis (talk) 17:11, 8 July 2014 (UTC)


 * Ubuntu's current package-management GUI application is Ubuntu Software Center (which looks and works like a somewhat basic app-store); other Ubuntu variants still use Synaptic (but USC works on Xubuntu and Lubuntu, and can be installed with Synaptic or apt-get). -- Finlay McWalterᚠTalk 21:56, 8 July 2014 (UTC)

Challenging browser problem
I want to have continuous vertical scrolling. Sometimes at Flickr or Facebook for example, when you go through a big list with hundreds or 1000 and more items (May it be images or People's lists), The window shows only a tiny bit every time you got to the end of the scroll bar. what I want to have is some kind of an automatic command that when i turn it on, The browser will automatically scroll down till the end of the items list (I don't care how much time it's gonna take)... Do you guys know any way I could achieve this goal? (I have no knowledge in Script writing). thanks. Ben-Natan (talk) 23:30, 8 July 2014 (UTC)


 * Although not a script, you can hold down the End key on the keyboard to keep jumping to the bottom of the page and loading new content. When you see all the content is loaded, you can let go and press the Home key to jump back to the top. Does this fit what you want? --Bavi H (talk) 00:16, 9 July 2014 (UTC)


 * I'm guessing he wants it to smoothly scroll past everything, so he can find what he's looking for quickly, while your method would jump past lots of stuff. StuRat (talk) 00:19, 9 July 2014 (UTC)


 * In my method, the first step (holding down the End key) is just so you can load everything, you aren't able to read comfortably in this step. Once you see the page has reached its true end and no longer loads new content, the second step is to press the Home key to jump back to the beginning. Now you're at the top of the fully loaded page, you can read and scroll at your own pace and the scrollbar will consistently indicate your true position in the fully loaded page. --Bavi H (talk) 00:20, 10 July 2014 (UTC)


 * That should work, but on a huge web page on a slow browser, that could take several minutes before you can read anything. StuRat (talk) 00:41, 10 July 2014 (UTC)


 * True, but Ben-Natan said "I don't care how much time it's gonna take", so I think it might fit what he wants. --Bavi H (talk) 02:18, 11 July 2014 (UTC)


 * Before the middle mouse button created new tabs, it usually started a continuous scroll mode. I know that IE6 did. You could click the middle mouse button and then move the mouse up or down. This should do the trick, without keeping a key or button pressed. (I didn't find that feature useful but it was there. No clue if newer browsers can do it, but it could be a compatibility setting. I wouldn't suggest IE6 for long lists (or long articles for that matter); it performs like O(N^2) with these when the competitors are closer to O(N log N).) - ¡Ouch! (hurt me / more pain) 06:16, 9 July 2014 (UTC)


 * With due respect, you should not use the algorithmic complexity notation to express your frustration with a slow user interface. In actual fact, correct compliance with the HTML document object model requires algorithms that are much worse than O(n2) for an HTML document with n nodes.  A fairly recent archived mailing on the World Wide Web Consortium's CSS list estimates that the algorithmic complexity for constraint satisfaction of the DOM for an n-node webpage is exponential in n : O(  Dn ) for D layout constraints and n document nodes.  Modern web browsers use heuristics to parse the document tree, essentially violating the exact to-the-letter specifications.  Arguably, older browsers are more standards-compliant than newer ones!  But in reality, the performance of the user interface of your browser probably is more limited by implementation details - like choice of graphics library, or efficiency of memory-use - than by theoretical limits imposed by the algorithmic complexity of the layout engine.
 * Here's a recent blog, which was featured in last month's Planet Webkit front page on http://webkit.org : Automatic Grid Placement algorithms with CSS, which is written at a level that most programmers can follow (without too many confounding implementation and theory details). There is much discussion of spec-compliance, and several links to the W3's canonical DOM definitions.
 * Here's a more theoretical guide to constraint programming concepts, Constraint Satisfaction, focusing more on the math (and with no emphasis on its application to solving the HTML or webpage layout problem). It's a great website to help familiarize yourself with the fundamentals.
 * If you enjoy spending a few hours reading about layout algorithms, you might as well read The Java Swing Architecture and the extensive links from that page. Swing, its layout, look-and-feel, and abstract windowing and UI backing layers, were designed to be very algorithmically efficient, especially compared to the HTML DOM.  Here's how layout management worked, with visual examples.  Yet, many people will anecdotally tell you that Java user interfaces are "sluggish" - evidence again that the user experience and latency is probably more limited by practical implementation details than by actual algorithmic complexity.  It is unfortunate that on many platforms, Java's windowing and UI were backed by software-emulation instead of high-performance native libraries; a very well-designed and efficient technology was in many ways lamed by its implementation-details - particularly on important operating systems like Windows.
 * Nimur (talk) 06:33, 9 July 2014 (UTC)


 * Middle clicking on anything that's not a link in a webpage will still give the "move mouse to scroll" behaviour (at least in Chrome and IE9, which are all I have access to at work). MChesterMC (talk) 08:26, 9 July 2014 (UTC)

Web page loading
I've never been happy with the way web pages load. There are several problems:


 * 1) They don't always load from the top down. So, you can't just start reading it as the rest loads.


 * 2) Pics don't load at their full size initially. I'd like a frame to appear at the proper size, then fill in the detail later.  This would prevent the text from jumping around as pics above the text are resized.


 * 3) They often load a limited amount, then wait for you to get to the bottom, and possibly hit a page forward button, to load the next page. This leads to a lag while you wait for the next page to load.  I realize they probably can't load all the pages, but loading the next page in background, after the current page is loaded, would make sense.


 * 4) Scrolling often doesn't work properly while it's loading a page. Either it doesn't scroll at all or scrolls in jumps and starts.  They need to allocate more resources to scrolling so this doesn't happen.


 * 5) Video often starts without me clicking a play button. This is annoying and wastes valuable resources.

I realize that much of this could be fixed by faster computers and internet connections, but that's no excuse to do things in an inefficient way. So, is there any browser which will load web pages as I've outlined above ? StuRat (talk) 16:55, 9 July 2014 (UTC)


 * WOW, lot's of information was given... any humble summing of a few words --- what exactly should one do? (please don't explain me of computer sciences, I am not familiar with this field). Ben-Natan (talk) 06:00, 10 July 2014 (UTC)


 * StuRat, what kind of connection do you have?! It sounds almost as if you are using something below 56k speed. Also, what kind of computer are you using?
 * 1) If it was plaintext, I might understand, but it's HTML, which needs to load the entire page to discover how it's going to render it completely.
 * 2) Pictures should be loading with their correct size even if they aren't completely loaded yet. What browser are you using?
 * 3) Resources. Why load something when you don't know if the user is going to look at it?
 * 4) Shouldn't happen. Again, what browser?
 * 5) I can agree here, except for cases like YouTube in which you're highly likely to want to watch the video whose page you just loaded.
 * Thanks. -- 140.202.10.134 (talk) 17:18, 10 July 2014 (UTC)


 * 3) Well, if I just read pages 1-5, it's a reasonable bet I'm about to read page 6, isn't it ? I can possibly see not pre-loading page 2, since most people who look at page 1 of a document may never go on to page 2.  It also somewhat depends on how sequential the data is, though. StuRat (talk) 14:10, 11 July 2014 (UTC)


 * 4) I use Firefox on my old Windows 98 PCs, and Chrome on my Windows 7 PCs. I also have Opera and IE, but tend not to use them much. StuRat (talk) 14:07, 11 July 2014 (UTC)


 * Just some qualifying points:
 * 2) Pictures: this is dependent on the person who coded the site. If they have not correctly specified the size of the image in their code, the browser cannot correctly size the placeholder until the image has been loaded. Well-written sites should not have this problem.
 * 3) Loading ahead: many browsers do preload the next page etc (e.g. Chrome). You can sometimes turn this setting on and off.
 * —Noiratsi (talk) 17:32, 10 July 2014 (UTC)


 * Where do you turn on the pre-load feature in Chrome ? StuRat (talk) 14:07, 11 July 2014 (UTC)


 * See this support article from Google. It is under Seetings > Advanced > Privacy > Predict network actions to improve page load performance. —Noiratsi (talk) 06:28, 12 July 2014 (UTC)