Wikipedia:Reference desk/Archives/Computing/2011 February 16

Computing desk
< February 15 << Jan | February | Mar >> February 17 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


February 16 edit

Did IBM Watson take the lead tonight on Jeopardy! ? edit

I missed the show. Thanks. 76.27.175.80 (talk) 01:24, 16 February 2011 (UTC)[reply]

Usually the relevant articles for such things are updated quite promptly with current events: See Watson (artificial intelligence software). Short answer: it ended the first match with over three times the amount of the next contestant, although it botched the Final Jeopardy question (In the category "U.S. Cities", it answered "Toronto"), losing only $947. -- 174.21.250.120 (talk) 05:24, 16 February 2011 (UTC)[reply]

Error while calculating Windows Experience Index in Windows 7 edit

I am using Windows 7 and when I try to calculate the WEI, I get the following error message:

Cannot complete assessment. The assessment or other operation did not complete successfully. This is due to an error being reported from the operating system, driver, or other component.

I tried updating all drivers by going to the Device Manager. I also do not have Kaspersky antivirus installed (I learnt from a forum discussion that Kaspersky might be the cuplrit). But I still have the error. Due to this error, I am not able to apply the Aero effects onto my system. How can I calculate the system rating? Please help me. Thank you very much. —Preceding unsigned comment added by 125.16.180.5 (talk) 07:51, 16 February 2011 (UTC)[reply]

extract text from PDF edit

Hi I have a PDF file with a lot of tabular, columnar information. There are columns with characters and some with numbers. I would like to extract all of this into text, but when I choose 'save as text' it saves all the text separated by spaces... so that's pretty useless because I don't know which column it belongs to. I was thinking about selecting a column at a time and copy/paste but the alt-left-click doesn't allow scrolling down the file. Any ideas on how to get the PDF into fixed-format text or even better, comma-delimited? Sandman30s (talk) 07:54, 16 February 2011 (UTC)[reply]

It depends on how the PDF was created. Sometimes it records structural information, sometimes not. My first run at it would be to use a PDF->HTML converter that would make column extraction easy. Failing that, I would look for or write some software that would guess at the columns based on where the bits of text are. --Sean 14:11, 16 February 2011 (UTC)[reply]
If the text is separated into columns by spaces, would the Text-to-Columns feature in Microsoft Excel work? Kingsfold (Quack quack!) 14:49, 16 February 2011 (UTC)[reply]
I suggest opening the PDF in Inkscape. You will likely have better access to getting the columns of text out of the text box itself than the document viewer. -- kainaw 14:57, 16 February 2011 (UTC)[reply]
If it is not too many tables I usually copy-and-paste the text into Notepad. Then it is easy to see if there is really a tabular structure to the text. If there is, and you are not lucky and get tab characters between the columns, I usually remove any spaces within fields manually (for example, "United States"-->"UnitedStates". Then, I copy and paste the data into Excel, where the "import text" feature comes up automatically and I can choose space as a column separator. Not too elegant, but always fastest for small operations. It's hard to standardize this in any case as the structure of PDF documents can differ a lot. Jørgen (talk) 20:25, 16 February 2011 (UTC)[reply]
Thanks for all the help. I managed to find a PDF to Word converter and got it into a program-readable CSV thereafter. To Jørgen, no, the file was way too big to use the method you described. Sandman30s (talk) 10:27, 17 February 2011 (UTC)[reply]

VuDroid edit

The pdf pages overlap vertically. Can you help me? --83.103.117.254 (talk) 08:03, 16 February 2011 (UTC)[reply]

Probably the best thing to do is to file a bug with that project here. You should link to or attach the document in question. --Sean 13:58, 16 February 2011 (UTC)[reply]

I have planned to buy the GPS monitor shown here.I am not subscribed to any GPS service or satellite etc. Will the monitor still work for me ? ( I am in south of (India's) Punjab)  Jon Ascton  (talk)

You do not need to subscribe or pay for GPS service. The satellites are already in orbit, and they broadcast their signals free of charge. Some users, particularly the military, use different signals (or, the same signals with different decoding technology that isn't available for sale to the general public), to obtain better quality position information - see military GPS signals - but anyone can receive all of the signals. Nimur (talk) 17:46, 16 February 2011 (UTC)[reply]
I don't believe that the miltary get better positional information - this was the case with Selective availability but this is no longer used. --Phil Holmes (talk) 10:31, 17 February 2011 (UTC)[reply]
However, some GPS mapping providers charge for updates to maps when new roads are built, and for live traffic information, to help you avoid the jams. Others do the routing on a central computer, instead of on the device (Android Maps for one). For any of these services, you'll need to take out the subscription. CS Miller (talk) 19:30, 16 February 2011 (UTC)[reply]

How does IBM's Watson choose the categories and dollar amounts on Jeopardy? edit

I haven't been able to find any answer or reference for this question, so hopefully someone can provide the answer or direct me to the right resource. Watson's algorithms for answering the questions themselves have been nicely discussed on many sites, including IBM's own Watson site and the recent PBS NOVA special. However, I haven't heard any explanation of how Watson chooses which category to select, nor how Watson chooses the dollar amount to select within the category. Is Watson choosing by itself? If so, how? Or is a human "behind the scenes" making the selections for Watson? Similarly, how is Watson choosing the dollar amounts to bet in both Double Jeopardy and Final Jeopardy? The amounts that Watson bet on the show last night were very strange--almost random, and not something that a human would choose. Just curious if there was some sort of coded-in strategy for this aspect of the game. --Zerozal (talk) 17:43, 16 February 2011 (UTC)[reply]

All the public information about Watson is available from IBM at the IBM Watson website. I haven't watched all the videos explaining how he works, but I would posit that estimating confidence in an answer is easy, based on statistical results of earlier answers - so to pick a category, Watson can select the category he is most confident in. Or, Watson may use an elaborate game theory model to defer using "preferred" categories until a more optimal time. Nimur (talk) 17:49, 16 February 2011 (UTC)[reply]
I looks like they do some basic game-theoretical stuff. If my calculations based on the information at j-archive are correct, Jennings' total score could have maxed out at $41200, had he wagered everything (and gotten it right, which he did), and Watson's wager was such that, had it lost, he'd've had totaled $41201. Of course, the reason Watson had to wager a wacky number to make that come out right was because it had a wacky score from Double Jeopardies where it wagered wacky amounts. Probably that's because they've worked out some kind of function that takes as input the Watson's estimated likelihood of getting the answer right, and then they round it to the nearest dollar. Paul (Stansifer) 02:35, 17 February 2011 (UTC)[reply]
This post describes Watson's wagering strategies for both Daily Double and Final Jeopardy questions. --LarryMac | Talk 17:59, 16 February 2011 (UTC)[reply]

My laptop runs Windows Vista, and whenever I want to view photos in Thumbnail mode, I am unable to see previews; I have to open each individual file to see my pictures. Do I need to change a setting to see the previews? Alternatively, is there a third-party application that I can install? Currently all I see is a generic view of an island. Hemoroid Agastordoff (talk) 17:55, 16 February 2011 (UTC)[reply]

In "Folder options" (the view tab), the option to "Always show icons, never thumbnails" needs to be unchecked. Showing thumbnails can slow down the browsing of folders, so this option is provided for slow computers. Dbfirs 09:45, 17 February 2011 (UTC)[reply]
Thanks, Dbfirs, you've made my day ! Hemoroid Agastordoff (talk) 15:31, 17 February 2011 (UTC)[reply]

Standby edit

  Resolved

Is there a way to make a computer running Windows 7 go into standby or some other similar mode where the hard drive and processor etc power down, but keep the screen on and displaying something? 82.43.92.41 (talk) 23:00, 16 February 2011 (UTC)[reply]

No, the processor is needed to generate the image you see on your screen. Looie496 (talk) 23:12, 16 February 2011 (UTC)[reply]
(ec) No. Even with fancy ACPI power management technology, and even with smart graphics processors, the CPU is still required to put pixels to the screen (or to command the GPU to do it by proxy). There are no "personal computer"-style machines that can power down the CPU and still keep an image on the display screen. Some e-book readers may have a capability to do what you want. Nimur (talk) 23:13, 16 February 2011 (UTC)[reply]

Ok 82.43.92.41 (talk) 23:15, 16 February 2011 (UTC)[reply]

Well, technically, most new monitors display something when they are powered on, but do not receive a signal. For example, something like this: [1]. decltype (talk) 09:36, 18 February 2011 (UTC)[reply]
True; technically, most monitors contain a fully configurable and programmable computer inside them, too: a Video Display Controller resides on the monitor-side of the connection, and buffers, and then decodes, VGA or DVI or HDMI signals. Ironically, this means that in a modern PC, using a VGA connection to an LCD monitor, the signal is digitally generated by software on the CPU, copied to RAM, sent through a peripheral bus like PCIe, sent to a GPU bus controller, copied to video ram, converted to an analog signal by a RAMDAC, sent in analog format to a video controller, re-digitized/resampled, re-buffered, and finally displayed on the screen. And this doesn't even count video/GPU processing. Every single pixel of every frame is passed through this pipeline when the CPU is in control of the screen buffer. Most users, running most operating systems, on most hardware, are not able to micromanage the entire video-pipeline, but conceivably, a skilled programmer with the appropriate driver toolkits and reasonably configurable hardware, could subvert the "conventional flow" to perform a task similar to what the original questioner wanted. Frankly, such technology baffles me, and I wouldn't want to be the person who has design it. Nimur (talk) 22:02, 18 February 2011 (UTC)[reply]
Actually this hits something I was thinking of. Doesn't this really depend on the monitor not th computer? I don't know if such a monitor exists but I could imagine one with a USB connection and some software on the PC you use to tell the monitor what to display in standby mode. Nil Einne (talk) 13:21, 19 February 2011 (UTC)[reply]