Wikipedia:Reference desk/Archives/Computing/2013 May 15

From Wikipedia, the free encyclopedia
Computing desk
< May 14 << Apr | May | Jun >> May 16 >
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.


May 15[edit]

Software for scanner[edit]

I'm looking for scanning software. The software that came with my HP scanner will not save any settings (at least not the ones that matter). I use the scanner to make PDFs, to scan photos and documents to files, and to act as a photocopier. What can fit that bill? Bubba73 You talkin' to me? 04:25, 15 May 2013 (UTC)[reply]

software to crop PDFs[edit]

I'm looking for software to crop PDFs that are scanned in. The ones I make from my scanner have a lot of stuff around the edges that are outside the original document - I'd like to eliminate that. (Some scans are of multiple pages.) Bubba73 You talkin' to me? 04:28, 15 May 2013 (UTC)[reply]

I use BRISS - GNU GPL http://sourceforge.net/projects/briss/ It's not as intuitive as claimed but really easy once you understand the process. The software merges/overlays all the pages in the document allowing you create a single crop area to be applied to all the pages. This is not useful to me so I need to tell the software to exclude page 2 to the end of the document from the merge. You can then specify a crop rectangle for each page. At about 9megs, free and no install required I love it. 196.214.78.114 (talk) 10:29, 15 May 2013 (UTC)[reply]
Resolved
I tried it on a one-page PDF and it worked pretty well, although the cropped file was the same size. But size isn't important - except when I did it in photoshop, the resulting file was more than 10 times as big. Bubba73 You talkin' to me? 16:59, 15 May 2013 (UTC)[reply]
Just a remark: As I understand PDF, "cropping" does not usually remove any image elements, it will just mask them out. If you have a broken renderer, or a malicious hacker who wants to see the hidden part, it can be, in principle, recovered. It's very simple to suppress pixel output outside the cropped area. It's much harder to actually determine which PDF elements are fully outside the visible area, and even harder (for the "optimal" case) to replace partially visible elements by equivalent ones that have the same effect in the visible area. Most people and programs go for "easy". --Stephan Schulz (talk) 12:50, 19 May 2013 (UTC)[reply]
The part I'm cropping out is the part of the scanner beyond the actual page being scanned. Bubba73 You talkin' to me? 18:22, 20 May 2013 (UTC)[reply]
I've primarily offered the comment as an explanation why the cropped version is usually not smaller than the original - the content is (often) still there, just invisible. --Stephan Schulz (talk) 12:57, 21 May 2013 (UTC)[reply]
Thank you, and just being not seen is sufficient for my purposes. I just want to not see the ugly, non-document part of the scan. Bubba73 You talkin' to me? 17:08, 21 May 2013 (UTC)[reply]

Break havoc in Linux without admin password[edit]

I understand that you cannot install programs without being admin, but isn't it a little bit insecure to be able to run scripts like rm whatever, or others, without any password? OsmanRF34 (talk) 13:22, 15 May 2013 (UTC)[reply]

You should not be able to remove a file for which you do not have write-permission. File permissions should be set correctly and managed by an administrator as part of a complete security policy. Nimur (talk) 13:40, 15 May 2013 (UTC)[reply]
You can only remove files from folders to which you have write permission. In practice, for an ordinary user on a normal config, that mostly means you can only remove stuff from your own home, and stuff you created in /tmp. It's also not entirely true that one has to have superuser permissions to install programs, because it's quite possible to install a program for a single user - in which case stuff only goes into that users' home. Unfortunately many programs' idea of being installed does envisage their config files being in /etc, so this isn't as true as it could be. -- Finlay McWalterTalk 13:41, 15 May 2013 (UTC)[reply]
http://xkcd.com/1200/ πr2 (tc) 22:30, 17 May 2013 (UTC)[reply]

md5sum of installed program in Linux[edit]

I know that when you download a program you do an md5sum check on the installation file, but can you do the same on already installed files, to check if they were not tampered with? OsmanRF34 (talk) 13:24, 15 May 2013 (UTC)[reply]

GNU coreutils (which is installed on just about every Linux system) includes a command line md5sum binary, so one can obtain the md5sum of any given file. One can verify that against what its source says it should be, if the source publishes that info somewhere. -- Finlay McWalterTalk 13:44, 15 May 2013 (UTC)[reply]
You can compute the checksum for any file on which you have read permissions. If this is part of a tamper-evidence security policy, there are other concerns: a mismatching md5 sum might not indicate malicious tampering. The whitelist of approved checksums might be large and must be secured. The integrity of your md5 program must be secured. These problems are part of the field of trusted computing., and application whitelisting and file integrity verification are common strategies applied as part of a total security solution. Nimur (talk) 13:45, 15 May 2013 (UTC)[reply]
"rpm -q --verify PACKAGENAME" does exactly that. --Sean 16:59, 15 May 2013 (UTC)[reply]
On Debian (apt) systems like Mint, Crunchbang, Knoppix, and Ubuntu, one uses debsums PACKAGENAME for this. -- Finlay McWalterTalk 09:58, 16 May 2013 (UTC)[reply]

Learning Data digitization[edit]

HI all, can anyone please help suggest me a website or any resource to help me learn Data Digitization ,any info or help would mean a lot as i tried searching online but was to no avail,.thanking you all in advance.regards jane — Preceding unsigned comment added by 117.199.64.5 (talk) 14:37, 15 May 2013 (UTC)[reply]

the reason your search does not succeed is that the term doesn't really mean anything definite. Are you trying to learn about analog-to-digital converters? Looie496 (talk) 16:36, 15 May 2013 (UTC)[reply]
The OP may be referring to digitization of data as in the digitization of documents etc. --Yellow1996 (talk) 01:41, 16 May 2013 (UTC)[reply]

3D maps[edit]

If I have a map of a particular area, with contour lines showing different heights, or shaded so each colour represents a different height, is there any way of converting that into a 3D model just by feeding the data into a computer program, rather than having to go to the trouble of sculpting every individual point on the surface? 213.104.128.16 (talk) 16:16, 15 May 2013 (UTC)[reply]

See "3D printing".—Wavelength (talk) 16:31, 15 May 2013 (UTC)[reply]
If your map is in the form of an image that shows labeled contour lines, then this is quite a difficult problem, and I don't think you'll find a general-purpose solution. If the color code can easily be decoded, then that's a much simpler situation, but it would still be necessary to write a special-purpose program to do the job. Looie496 (talk) 16:39, 15 May 2013 (UTC)[reply]
There may be a Hackerspace in your area with a 3D printer. They may be able to help convert it to a printing format as well. http://hackerspaces.org/wiki/List_of_Hacker_Spaces has a large list.--Canoe1967 (talk) 17:46, 15 May 2013 (UTC)[reply]
OP, do you want a physical, concrete model? All the responders so far seem to think so, but let us know if you just want a 3D software model. In either case, some data extraction program (such as data thief [1] would probably help convert your contour lines to a XYZ mesh. SemanticMantis (talk) 18:05, 15 May 2013 (UTC)[reply]
Inkscape can also vectorize raster-images using the Inkscape Potrace feature. Once your contours are vectorized, you can proceed to migrate them to a more convenient 3D modeling format; and if you actually do intend to print or cut a physical version of the model, you can convert to an appropriate input format for your printer or cutter tool. Nimur (talk) 18:26, 15 May 2013 (UTC)[reply]
If you mean a "software" model, then I think what you are after could be fairly easily acheived using a Heightmap, the article has links to software you can use. Vespine (talk) 23:02, 15 May 2013 (UTC)[reply]
Yea, I did mean a 3D computer model of the map, I have a 3D printer that could then print that out easily, though it would be too small to be much use for anything, I think. What I have so far is a map where each layer is a different colour for a specific height. Though, I'm thinking, that would make it go up in steps rather than smoothly, once I have a program that can turn it into a 3D form, would it be possible to set it so a certain edge of each space is that specific height, and then slope down from there? 213.104.128.16 (talk) 23:03, 15 May 2013 (UTC)[reply]
If you want something bigger than a 3D printer you can move a small drill using a screw driven by a stepper motor for the z in a big xy plotter arrangement and shape the profile out of polystyrene. It is easy enough to program the intermediate heights to be a reasonable but usually I'd have though you'd have started with the measured heights of various points in the area. Dmcq (talk) 00:09, 16 May 2013 (UTC)[reply]
Did you look at the heightmap article? I played around with height maps using the farcry level editor (a computer game). If you can make your image grey scale with white value representing height, you can even use photoshop tools to edit the image however you like. For example, if you have distinct steps but want a smooth gradient, you can simply use the blur tool on the lines until they are a smooth gradient, or you can even try the gradient tool if you work out how to use masks and stuff. It will take a bit of playing around depending your photoshop skills and how nice you want to get the result, but it's definitely not the most impractical way of doing it. Vespine (talk) 04:12, 16 May 2013 (UTC)[reply]
There's various interpolation techniques but if you want to do something simple yourself scan in a map. Set up a binary array of where the contour colour changes. Go across all the lines of the image linearly interpolating between successive points like that. Then get the average with doing the same in the columns then finally do a few sweeps of averaging the four closest points with a point except where the original binary array is set. That should get you a pretty good linear smoothed version just by scanning in a map. Dmcq (talk) 11:29, 16 May 2013 (UTC)[reply]
I was thinking more setting it to interpret certain edges of pixels rather than their whole width as a set height. Thing is, it's already a 60 megapixel image, in which some height differences are a single pixel wide as it is, extending that to a sufficient resolution and then interpolating between every single line of that seems like a huge job. 213.104.128.16 (talk) 14:59, 16 May 2013 (UTC)[reply]
Well I think I'd use something faster than Visual Basic for that ;-) But computers are quite fast enough and big enough to crunch through all that with the way I said. Dmcq (talk) 11:31, 17 May 2013 (UTC)[reply]