Wikipedia:Reference desk/Archives/Computing/2014 December 16

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

December 16

Can I install two operating systems on one internal 3.5 inch hard disk

That's my question. Thanks --AboutFace 22 (talk) 01:14, 16 December 2014 (UTC)

That depends on the operating systems but it is possible. See boot loader. Dismas|^(talk) 01:18, 16 December 2014 (UTC)

One OS is Windows Server 2008 and the other one, that I want to install also on the same disk is Ubuntu. --AboutFace 22 (talk) 01:53, 16 December 2014 (UTC)

Both O/S's would need to support the format used on the hard drive. And there might be other problems, like different O/S's using different paging space on the hard disk. I certainly wouldn't recommend it, but it might be possible. StuRat (talk) 02:38, 16 December 2014 (UTC)

This is certainly possible. You need to install the two operating systems to separate partitions within the single physical disk. Ubuntu has a good guide for going about this here. GoldenRing (talk) 03:28, 16 December 2014 (UTC)

Yes, Win first, then Ubu. But troubleshooting will be more complicated. Swapping 2 drives is less complicated. --Hans Haase (有问题吗) 12:12, 16 December 2014 (UTC)

I will second that. A second drive is well worth the money (and easy to instal, plus more storage, plus a convenient place to back up). <Linux plug>Also, you will then realize that of your precious time spent on maintenance and trouble shooting, is nearly all wasted on the drive hosting Microsoft. If window was released for the first time today, it wouldn’t stand a chance against the modern Linux family of no hassle distros. </Linux plug>--Aspro (talk) 22:58, 16 December 2014 (UTC)

Hosting English Wikipedia for a group of Users

I have installed MediaWiki 2 years ago.But since MediaWiki can't handle more than 1 GB(not sure) I changed to XOWA app.But using XOWA for a single user might be good but when it comes to a group of users I think MediaWiki might be more efficient since online wikipedia uses this software.Could you tell me are there any memory constraints for MediaWiki(asking whether it can handle a wikipedia dump of 10GB)?I once tried to install the dump by increasing the memory limit in php configuration file(apache too) in WAMP server but I think it can handle around 1GB as I tried but when it comes to the dump file(10GB) it usually gives some error.Since I couldn't get a lot(memory) from MediaWiki should I need to install any other software?Could anyone help me.JUSTIN JOHNS (talk) 03:52, 16 December 2014 (UTC)

What limit are you talking about, the media upload size limit?

Do you know about Wikipedia Zero and the project wiki on a stick or wikipedia offline? --Hans Haase (有问题吗) 12:16, 16 December 2014 (UTC)

@Hans Haase:No.I'm taking about the memory limit in MediaWiki for uploading(MAX_UPLOAD_SIZE in MediWiki).Also 'Wikipedia Zero' is an mobile application.I'm really asking about an software used in computer to allow a group of users access the wikipedia in a network.I have asked this question previously at :https://en.wikipedia.org/wiki/Wikipedia:Reference_desk/Archives/Computing/2014_December_10#Hosting_English_Wikipedia_in_College.

Could you tell me is there any constraint in uploading a 10GB file(after uncompressing 40GB) into MediaWiki.Is there any other software that might be good to host an English Wikipedia for a group of users.I know about XOWA but using it individually might be good but when it comes to a group of users the addon(firefox) it provides doesn't work and the http connection is slow too.JUSTIN JOHNS (talk) 04:33, 17 December 2014 (UTC)

Trying to pronounce ð and θ

Hi there,
I've tried to pronounce those consonants,
I would like to hear comments about my pronunciation,
Beginning with accent, and particularly the pronunciation of the THs.
It should be "they think".
http://picosong.com/4MHE/
Another issue is that I noticed that most of the native-English speakers, put their tongue out.
What I do, is putting my edge of the tongue on my front teeth,so another part of the tongue is sliding out from the "teeth box".
It's like my tongue is in a C-like shape. Is it okay?
15:14, 16 December 2014 (UTC) — Preceding unsigned comment added by 84.229.104.133 (talk)

If you were a computer, your post would have been justified here, otherwise, perhaps the Help desk would be the best place for it? --AboutFace 22 (talk) 15:44, 16 December 2014 (UTC)

You mean the Language Ref Desk, I assume  ? If so, I agree, unless they want an audio file or software tool to help them pronounce it (although they did mention C !). StuRat (talk) 16:40, 16 December 2014 (UTC)

The sound is faint for me, but I hear nothing wrong with it. —Tamfang (talk) 23:11, 16 December 2014 (UTC)

TV color adjustment

Faces on my TV have yellow patches on them. I have the following color controls for each color (R, G, B): brightness and contrast. So, since there's no option to turn yellow down, which sliders should I change ? It's an LCD TV with LED backlight, if that matters.

Thanks,

StuRat (talk) 17:32, 16 December 2014 (UTC)

The RGB color model is additive, and yellow is created by adding red and green - you should either decrease both, or increase the blue. AndyTheGrump (talk) 17:44, 16 December 2014 (UTC)

Since they are only yellow patches, not the whole image being too yellow, does that mean I should turn up the blue contrast and/or turn down the red and green contrast ? StuRat (talk) 17:53, 16 December 2014 (UTC)

Have you tried a test pattern? Many modern TVs have them built in, or you could pipe one in from a computer. I am suspicious that only patches of faces seem off. Other issue: what is the input signal? Broadcast HDTV? DVD? etc. It could be a problem with the input and not the display. The closest thing I can think of is when an old film noir movie has been poorly converted to DVD. When the MPEG encoding/compression isn't done right, you can clearly see clean squares of dark purple or maroon in what should be smooth deep shadows. Anyway, from your description, it's not clear to me that simple color adjustment will solve the problem, though it may help or at least make it less noticeable. SemanticMantis (talk) 18:35, 16 December 2014 (UTC)

The yellow patches aren't blocks, the seem to be areas of the face that maybe should be a bit more yellow, but are much more yellow instead. I turn up the blue contrast to max and turned down the red and green contrast to min. That certainly reduced the problem. BTW, this is broadcast HDTV (1080i). StuRat (talk) 22:03, 16 December 2014 (UTC)

I don't feel it's a good idea to make an adjustment to fix a specific problem; in my experience, that just creates another problem. Your problem might not even be in the color settings. I think a more "scientific" or "holistic" approach, for lack of a better word, works better over all, and there are a lot of web resources for "tv picture calibration". I was lucky with my current TV, a low-end HD LED model. Consumer Reports published a full set of picture settings (not only color) recommended for that model by their engineers/testers. I tried their recommendations and after two years I've been extremely happy with the results. ‑‑Mandruss ☎ 19:39, 16 December 2014 (UTC)

LOL, it sounds like it took two years to do the color calibration. StuRat (talk) 22:03, 16 December 2014 (UTC)

Eh? I don't know how long it took the engineers to determine the settings. It took me 5 minutes to set them and several months to decide that I was extremely happy. ‑‑Mandruss ☎ 22:08, 16 December 2014 (UTC)

Generating lists from Excel

I have something similar to the following table in Excel and I'm wondering if there is a way to generate lists from it in Excel or Word, such as, a list of Mr. Smith's math students, a list of Ms. Jones's science students, etc.

Name of student	Class 1	Teacher 1	Class 2	Teacher 2	...
Alison	Math	Mr. Smith	Science	Ms. Jones
Barry	Gym	Mr. Fisher	Art	Mr. Fraser
Charles	Math	Mr. Smith	Art	Mr. Fraser
Dawn	Gym	Mr. Fisher	Art	Mr. Fraser
Edward	Gym	Mr. Fisher	Science	Ms. Jones
Fred	Drama	Ms. Taylor	Art	Mr. Fraser
Greg	Drama	Ms. Taylor	IT	Ms. Carter
...					...

Carveroa14 (talk) 21:48, 16 December 2014 (UTC)

A simple way to do it in Excel is to use the Auto Filter function: filter "Teacher 1 = Mr. Smith" and "Class 1 = Math", then cut and paste the result into a new worksheet or into Word. --Canley (talk) 00:33, 17 December 2014 (UTC)

Yes and to export the sorted list should not be the problem. An other solution: Mark the whole table. In menue, select data, sort data and specifiy the sort order of the colmns. Copy an paste the result. If you do not want to change the excel file, close and don't save it. --Hans Haase (有问题吗) 16:31, 17 December 2014 (UTC)

Minimize Backtracking Algorithm

So today my teacher was talking about Depth-First Search and Breadth-First Search. He was saying how when finding the path you have to backtrack for DFS and go to the next available branch. When he finally had found the tree his tree seemed mangled and had multiple sections with non-sensical back-trackings (degree 1 nodes in the final tree). My question is "What is the easiest way to find a tree such that the minimal amount of traversing is required to pass through the tree. (explanation--when picking a new location how can I continue the line I am on without creating a dead end before passing over all nodes/vertices). Idk if this question has already been answered so if it has just post a link. Thnx -- Jetstream5500 (talk) 22:37, 16 December 2014 (UTC)

I think this is the "minimum leaf spanning tree" problem, which is NP complete since it has the Hamiltonian path as a special case. NP-complete problems aren't always intractable in practice, but this one doesn't even have a constant-factor approximation (i.e., a fast way to get within a factor of N of the minimum amount of backtracking). That probably means that a backtracking-minimized DFS would be much slower than an ordinary blind DFS. -- BenRG (talk) 00:24, 17 December 2014 (UTC)

Could you make an algorithm that could give not the correct answer but a possible answer that is 95% correct or would this still be difficult. In more general terms: Is there an algorithm that could make several spanning trees with 0 backtrackings and then somehow connect them to result in a fairly good answer.Jetstream5500 (talk) 03:15, 17 December 2014 (UTC)

Based on the info you've given, it sounds like you have no way of knowing when you are close to the target state. In that case there's really no efficient way to get there, other than an exhaustive search.

But, let's say you were solving a maze, and the desired state is being at the exit. In that case, you could use the Manhattan distance from your current location to the exit as an approximate measure of how close you are. Of course, it's not always correct, because there could be a wall in the way even though the distance is less. But this would allow you to prioritize search down branches with lower distances, and usually make your search more efficient.

I've also found that searching from both directions is more efficient yet. So, while you are branching from the maze entrance towards the exit, also branch from the exit towards the entrance, and when your two trees meet in the middle, that's your solution. StuRat (talk) 03:49, 17 December 2014 (UTC)

Could you break apart the maze anyway? When 2 graphs are connected by a bridge you can run the program on both sections thus starting in 4 locations right? This would be even quicker. --Jetstream5500 (talk) 04:48, 17 December 2014 (UTC)

Yes, but you'd need to know an intermediate state you will need to pass through (the "bridge") in order to break it down further. I was assuming the only two states which are known for sure are the start and end states (entrance to the maze and exit). Now, you could also have a system with multiple starting and ending states (entrances and exits, in the case of a maze). If the goal is to find the shortest path between any starting state and any ending state, then simultaneously building trees from each of those might also make sense. When the first tree from a starting state intercepts the first tree from an ending state, that's your solution.

Oh, and if you haven't already discussed pruning algorithms, the obvious one here is to prune off any branch that leads to a state that's already in that tree. In the case of a maze, that means we don't go back to any place in the maze we've already visited.

For a real-world example, lets say we are writing a program to find the shortest path between a start location and destination on a map (like Google Maps). We'd want to start a tree at each end, and also start trees at intermediate stops, if any. (In practice we'd completely solve one segment before starting the next, unless we are using some type of parallel computing architecture that would make solving all segments simultaneously faster.) We'd also want to weight each step by if it moves you closer or father from the goal state (location), and explore those nodes that get us closer first. As in the maze example, this is no longer guaranteed to find the best solution, because a route which initially takes you farther from the goal state may actually turn out to be the best (here perhaps that might mean driving to the nearest highway entrance, even though it's out of the way). Weightings could also be altered based on type of road, etc. StuRat (talk) 06:01, 17 December 2014 (UTC)

If you actually want to find a goal node quickly then A* search is often a good way to do it, but in this question there is no particular goal node. The goal is to explore the entire graph without backtracking. "Pruning off any branch that leads to a state that's already in that tree" is how DFS always works. I get the impression that you've never taken the standard algorithms course that Jetstream5500 is taking. -- BenRG (talk) 07:41, 17 December 2014 (UTC)

I just wasn't sure if he'd gotten to that part of the class yet. I took the class a while back, but didn't recall if they taught the pruning right up front, or first teach the difference between DFS and BFS and then add in pruning methods. StuRat (talk) 07:49, 17 December 2014 (UTC)

Im in a Discrete Math course in high school. Now I was wondering you can't find a path that minimizes back-tracking and you said this is similar to hamiltonian paths and cycles. What I was wondering is can you (like hamiltonian paths and cycles) determine if such a path through the nodes exists w/o actually knowing the exact path that would need to be taken. I like the idea of pruning and the above mentioned algorithms also. Also if you have 2 bridges to the same point you can eliminate a edge that passes between 'b' and 'c' if 'a' is part of graph 1 and 'b' 'c' are part of graph 2. Thank you for the responses. This problem makes a lot more sense now. Close approximations are the best you can do unless a new algorithm / idea comes along on hamiltonian paths/cyclesJetstream5500 (talk) 20:07, 17 December 2014 (UTC)

As far as a pair of bridges go, there are some different possible arrangements. One is if you can cross one or both bridges, as in two shores on opposite sides of a river with 2 bridges across it, where you start on one side of the river and must finish on th other. In this case the bridge just behaves like any other link between nodes. However, I assume you meant a case where there's an island in the river, and one bridge from the island to each shore, so that you must take each bridge exactly once to get from one shore to the other. In this case the problem breaks down to 3 distinct problems to be solved, optimizing the left shore navigation, island navigation, and right shore navigation.

BTW, edX offers an excellent class on Artificial Intelligence through the University of California at Berkley which covers this and much more, online, for free, the last time I checked. StuRat (talk) 10:46, 18 December 2014 (UTC)