Wikipedia:Reference desk/Archives/Computing/2009 November 27

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.

November 27

GPS-based alarm

I would like to know if there's a program that could alarm you when it reached certain point on the map. Suppose I am sleeping while taking a train to a point B and I would be awoken in point B-1km. Have anyone heard of such a program? Thanks... roscoe_x (talk) 01:27, 27 November 2009 (UTC)

On what sort of system are you asking this for? It is a mobile phone, an iTouch, a computer with an attached GPS device? Magog the Ogre (talk) 18:24, 27 November 2009 (UTC)

I almost wonder if you're fishing here because you've described exactly an app that's on the iphone but whose name I can't remember right now. It's called iNap, and read this for more. I don't think it somehow makes you get a GPS signal underground.... Shadowjams (talk) 09:06, 28 November 2009 (UTC)

Difference/s with Suse, Ubuntu, etc along those lines

Basically, it was in one of the answers to my previous question on here. —Preceding unsigned comment added by Jessicaabruno (talk • contribs) 04:25, 27 November 2009 (UTC)

Are you asking what the difference between Linux distributions is? If so, it is primarily the package managers. Choose a package manager you like. Pick the popular distribution for that manager. For example, APT and YUM are two very popular package managers. If you like APT, you will likely choose a Debian-type Linux like Ubuntu. If you like YUM, you will likely choose a RedHat-type Linux like Fedora. -- kainaw™ 05:07, 27 November 2009 (UTC)

Distributions also seriously vary the license terms, so pay close attention to those if you have special commercial or legal needs. Most home users and many commercial uses are fine with "free", and need no further detail. Nimur (talk) 15:35, 27 November 2009 (UTC)

SUSE uses .rpm files to install programs and uses Zypper (not YUM) to install them via the command line. It uses YAST to install them graphically. It has traditionally used KDE as its interface, but you can also choose GNOME during installation. SUSE has been around since 1992.

Ubuntu is based on Debian, so it uses Aptitude as the package manager and .deb files to install programs. It uses Synaptic to install them graphically. It uses GNOME as the GUI, but you can also use KDE by installing Kubuntu. But Ubuntu works the best with GNOME. Ubuntu has been around since 2004.

Kainaw: I wouldn't recommend Fedora as it is used by Red Hat to test technologies before they're included in Red Hat, so I don't consider it stable enough for use by a newcomer. Likewise, I would recommend Debian over Ubuntu as Ubuntu is based on the unstable version of Debian. I've also used SUSE and it's very stable and easy to use.--Drknkn (talk) 06:15, 27 November 2009 (UTC)

Which technologies are tested in Fedora before inclusion in "Red Hat" (presumably Red Hat Enterprise Linux)? Nimur (talk) 15:37, 27 November 2009 (UTC)

There is a rather widespread, but completely false, belief that Fedora is the testing version of RedHat. The real deal is that RedHat is supported. So, every package officially included in RedHat is supported. You have to pay for that support. Fedora is not supported. You take it or leave it. So, it is not necessary that every package be supported (since none of them are). The end-result is that packages are quickly added and updated in Fedora. They are slowly added and updated (with the exception of security fixes) in RedHat. Because RedHat lags behind Fedora, most people believe that the packages are tested in Fedora for the purpose of including them in RedHat. So, you get comments such as "technologies are tested in Fedora before inclusion in RedHat." -- kainaw™ 21:45, 27 November 2009 (UTC)

Here are some citations to back up my claim: [1], [2]. Do you have any proof to back up your claim?--Drknkn (talk) 21:21, 28 November 2009 (UTC)

Ummm... yes. See Fedora Project. While RedHat has a lot of control over the Fedora Project, the Fedora Project steers development of Fedora. They do not simply test new packages to see if they will work on RedHat. Most of the packages available for Fedora since Fedora Core 1 in 2003 aren't even available in an official install of RedHat. So, if the claim is true that Fedora is a test-bed for Redhat, why are all those Fedora-only packages missing from RedHat? -- kainaw™ 21:43, 28 November 2009 (UTC)

Since we are using original synthesis in this debate, I should mention an anectdote of my own. I have used both Fedora and Red Hat Linux. After installing new releases of Fedora, I would then have to download several hundred megabytes of updates. They were clearly beta-quality releases. Admittedly, I've had more issues with Ubuntu than Fedora. Red Hat, on the other hand, has tended to be very stable, with fewer updates. Which one is the test bed? You decide.--Drknkn (talk) 22:30, 28 November 2009 (UTC)

If there's still any uncertainty, here is the Red Hat Enterprise Linux support schedule. "Test" packages are tested prior to the "General Availability" date for any release, in the developer alpha or beta period. You need to be pretty tight with Red Hat to get in on these things (usually, this means you need to be a research group or corporate customer already subscribed to some of their other products). The notion that Red Hat "tests" anything on any platform other than the main line RHEL is just silly - it would defeat the purpose of testing and compatibility certification. The "test bed" is exactly that - a pre-release RHEL. Nimur (talk) 22:35, 28 November 2009 (UTC)

Sorry, I failed to comment directly on your references. Had two screaming kids in my lap. The first one is correct only because it is rather old. When it was written, RedHat (and most of the world) saw Fedora as nothing more than a test-bed for RedHat. That view has changed. The second reference is also correct. RedHat likes to think that they own Fedora and they can use it to test things for RedHat. However, the Fedora Project doesn't sit around wondering how to make RedHat better. They are concerned with Fedora. If RedHat (or any other distro) benefits from their work, so be it.

I don't feel it is fair to call any distro (Fedora or Ubuntu) a "test-bed" just because it requires a lot of updates. There are two common mindsets for distros. Some are cutting-edge. When a program is updated, the package for that program is updated. Others hold back. Only when a program has a completely stable release will the package be updated. Since most distros use open-source software, the updates are continuous. It isn't like GIMP or Firefox are going to hit a "final version". They keep getting developed. If you demand to have the most recent version of all the programs, use a cutting-edge distro. If you require absolute stability, use one that doesn't update packages very often - and when they do, it is to a rather old version. As for test-bed, both Fedora and Ubuntu have a test-bed, which is the "testing" package repository. If you want real cutting edge with all the headaches, switch to the testing repo. -- kainaw™ 22:41, 28 November 2009 (UTC)

OP: I wrote a very long response to a similar question back in May. For your convenience, I will repost it here. In brief summary, Kainaw's description pretty much hit it on the head - most people know their Linux distribution for their package manager. Most linux and unix users would have a hard time actually knowing which distribution they were actually on, if you took away the logos. Only a few esoteric details, and some nitpicky details of configuration, separate the major linux distributions today; in fact, even most Linux/Unix/POSIX-like operating systems are pretty similar in 2009. So with this being said, I re-post my May 2009 response:

I spent a good portion of my undergraduate time working between different unixes and linuxes and qnixes and things you've never heard of. Boy, is it confusing! First of all, you've made the important first step in comprehending that the front-end user interface (GNOME or KDE or fvwm or whatever) is not the operating system distribution. (In fact I've run all of the above environments on all of the above *nixes and sometimes as a result I can't tell which machine I'm currently on!). And, your csh and bash and tcsh and zsh will probably run on all of the above as well. So... what's it all about? What's the difference between the distributions? (Linux distribution might help out here, but seriously... what exactly is a "distro" anyway? Why is Debian different than Ubuntu, if they both use the same package manager, same shell, same GUI, same libraries, ...)

Well, first of all, the Linuxes are all running the linux kernel, while the Solarises and BSDs and Mac OSXes are not. (And QNX? Well, just suffice to say that although it presents you with a POSIX-like shell and a lot of the standard system-calls, it's... not very much of a linux at all!) But all of them are POSIX compliant, and support networking and multithreading and encryption and so forth. But if you are going to remap your memory system for a custom coprocessor and need to recompile your kernel memory-module to handle variable page sizes based on current coprocessor instruction, you're going to need to choose your kernel carefully (I've heard, from people who would know, that CENTOS and Solaris make this task "easier"). And if you were planning to do something more benign, like maybe mixed shared memory programming with OpenMP and a little pthread code in the same program, you might actually find that there's a difference in the dynamic scheduler capability for different incarnations of the kernel. Or maybe you've got some files mounted on an AFS drive and you want to ensure that the network traffic stays encrypted, all the way through the machine, past the network, up to the shell, through the user-space, and decoded at the point-of-use in some kind of protected memory. Then you better have a kernel with libPAM module support! Are you doing these things? If not, you may never really notice your distribution.

Backing up a notch or two, at the "intermediate" level, you are going to want to install or compile some program some day which is going to have some dependencies. A lot of libraries are pre-packaged and precompiled for the common distributions (in the form of a DPKG or an RPM or sometimes even straight-up .so files). Pick a distribution that's going to be used by people who work with things that you work with... that way, you'll have a community which has already prepared the sort of tools you are going to need. It's not often worth anybody's time to trace back seven levels of library-dependency when you just want to get a standard tool to run.

Compiler support may be an issue between vendors. Some of the more esoteric optimization flags and the less standard extensions (like some c99 complex-math support) turns out to be not very platform-portable - this usually means that it's getting linked in with some system library (like libm.so).

So, what's the moral here? Distributions make a big difference if you're doing non-standard things; but if you follow "best practice" and write code that doesn't link with weird libraries, and doesn't jump from high-level logic to operating-system calls in the same module, you'll be better off and spend less time tracking down portability problems. I would stick with Ubuntu if I could, but some of my tools are only available on other linux platforms (and aren't worth the hassle of porting).

Hopefully this will give you some perspective - use "whatever distribution is easier." If you actually get to a point in your professional or academic development when you can decisively state that "the Solaris cilk scheduler gave me a 20% speed improvement" or "the network stack on QNX was insufficient to handle packet buffers for gigabyte-sized files using https" or some other distribution-specific issue, you're probably going to care what distribution you are using. Until then, pick a good shell, pick a good user-interface, and use as much standard unmodified software as you possibly can.

Nimur (talk) 15:43, 27 November 2009 (UTC)

Thanx for all the info. Sorry, for all the confusion that I caused with this and etc. To me all of your answers were confusing. Went with Ubuntu because it seems more stable then Suse or etc along those lines. —Preceding unsigned comment added by Jessicaabruno (talk • contribs) 21:03, 27 November 2009 (UTC)

Sorry for the confusing answers. Sometimes I give too much information, with the goal of completely explaining everything; a lot of people just need a quick summary answer. I think Ubuntu will be a good choice for most of your needs. Good luck, Nimur (talk) 05:00, 28 November 2009 (UTC)

I actually said that Ubuntu is less stable than SUSE.--Drknkn (talk) 22:30, 28 November 2009 (UTC)

Home server and net security

Hi, if I set up a computer at home to run as a Web server with Apache and Mediawiki and such, and I only aim to access this at home (all the computers will connect to the same router), am I safe from random people from the Internet accessing this server as well? In particular, is it sufficient to run one of these "port scanners" on the Web (I think Symantec or someone has one, at least), and if I'm all "stealth", I'm good? I haven't opened any ports on the router. As an add-on question, is there any easy and safe way to then open the computer to access from outside with a reliable password (or other identification) system? Thanks! Question inspired by SteveBakers MediaWiki marketing above Jørgen (talk) 12:38, 27 November 2009 (UTC)

"Safe" is relative. But if your router is not forwarding the ports, there is little additional risk. The easiest way to allow access from the outside is to forward just the ssh port, and to have strong passwords on the machine - or, even better, only allow the use of sufficiently strong public key exchange for ssh. Again, this is not absolutely safe, but likely to be safe enough for most people. If you need absolute safety, write on wax tablets and burn them before reading. ;-). --Stephan Schulz (talk) 12:51, 27 November 2009 (UTC)

Of course you could provide an extra layer of security by setting apache to only allow people from your network in. E.g., place in your .conf file:

allow from 196.128.100
deny from all

Magog the Ogre (talk) 18:22, 27 November 2009 (UTC)

Thanks both! Opening only SSH sounds like a good idea to start with. Now, would you consider apache or MediaWiki user control combined with an open HTTP port inadequate? No-one's after me, I'm just seeking protection from random hackers. Jørgen (talk) 20:09, 27 November 2009 (UTC)

Only open port 80 if it is only a web server and you aren't doing HTTPS stuff. Then, hackers can only get in on port 80. There is still the issue of scripts running on the server. Keep up on what you are running. Many scripts have vulnerabilities that come up over time and need to be updated. While the user can't alter your router through one of these vulnerabilities, they can alter the server itself and turn it into a spam-bot rather easily. A good log file manager will help you keep up on what vulnerabilities the hackers are looking for. For me, they are about 90% Windows vulnerabilities, but at least 10% are aimed at popular PHP, Python, and Ruby scripts. -- kainaw™ 20:57, 27 November 2009 (UTC)

What have I done to Firefox?

Everything I type into FF now comes out like ூபகே. By everything I mean Wikipedia (ஐகககஜா்கோ), the address bar and Google (உததுதா) searches. I have to type this into Notepad then copy and paste it here. I can't see anything obvious that I did to change anything but I really need to change it back. Thanks. Enter CambridgeBayWeather, waits for audience applause, not a sausage 13:31, 27 November 2009 (UTC)

Looks like you have somehow changed your keyboard language. What's your operating system? --Mr.98 (talk) 13:41, 27 November 2009 (UTC)

For what it's worth, you're typing in Tamil. Algebraist 13:44, 27 November 2009 (UTC)

I'm using XP and it appears to be confined to FF as I'm now using IE. I thought it looked like Tamil but I wasn't sure. Wouldn't have been so bad if this was mine but it's the work computer. Enter CambridgeBayWeather, waits for audience applause, not a sausage 13:57, 27 November 2009 (UTC)

It's OK. I should have thought and rebooted the computer. Still have no idea what I did to cause that but rebooting made the problem go away. Thanks. Enter CambridgeBayWeather, waits for audience applause, not a sausage 14:08, 27 November 2009 (UTC)

Has a Tamil spealker at your work, used your PC and inadvertantly (or as a joke) changed the input language to Tamil? Astronaut (talk) 14:35, 27 November 2009 (UTC)

No, it had to have been me. I've been the only person here for several hours and there was no problem until then. In fact the only reason that any other languages are on this computer is so I don't have to put up with ? or ? in articles. I don't suppose that anyone else even knows there are other languages installed. I checked and the computer doesn't have an hotkey setup to change to any language. Enter CambridgeBayWeather, waits for audience applause, not a sausage 14:43, 27 November 2009 (UTC)

If you have multiple keyboards/languages defined, there is a hot-key sequence to switch between languages - on my PC Alt-LeftShift will switch between English and Japanese input. First check to see if you can make the language bar appear: right-click on the taskbar and see if "language bar" is one of the toolbars you can activate. if not, you will have to look for it the Control Panel. When activated, there should be an option for the language bar settings which include ket settings. Unfortunately, I cannot guide you more accurately because I know that is one of the things that was changed in the move from XP to Vista. Astronaut (talk) 16:52, 27 November 2009 (UTC)

Thanks. I activated the language bar right after the Tamil started appearing. It shows that we can input in Tamil, two other languages and US English but was set to English and none appeared to have a setting for a hotkey. Right now I'm at home so I can't check to see what the other two were. I thought that changing that would make the default input into all programs the same and not just one. I wouldn't really be that bothered but if I can do it that easy then so can someone else. What makes it worse is just how long it took me to think and reboot. I should have tried that first! Enter CambridgeBayWeather, waits for audience applause, not a sausage 00:11, 28 November 2009 (UTC)

Ah, now I'm on an XP machine. If the language bar is not visible, go to control panel and choose "Regional and Language options". On the "Languages" tab, click "Details" to make the "Text services and Input Languages" box appear. At the bottom of the "Settings" tab, the "Key Settings..." button lets you set up the hotkey sequences. If the language bar is visible, choose "Settings..." to bring up the "Text services and Input Languages" box. Astronaut (talk) 01:51, 28 November 2009 (UTC)

Thanks. I started it up yesterday thinking it would be the problem but the input was set to US English. The other options are Tamil and Sinhla. I'm at work so I just checked and there are no hot keys set, so it's hard to understand how it could have changed if the language bar wasn't running. I also found that the input setting is per program, the top one, and not system wide, thank goodness. Please don't spend a lot of time on this as it's not likely that anybody else will manage to do that. As the reboot solved it I'm just curious as to how it happened. Thanks. Enter CambridgeBayWeather, waits for audience applause, not a sausage 04:32, 28 November 2009 (UTC)

Difference between Intel i7 CPUs

What is the (practical) difference in performance between Core i7-860 and Core i7-930? --Andreas Rejbrand (talk) 13:37, 27 November 2009 (UTC)

The i9xx series uses QuickPath Interconnect, which will yield higher performance to main memory and to peripheral devices. You can compare other specs at our Core i7 comparison article. I don't believe our article's un-sourced price estimate for the i930 - it is a higher performance CPU and will probably be more expensive than the i870. I also recall reading some rumors online that there will be serious reorganization of the i7 brand to re-label the i9xx series as high-end server processors "because they are too cheap" for their performance. (I can't find this forum post now). But you can look at the Xeon Harpertown E5xxx series (almost the same silicon, ~$1000 and up) and compare for yourself. Anyway, here's some more detail on the i920 for comparison. Nimur (talk) 15:24, 27 November 2009 (UTC)

Memory on the computer

Where do I look to find out how much memory I have and used on the computer? —Preceding unsigned comment added by 68.0.157.163 (talk) 22:46, 27 November 2009 (UTC)

What operating system? --LarryMac | Talk 22:50, 27 November 2009 (UTC)

If it's windows, task manager and process explorer will tell you. In task manager it's under the Performance tab, and in Process Explorer go View -> System Information —Preceding unsigned comment added by .isika (talk • contribs) 22:57, 27 November 2009 (UTC)

To find out how much memory is installed, simply press Win+Pause. --Andreas Rejbrand (talk) 00:12, 28 November 2009 (UTC)

Pause? Not OP but I have no pause on my keyboard. —Preceding unsigned comment added by 82.44.55.75 (talk) 00:35, 28 November 2009 (UTC)

Almost all keyboards have a Pause key; probably yours too. --Andreas Rejbrand (talk) 01:11, 28 November 2009 (UTC)

It is also labelled "Break". Astronaut (talk) 01:36, 28 November 2009 (UTC)

And it's usually on the far right of the very top row of keys. StuRat (talk) 07:36, 28 November 2009 (UTC)

...and if it's Linux (or Unix) you can type 'free' at the command prompt to get an overally summary or 'top' to get a task-by-task breakdown that updates as programs come and go and change their memory usage). SteveBaker (talk) 01:39, 29 November 2009 (UTC)

That was a good answer, Steve, overally. :-) StuRat (talk) 03:23, 30 November 2009 (UTC)

Closed Source and Open Source

Recently uninstalled Ubuntu again because having such a hard time with it. Now, I'm wondering is there any difference/s between closed source and open source. Since I already used Firebox and Wikipedia.

Think thats it for now.

Thank you, in advance, again. —Preceding unsigned comment added by Jessicaabruno (talk • contribs) 23:07, 27 November 2009 (UTC)

Well, the obvious difference is that open source releases its source code. Do you mean in performance? Thanks, gENIUS101 00:03, 28 November 2009 (UTC)

Closed source = recipe is secret, Open source = recipe is shared and published for all to see. —Preceding unsigned comment added by 82.44.55.75 (talk) 00:09, 28 November 2009 (UTC)

Thanx Basically, in performance and etc as well.--Jessica A Bruno 01:14, 28 November 2009 (UTC) —Preceding unsigned comment added by Jessicaabruno (talk • contribs)

Both have strengths and weaknesses. A strength of open source is a quality of the widely used applications (Linux, Emacs, gcc, Firefox...). A weakness is integration and polish, although Ubuntu has done a pretty good job in that respect. In general, most internet infrastructure is open source, and much more robust than the corresponding closed source programs. Personally, I use closed source for stuff that does not matter (iTunes, DVD player, Safari), and open source for stuff that does matter to me (emacs, gcc, LaTeX, make, Python), first because I don't want my work to be hostage to any particular software maker, and secondly because those tools simply work better for me. --Stephan Schulz (talk) 02:05, 28 November 2009 (UTC)

Thanx for the latest response. Interesting because how both of them work anyway.--Jessica A Bruno 17:56, 28 November 2009 (UTC) —Preceding unsigned comment added by Jessicaabruno (talk • contribs)

Good lord, the Reference Desk has lost its purpose in life: to provide references. See the articles open source and closed source. Tempshill (talk) 18:47, 28 November 2009 (UTC)

It is also worth mentioning that open or closed source is related to the licensing issue. The corrollary of open- and closed- sourecodes are free software and proprietary software (roughly speaking). Since Jessica A. Bruno (the OP) is probably not a programmer, this is most likely a more useful categorization than "open-source" or "closed-source". Usually, only programmers need access to the source-code; end-users want to know whether their programs are free and free. Nimur (talk) 22:50, 28 November 2009 (UTC)

Sorry, tempshill for all the trouble that I caused here. Understand were you are coming from on this.

Thanx, Nimur for your answer to my question here.--Jessica A Bruno 22:09, 30 November 2009 (UTC)

Cropping an image

I've officially gone stark raving mad over this now. My mom got me a digital photo frame for Eid al-Adha, and I've been messing with it for over an hour. I can get over the fact that it only supports JPEG. I can even get over its rather limited settings and options. But what I can not get over is its aspect ratio.

All 12 of its sample pictures are sized 480x234. That's 480 wide, and 234 tall. And yes, at ~2.051:1, that is a nonsense aspect ratio. I doubt that I have any images in the 1000+ on my computer that quite match that specification. The obvious solution is to crop some copies of images, but there rises a difficulty: I don't know of any programs that can do so appropriately. Sure, Windows Live Photo Gallery can crop images, but only at fixed aspect ratios / no aspect ratio at all. I've perused through search engine results to no avail. Is there any program I can use to crop a photo according to a certain aspect ratio?--The Ninth Bright Shiner 23:41, 27 November 2009 (UTC)

Yes, GIMP will. To do it, open the image, click on the rectangular select tool, in the toolbox click the "fixed" radiobox and select "aspect ratio" (as opposed to width, size, height) in the adjacent dropdown box. Put the ratio you calculated above into the field below that (with a colon delimiter) and you're done. Then using the select tool in the picture's window makes for a select that's always of that ratio; you can drag it around to move it, and drag its sides and corners, and it'll always be of the desired ratio. -- Finlay McWalter • Talk 00:06, 28 November 2009 (UTC)

Oops, I should add: to actually crop the image to the box you've selected, it's simply image->crop-to-selection -- Finlay McWalter • Talk 00:08, 28 November 2009 (UTC)

And it's sensible to then scale the image in GIMP to the size your frame wants (both to save storage space on the flash card and because GIMP's image resizer is very likely to be superior to that in the frame's firmware); you simply do image->scale-image in GIMP. -- Finlay McWalter • Talk 00:15, 28 November 2009 (UTC)

Oh, and yet another thing (you can tell I've one exactly the task you're asking about): when you scale the image, click the little chain that spans the width and height boxes, and make sure the chain is broken. That chain constrains the scale to the current proportion, but (in part because of that weird ratio you're having to use) you might end up with images that would naturally scale to 480x233 rather than 480x234 (due to rounding issues). This way, again, you're using GIMP's excellent scaling algorithm, and not risking the kludgetastic one in the frame seeing a 480x233 pixel image and doing a bad job of scaling it, for want of that one row. -- Finlay McWalter • Talk 00:27, 28 November 2009 (UTC)

Just curious: What happens if you give the frame an image in a more normal size and aspect ratio, say: 1280 x 1024? Astronaut (talk) 01:40, 28 November 2009 (UTC)

Nearly every digital frame on the market will show an image in any size/ratio, but there will be a significant delay while it resizes the photo. For that reason, it is preferred to manually resize and crop photos before putting them on digital frames. -- kainaw™ 03:22, 28 November 2009 (UTC)

Actually, in my frame's case, it tended to zoom in and cut off portions of the picture. Granted, that's what I'm doing manually, but I'm doing it properly. And by the way, GIMP works wonderfully. Thank you!--The Ninth Bright Shiner 03:35, 28 November 2009 (UTC)