Computing desk
< November 13	<< Oct \| November \| Dec >>	November 15 >

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 14[edit]

Subscripts[edit]

I have an old Windows XP PC. I'd like to see unicode U+2093 "_x", ~~U+208A "₊", U+208B "_-"~~. Any ideas? Regards --Chris.urs-o (talk) 11:22, 14 November 2013 (UTC)[reply]

Which font has U+2093 "_x"? --Chris.urs-o (talk) 04:43, 16 November 2013 (UTC)[reply]

Looks like when Internet was new[edit]

This morning a web page was slow to come up and it started out looking like web pages did when the Internet was new, with blue links. This is similar to what happened here, except thepage eventually changed to its the normal appearance. The explanation may be different from the one I was given before.— Vchimpanzee · talk · contributions · 14:03, 14 November 2013 (UTC)[reply]

Or it might be the same as (or very similar to) my comment of 20:33, 8 November 2013 (UTC). If the whole style sheet hasn't yet arrived, the browser may attempt to display what it can; when the style sheet finally arrives, the browser may redraw the page with the correct styling. --Redrose64 (talk) 15:33, 14 November 2013 (UTC)[reply]

Yes, I often get this effect, including on Wikipedia pages, because my internet connection is sometimes very slow. Dbfirs 15:57, 14 November 2013 (UTC)[reply]

Online maps and aerial photographs[edit]

I'm trying to make a large, detailed map of an area by piecing together screenshots of online maps, but I can't find any that work. Google Earth is nice, but they've rendered the earth's curvature accurately enough that the edges don't line up properly, whilst both Google and Bing maps work properly in 2D, but I can't get the right level of detail. I'd like it to show county boundaries, but that's not an option. I'd like not to have brightly coloured roads everywhere, can't do that unless I remove all the place name labels as well...

Anyone know where I can find a good, high resolution satellite map style image, with the county boundaries and place names but no other markings?

213.104.128.16 (talk) 18:02, 14 November 2013 (UTC)[reply]

Do u know these? OpenStreetMap Wikipedia:Graphics Lab/Map workshop --Chris.urs-o (talk) 18:46, 14 November 2013 (UTC)[reply]

That Open Street Map looks interesting, but I have to admit I'm not overly fond of it, they've piled so many different labels and areas on that it's difficult to read with all the colours jumbled together, is there nothing where I can choose which layers of data to include or not? 213.104.128.16 (talk) 00:19, 16 November 2013 (UTC)[reply]

Do you want to ask Wikipedia:Graphics Lab/Map workshop or the german sister? I think they have the basic map without labels. Commons "Category:Maps of the world" is a nice collection too. --Chris.urs-o (talk) 04:48, 16 November 2013 (UTC)[reply]

As I said at the start, I can get a map without labels, or a map with all the labels, on google, what I want is a map with only some labels and not others. I don't suppose there's any way of exporting data from Google Earth without taking a screenshot? Something that can get a large area of map, a lot more than can be covered in just one screen? 213.104.128.16 (talk) 17:49, 16 November 2013 (UTC)[reply]

Rammstein code[edit]

I have lately seen this image on Facebook and immediately got the joke: it's C++ keywords written so that they sound like Rammstein's Du hast. But I immediately also saw that this code won't compile. Now I got to thinking, is such a thing possible? Here's what I am thinking of:

Every single line must sound like the corresponding line of Du hast. By "sounding like", I mean that at least the count of words, and the count of syllables within the words, must match. Individual vowel and consonant sounds aren't that important, to make this easier.
Every single line must be compilable C and/or C++ code. To make this easier, each line needs only be compilable code by itself, the whole text need not be compilable.
Punctuation marks are allowed where ever necessary. However, words must be C and/or C++ keywords, identifiers aren't allowed. (If they were, we could just as well borrow Rammstein's original German lyrics.) Numbers aren't allowed either.

I got as far as:

do

do { break;

do { break; if

But then I noticed that the second statement in the block would never be executed, and was unsure if C and/or C++ allow this.

So is this in any way possible? JIP | Talk 18:27, 14 November 2013 (UTC)[reply]

Yes, code structure that would never execute is allowed. That includes declaring functions without invoking them, conditional instuctions like

if( 0) { neverDone(); }

or

while( false) { neverDone(); }

skipping some code with return, break or goto. Modern compilers typically warn you when they find such inaccessible code, but that's not an error (unless you set a restrictive mode 'treat all warning as errors'). --CiaPan (talk) 18:47, 14 November 2013 (UTC)[reply]

Thank you. In fact, that sample allowed me to go as far as:

while

while (true)

while (true) { break;

but then, I would need a two-syllable keyword, which can start a statement or a variable declaration, to sound like German gefragt ("have asked"). public or private would do, but neither is allowed inside the body of a function. JIP | Talk 18:52, 14 November 2013 (UTC)[reply]

How about while (true) { break; goto? I know the keyword goto must be followed by a label, and I said identifiers aren't allowed, so labels aren't allowed either, but then I did also say that each line only needs to be compilable by itself. Can the label be on a separate line? But after this line we run into trouble. The next line would have to sound like du hast mich gefragt und ich hab' nichts gesagt (ignore the apostrophe, it's only to mark colloquial German hab' instead of standard German habe, to make the word one syllable only), which would mean that goto would have to be followed by another keyword, which is not allowed. JIP | Talk 19:00, 14 November 2013 (UTC)[reply]

If I don't have any malware installed, why are open ports dangerous?[edit]

Why would a site with open port be vulnerable? You don't have to execute any package arriving at your ports. Why can't a web-site with open ports be perfectly protected from internet hackers? OsmanRF34 (talk) 19:23, 14 November 2013 (UTC)[reply]

Having a port open means that the computer responds to messages that arrive on that port. (A port, in case you don't know, is really just a number that is encoded into a specific part of an internet packet -- it isn't a physical thing.) There is no difference between closing a port and not responding to messages on it. Looie496 (talk) 22:15, 14 November 2013 (UTC)[reply]

Sure, but what makes it intrinsically unsafe to respond to messages? You don't have to execute as code anything arriving at your end. Indeed, the service listening to a port doesn't have to be able to do anything malicious (like deleting or updating your files). Why go to the extreme of not answering at all? OsmanRF34 (talk) 23:34, 14 November 2013 (UTC)[reply]

Responding to a message tells the hacker that there is a program responding to that port. If the hacker knows of any security holes in that program that can be exploited to gain access to the machine or escalate privileges, you have a potential security problem. Sometimes just responding can cause a problem--if the response takes sizable computational or network resources, it is a potential target for a denial-of-service attack. So you are correct, responding is not intrinsically unsafe, but there are plenty of practical problems that could arise. Minimizing the number of open ports/responding services is an example of reducing the attack surface of a system and is generally a good idea. --Mark viking (talk) 23:56, 14 November 2013 (UTC)[reply]

Accepting the message rather than just ignoring it completely with a closed port means that your system is doing some sort of work with the packet that comes in. That means you have one more piece of code that is doing something with network-originated traffic. Although just answering and not really doing anything should be safe, there is always the chance that there is some exploit in the network stack, your simple "do-nothing" code that is still at least doing a little bit, or somewhere else in the system that has to deal with that traffic. If you don't actually have a reason to let the traffic in in the first place, then it is safer to ignore it completely - otherwise you're increasing your system's attack surface (even if only slightly) in order to do something that has absolutely no purpose or benefit to you. Katie R (talk) 14:35, 15 November 2013 (UTC)[reply]

Spending time responding to obviously junk packets is undesirable because it makes you more vulnerable to denial of service attacks. So you want to discard them as early as possible in the pipeline - preferably in your network router - before they even reach the server. I'd also suggest (without evidence) that leaving unused ports open suggests a poorly defended computer - and that may encourage the evil-doers to spend more time pushing your computer to look for vulnerabilities than if you gave the impression of running a tight ship by shutting down everything you don't need. On my web servers, I even go so far as to go in and delete every piece of Linux software and every library that's not needed to serve the purpose of serving web pages, lock down every port and so forth. SteveBaker (talk) 16:32, 15 November 2013 (UTC)[reply]

That's the same thing I do for the Windows Embedded Standard 7 systems that I work with - it is componentized, so I can only install the parts of Windows that are required for the software we install on the systems. (Number one customer complaint after I started working and migrated us to WES7: "what happened to solitaire?") The systems are for product-testing in industrial situations, and many will end up connected to networks but denied internet access, or in a facility that requires a long and expensive recertification procedure whenever something about the system changes, so we can't expect Windows Updates to be run on the systems (even though they can be done if a customer wants to). After the first year of shipping, I was able to go through the list of security patches released since we started shipping and demonstrate that each vulnerabilty on it did not affect our system because they affected components that were disabled, or would be stopped by our aggressive firewall settings. If it was running a normal Windows 7 install, there would have been all sorts of potential vulnerabilities without those updates and with the default firewall configuration. That's the value of reducing the attack surface by shutting off anything you don't actually need - a system that seems secure when you set it up can end up insecure due to flaws discovered in any of it's exposed components, and reducing the number of exposed components reduces the risk of a flaw being found. Katie R (talk) 20:40, 15 November 2013 (UTC)[reply]

Allrovi down?[edit]

It seems that allrovi is down for a while, however, I fail to find any announcement about that. The [| Downforeveryone] site can confirm that it's not a problem on my end. I wonder if there is any way of getting feedback about what happened, or if all wiki links should be forwarded to allmovie. OsmanRF34 (talk) 23:29, 14 November 2013 (UTC)[reply]