Wikipedia:Reference desk/Archives/Computing/2017 May 15

= May 15 =

thumb drive in glove compartment
A thumb drive kept indefinitely in a car's glove compartment, parked outside in an ordinary USA environment (hot in summer, cold in winter). Is there any reason to think its data might get damaged? 2602:304:CDA0:9220:24F3:F6D9:958E:D66 (talk) 07:38, 15 May 2017 (UTC)


 * People use USB sticks as USB dead drops, which is a much more exposed environment than inside a vehicle. So it will probably be okay, at least for a while. But I wouldn't make this my only means of storing any data that I wanted to keep safe. -- Finlay McWalter··–·Talk 07:54, 15 May 2017 (UTC)
 * What is an ordinary USA environment? Florida and Maine have very different climates. The vast majority of the northern hemisphere is hot in summer and cold in winter. But anyways, according to this USB flash drives are designed to operate at between 0 and 70 degrees celsius, but per this they aren't likely to have much problems at below 0 degrees celsius. Perhaps you should look at the manufacturer's specifications for the minimum storage temperature. Alcherin (talk) 09:50, 15 May 2017 (UTC)
 * Duh! Thanks for reminding me that the USA has a lot of microclimates.  I was trying to be broad and simple, and chose the locale of millions of the English-speaking readers who might answer.  Oh, is it objectionable to suggest that some of the responders are English-speaking?  Forgive me, and allow me to include enthusiasts of Indo-European, etc.  I cannot believe this board has devolved into such mindless pedantry. 2602:304:CDA0:9220:1D09:59C1:2EE8:921B (talk) 10:18, 16 May 2017 (UTC)
 * The majority of the southern hemisphere is hot in summer and cold in winter too.Hayttom (talk) 18:21, 15 May 2017 (UTC)
 * Tendentious and pedantic much? Jeez, man, not EVERYTHING is an example of chauvinism.  I was just trying to simplify the discussion.  If you'd prefer to imagine Venezuela as the locale, do so.  But keep your knee-jerk wounded universalism out of a board meant to provide simple information. 2602:304:CDA0:9220:1D09:59C1:2EE8:921B (talk) 10:14, 16 May 2017 (UTC)


 * Tin whiskers. Andy Dingley (talk) 09:54, 15 May 2017 (UTC)

This can be prevented with plastic spray, but not so on connector pins. There's also a difference between operating and storage temperature. A coating drops the maximum operating temperature due its thermal resistance. Before applying the plastic spray (a kind clear paint with electrical insulating properties), it is a good advice to wash the the solder flux from the PCB and extensively dry the PCB which can be a tissue on its cutting edges. Protect the USB jack on applying the spray. -- Hans Haase (有问题吗) 12:31, 15 May 2017 (UTC)


 * I've kept a thumb drive in my car for years (through three cars) and not had a problem. But it doesn't get very cold here.  Also, I've never tested it to see if it retains data because I use it for a temporary transfer of data.  Bubba73 You talkin' to me? 19:22, 15 May 2017 (UTC)
 * Data that exists in only one place is just waiting to be lost. If you have the data backed up elsewhere like you should, what happens to the drive shouldn't matter. In any case flash memory slowly "leaks" charge, so data on it has a finite lifetime. --47.138.161.183 (talk) 20:27, 15 May 2017 (UTC)
 * Re-read the question. Why are you answering something that was not asked?  "Data that exists in only one place is just waiting to be lost. If you have the data backed up elsewhere like you should, what happens to the drive shouldn't matter."  Oh, really?  Wow, that NEVER occurred to me!  I sure should have mentioned in the question that the data was stored elsewhere, TOO.  Now give me a lesson on how it should be stored in the cloud, as well, since obviously I don't know that. 2602:304:CDA0:9220:1D09:59C1:2EE8:921B (talk) 10:21, 16 May 2017 (UTC)
 * It might matter a lot if you expected the thumb-drive to act as some sort of emergency data store of files you might suddenly need to recover quickly.
 * I'm not sure what kind of data you might need in an emergency that couldn't be placed online, but perhaps the question-asker has a use-case.
 * (Edit:Of course, it needn't even be "emergency" data. Just infrequently used data. A traveling service technician might want to keep drivers or something handy on a flash drive, for those rare cases when internet is not available. It might be months between uses, and the rest of the time it would just sit in the tool box.)
 * ApLundell (talk) 21:58, 15 May 2017 (UTC)
 * WHY ARE PEOPLE EVEN DISCUSSING THIS? Maybe the question was about a character in a novel.  Maybe it was about someone with a car and an IQ of 57.  This board is meant to answer specific questions.  Here the question was very simple:  given a thumb-drive in a car's glove compartment in both hot and cold weather outside, will the drive lose data?  WOW, this is a bad reference board. 2602:304:CDA0:9220:1D09:59C1:2EE8:921B (talk) 10:25, 16 May 2017 (UTC)

Wikipedia talk:WikiProject JavaScript
I've been stuck on a viewport repositioning problem. If this is something you have familiarity with, I could sure use your help. The Transhumanist 17:26, 15 May 2017 (UTC)

External Solid-state harddisks (for backup use)
1.Being easier to built they ought to be cheaper than the moving-parts types, but it's otherwise. Why is that, please? 2.Are they as reliable as the moving-parts harddisks ? 124.253.249.145 (talk) 18:09, 15 May 2017 (UTC)


 * (1) the memory for an SSD is more expensive than the parts for a HD. (2) I think that they are at least as reliable.  Bubba73 You talkin' to me? 18:20, 15 May 2017 (UTC)
 * Echoing the previous reply: do you know with certainty that flash memory is easier to build than a hard drive? A semiconductor fab costs billions of USD to build and run. You don't need a fab to make a hard disk. Now, hard drives do contain chips (which of course have to be made in fabs) for controlling the drive and talking to whatever it's hooked up to, but those chips are relatively simple and cheap. The driving factor behind the price of flash memory is the demand to cram ever more bits onto the chips, which means you have to use smaller fabrication processes, which are more expensive and complex. You can get small amounts of flash memory for practically nothing, since they don't have to use these processes, but you probably wouldn't find too much use for a 512 KB flash drive! --47.138.161.183 (talk) 20:40, 15 May 2017 (UTC)


 * Some SSDs, even when blocks are worn out due extensive writing not not keep enough space for compressed data due internal write amplification. Such SSDs usually fail in raid applications when utilized 100%. The chance of damage by dropping the drive is lower. But an electrical failure such as over or negative voltage causes the problem of any possible recovery due internal compression and encryption as well as virtual addressing blocks due wear. -- Hans Haase (有问题吗) 10:08, 18 May 2017 (UTC)

Partitioning an external hard drive?
I want to buy a large external hard drive, perhaps 5T and partition it. Is it possible? Thanks, --AboutFace 22 (talk) 18:55, 15 May 2017 (UTC)


 * You can just google this yourself ! Aspro (talk) 19:03, 15 May 2017 (UTC)


 * A quick google search shows the Seagate STDT5000100 5TB external drive is sold by Amazon, Best Buy and Newegg for roughly the same price.
 * As far as partitioning goes, MBR partitions only support a maximum of 2TB volumes so you'll want to make a GPT partition if you want the whole 5TB to be a single volume. After that, just format it with whatever filesystem you want to use (probably NTFS for Windows, ext4 or maybe XFS for Linux, not sure what Mac folks are using these days.) and you'll be good to go. Seagate usually ships their disks with some kind of tools to do this for you but you can always just do it yourself using the built in OS tools instead. Random character sequence (talk) 19:53, 15 May 2017 (UTC)
 * The only potential issues will be with old software and hardware. See the above-linked GPT article. If you want to boot from the drive you may have problems with systems that aren't fairly new. But, if you just want to use it to store stuff, the only likely problem is if you're using Windows XP, which doesn't support GPT partitions at all. But XP is no longer supported anyway, so you really need to switch to a newer OS if that's the case. --47.138.161.183 (talk) 20:44, 15 May 2017 (UTC)
 * Note that some external drives expose a different sector size, perhaps 4k (which is probably the real sector size, but the drive tends to be just a SATA drive which is exposing a 512 byte sector size to the USB controller) which means there's actually no problem with a single partition larger than 2TiB. Nil Einne (talk) 12:39, 16 May 2017 (UTC)

Thanks, my OS is Windows 10, I have good hardware too. I did a backup yesterday and it took more than a half of my 1T external drive and said there is no space for the next one. This crazy talk about the ransom ware and the like, prompted me to try to safeguard my computer and the home network as much as I can. --AboutFace 22 (talk) 20:57, 15 May 2017 (UTC)


 * I partition external drives. I've found drive failures to increase with logical drive size. As an anecdote, I managed a computer with a 2T external drive. It had failures on the external drive every month. Sent the stupid thing back over and over. Had to back it up daily. I then partitioned it into 4 drives, 500GB each. In the OS, I merged the four 500GB drives as a single drive. I never had a drive failure again on that external drive. The other IT guys say it was just coincidence, but I have a slew of anecdotes to share. 209.149.113.5 (talk) 11:42, 16 May 2017 (UTC)
 * I think partitioning 2TB drive into 500MB partitions is a bit excessive, but perhaps that's just me. Nil Einne (talk) 12:39, 16 May 2017 (UTC)
 * Typo.. I got GB right the first time. Correcting... 209.149.113.5 (talk) 14:42, 16 May 2017 (UTC)
 * I've frequently replaced the partition scheme on external drives, changing them from MBR to GPT if needed when they are over 2TiB. However I normally just keep one partition, or maybe leave a little space for a small FAT16/32 partition. I've had a single failure after about 1.5 years. If you have good reasons to want to partition, go ahead, but make sure these are good reasons. I wouldn't rely on anecdotal reports of reduced failures of the drive, I'm sure you can find just as many people who say there are increase failures with more partitions. (The only failure advantage is that although file systems like NTFS etc are generally more resilient, it's still possible to for data to be lost or difficult to recover due to file system corruption and it's probably less likely you're going to suffer file system corruption of multiple partitions at the same time or close to it.) Nil Einne (talk) 12:39, 16 May 2017 (UTC)

Thank you guys. It really pays to post silly questions. A wealth of information. --AboutFace 22 (talk) 14:10, 16 May 2017 (UTC)


 * Some adapters block writing the partition table or several information. Whenever, remove the drive from its external case and connect it to a desktop computers mainboard. Setup the partitions and format them. Shutdown the desktop and place the drive into its external case again. -- Hans Haase (有问题吗) 10:00, 18 May 2017 (UTC)

Ransom ware question.
I simply want to understand some aspects of it in depth. I read this article. 'Accidental hero' halts ransom ware attack and warns: this is not over. I am not sure some people who write such articles understand what they are doing. This is an obscure internet site. In nutshell, a technician in England discovered this malware and saw that the domain was not registered. Domain was there but not registered with an official domain registrar, like godaddy.com. Is it possible to create a fictitious domain and send email around in such a way that the client computers will contact this domain and get infected. Isn't it what happened?

Then that tech took the domain name and registered it instantly!!! thus preventing further damage. A version I read on Friday said that the fact that he took control of the domain prevented the spread of the virus in the United States only. He also says that he wants to remain anonymous. I doubt that anyone can remain anonymous. You have to give your real name, use a credit card to pay for the domain name, everything about you gets published. Suppose he used his company's credit card, then perhaps it is possible this way.

Internet is a complicated business and I want to understand as much as I can. Thanks, --AboutFace 22 (talk) 21:13, 15 May 2017 (UTC)


 * The strange domain name was coded into the malware. It didn't exist until someone registered it.  We can only guess at what the author of the malware intended to do with the domain, but the most likely explanation is that once sufficient ransoms had been paid, the author intended to stop the spread of the malware to reduce the level of investigation and thus to reduce the probability of detection.  There might be other reasons, of course.    D b f i r s   21:35, 15 May 2017 (UTC)


 * You can register domains by proxy, and stay mostly anonymous. The registrar knows who you are, and could turn you in if the court ordered them to, but the published contact info just leads to the registrar. For the altruist engineer who registered the domain and wound up stopping the attack, that level of anonymity is probably enough. (Presumably the criminals would have registered the domain using stolen credit cards, or something like that.)
 * ApLundell (talk) 21:43, 15 May 2017 (UTC)


 * Plenty of domain vendors and hosting companies take bitcoin, which the operator of this malware is evidently very familiar with. While stolen credit card info is available on carder forums on the darknet, experienced criminals generally use them to buy some good (often digital goods) that can be readily exchanged for real money. Buying an ongoing service like a domain or hosting with a stolen credit card is daft, because once the fraud is detected, the service provider will quickly shut down the service. -- Finlay McWalter··–·Talk 22:06, 15 May 2017 (UTC)
 * Personally, I'd use my own CC. Then when the cops showed up, I'd grab my oldest son and start yelling at him until they got bored and wandered off. Then I'd buy him a video game with my ill gotten funds. MUAH HA HA HA HA!!! ᛗᛁᛟᛚᚾᛁᚱPants   Tell me all about it.  22:17, 15 May 2017 (UTC)


 * The domain is registered to "Botnet Sinkhole" and is through a registrar that accepts bitcoin, so if they did buy the domain with bitcoin and did it through a VPN that doesn't log user traffic, there may not be a way to find out MalwareTech's identity.
 * As far as the more technical details of how they stopped the spread of the attack, their blog post about the event probably has what you're looking for. Random character sequence (talk) 22:28, 15 May 2017 (UTC)

Well, in order for a domain name to be recognized on the network it has to be registered, otherwise DHCP servers could never find it. And I think for an operation like this a static IP address is necessary also. A static IP address must be registered and it is available publicly. Am I confused? --AboutFace 22 (talk) 22:36, 15 May 2017 (UTC)
 * It's actually the DNS servers that tell other systems how to find it, not the DHCP servers. The actual IP can be either static or dynamic, as long as the DNS server has the current IP on file. Dynamic DNS has more detailed info on the subject.
 * Based on their DNS info, it looks like their DNS A record is pointed to a cloud computing company. If I were to guess, I'd say they're using Namecheap to point to Amazon's DNS which points to a virtual server in Canada. None of that requires any real dedicated hardware, they probably just paid a few dollars to have the different companies point to each other and made a quick home page for the domain. I didn't bother to check to see if Amazon and OVH accept bitcoins as well, but I wouldn't be surprised if they did. Worse case scenario they'd have to pay a couple bucks to set up a prepaid debit card and use that to buy the Amazon and OVH subscriptions. Random character sequence (talk) 22:54, 15 May 2017 (UTC)
 * I think you're misunderstanding some things. What happened: the ransomware was programmed by the authors to ping a certain domain name (I'll use "example.com" as a stand-in) and stop working if it succeeded. Our "hero" dug through the code, found this, and went ahead and registered "example.com" themselves. I think you said DHCP when you meant DNS; the DNS is what your device uses to look up "example.com" when you tell it to go there.
 * Regarding your initial post where you said: Is it possible to create a fictitious domain and send email around in such a way that the client computers will contact this domain and get infected. Isn't it what happened? No, it isn't. The "kill switch" domain was, before our "hero" registered it, unregistered and had nothing to do with how the exploit spread. It spread by exploiting a Windows vulnerability, which when exploited tricks Windows into executing code that the attacker wants it to. --47.138.161.183 (talk) 23:04, 15 May 2017 (UTC)
 * Note: 's previous post was made after I opened the edit window, but Mediawiki apparently automatically merged my post in instead of triggering an edit conflict. I moved my post below. --47.138.161.183 (talk) 23:08, 15 May 2017 (UTC)

Thank you. By posting here I had an unstated desire to understand the process to such a degree that I could myself perhaps buy a hardware firewall and program it or perhaps build defenses in other way. I still don't understand anything. I want to only add one thing I observed though. It was a snapshot of a ransom note in English in one publication and also a similar note in Chinese. The latter had a page full of their characters. I doubt East European hackers could have prepared messages in so many languages to spread around the world. The message should be in Portuguese, Spanish, French, etc. An article in the New York Times this morning says it is all North Korean doing. It is much more realistic. Thanks. --AboutFace 22 (talk) 13:32, 16 May 2017 (UTC)
 * It might be someone from North Korean, but I think this is a serious smoke screen. I18n is no magic anymore, and a hacker in Seattle or Budapest or Berlin can run a text through Google Translate as well as any Korean. Korea nowadays uses Hangul, and the Korean language is an isolate, so not particularly similar to Chinese. As for the WannaCry attack itself: The domain "kill switch" has nothing to do with how it spreads. It's not quite clear how WannaCry initially attacks a local network, but once in, it then uses the EternalBlue bug in Microsoft's SMB protocol to spread to all computers in the network. The "kill switch" is an explicit test it performs before it spreads. It performs a Domain Name Service lookup for a freakish domain name. That can come either back with an NXDOMAIN error (if the domain is not registered), or with a list of IP addresses. If it comes back with NXDOMAIN (or no answer at all), the virus keeps  spreading. It does not need the domain or the computer behind it - it does not even care what's behind the domain name or the IP address, and it never even queries the site - it just checks if the name is registered. A firewall will not stop this - indeed, some firewalls that stop "suspicious" DNS requests kept the kill switch from working, and thus kept it spreading further. --Stephan Schulz (talk) 06:56, 18 May 2017 (UTC)
 * "does not even care what's behind the domain name or the IP address, and it never even queries the site"? Are you sure that's right? My understanding from the code (limited) and what I've read e.g.   (the last one in particular) is that it does need to establish some sort of TCP connection. It doesn't care what the server actually replies, it doesn't I think even need to be a HTTP connection. But if there's no response at all to queries, the test fails and WannaCry will still try to spread. I believe this means if it's pointing to an IP and the device behind the IP is filtering all requests to port 80 and so there's never any response simply a timeout, the test fails. If the IP is responding to requests on port 80, even if it's just a RST because the port is closed, then you are fine. Nil Einne (talk) 05:36, 19 May 2017 (UTC)
 * Actually reading and thinking, a bit more, I'm not sure if even that's correct. E.g. . I think the IP that the domain resolves to needs to have port 80 open with a server of some sort behind it, so a TCP connection can be established. It doesn't matter what this is, it doesn't really have to be a proper server (and definitely not a webserver), provided a TCP connection can be established. If the TCP port is closed for that IP, this is also a failure and the malware will continue to spread. I admit though, this is starting to get beyond my understanding and I couldn't find a ref with properly explains what the code requires, so I may be wrong. So definitely do not rely on this for anything useful. Nil Einne (talk) 06:25, 19 May 2017 (UTC)
 * You are right. Apparently, it uses some Windows API function call that tries to open a connection, and it continues if that fails. That's different from what most early popular press articles reported (which should teach me something...). --Stephan Schulz (talk) 14:48, 19 May 2017 (UTC)

Autofill form fields in Chrome
There's a web page that I use for work a couple times a week. The first couple fields that I have to fill in are my name, employee number, a pull down menu that I always select the same option for, and a radio button which I always choose the same option for. Is there some way of automating this? I'd use Chrome's autofill but as soon as it sees me typing in my own name, it wants to fill in my personal contact info into the rest of the fields farther down the page such as mailing address and email which, for this page, are always different since I'm not filling in my address and email. Thanks, † dismas †|(talk) 22:39, 15 May 2017 (UTC)
 * AutoHotkey can do exactly that. Take a look at the article and if you decide to download, comment back here. I'll happily share tips about writing a script that would do that for you. ᛗᛁᛟᛚᚾᛁᚱPants   Tell me all about it.  22:46, 15 May 2017 (UTC)
 * , how would I go about doing what I outlined above? Thanks!  † dismas †|(talk) 11:26, 18 May 2017 (UTC)
 * Okay, first, you need a way to get the screen coordinates of the form controls. So the first thing you do is download something that can show you the current mouse position (you don't actually need to do this, you can use trial and error, but this way is faster). There's one here that a friend of mine recommended (I programmed my own, it took about 5 minutes and I can share the source code tomorrow if you would like it.)
 * Next, you need to mark down what they are. You may need to scroll down on the page a bit to get them all on screen at once. If you do that, use the mousewheel to scroll down so you can simulate it using AHK.
 * Now, it's time for the script. Make sure you have AHK installed, then right click on your desktop and choose . Give it a name, then right click on it and choose   to open it in notepad. You will want to write a script that is something like this:


 * The italic green text ( Example ) is comments, you do not need to include them. They are there to explain what each line does. Once this is done, save the file, then double click on it. You will see a green square with an "H" in it appear in your system tray (bottom right of the screen); this lets you know the script is active. The just load the website and hit the numberpad enter key to trigger it. You can change this to any key on your keyboard, any mouse button (or scroll wheel) or any joystick button. If you need to scroll down before the fields, what you want to do is insert the following line underneath the  line, once for each click of the wheel.


 * I hope this helps. ᛗᛁᛟᛚᚾᛁᚱPants   Tell me all about it.  00:41, 19 May 2017 (UTC)

”Autostarted” on smart phone
Some unnecessary programs autostart every time I turn my smart phone off and on, for example games… How do I keep them off and fully without uninstalling them? 103.67.159.240 (talk) 22:50, 15 May 2017 (UTC)
 * Unless your phone is rooted, you don't have a good way to prevent services from being launched by an app. (Except uninstalling the app completely.)
 * There are apps that "auto-kill" services, but I strongly recommend that you don't use them. They do far more harm than good. (Most well-written services are pretty harmless when they're just running, but they use CPU and battery when they restart. So if you keep killing them, and they keep restarting....) ApLundell (talk) 14:33, 16 May 2017 (UTC)

”Force stop” on smart phone
Sometimes force stop button doesn’t work in the “App info”. Some apps provide a second force stop button inside “Battery usage” feature situated in the “App info” list which works for some and doesn’t for some. How do I stop a program manually with having to uninstall. 103.67.159.240 (talk) 22:50, 15 May 2017 (UTC)


 * Could be a buggy app, or it keeps restarting itself. You could use a task killer for the latter, but this can cause crashes, instability, battery drain, etc. Is this app a core part of the OS? Thanks ツ Jenova   20  (email) 13:03, 16 May 2017 (UTC)


 * It may be that there's an app and a service. (Some apps have one of each. Like an email program for example. The app is the part you actually run and interact with, but the service stays in the background and checks your emails every couple minutes.)
 * The "force stop" button should kill both the app and the service, but the service can restart without any active step on your part. Most services are well behaved and don't effect your device's performance, but some games will come with really abusive services that are part of their advertising package. ApLundell (talk) 14:43, 16 May 2017 (UTC)

External Storage
How do I directly download from ‘‘google play store’’ and install it in the external storage place? It always seem to install in the internal storage place even when I change it manually… 103.67.159.240 (talk) 22:50, 15 May 2017 (UTC)


 * By design Google Play only allows installation on the device. I'm pretty sure you can't change this. Thanks ツ Jenova   20  (email) 10:32, 16 May 2017 (UTC)


 * There's theoretically a setting for this, but virtually all phones hide it, and most phones keep it locked even if you use the debug command line. (If you're brave, the procedure is here. Don't recommend it.)
 * Another option is Android Marshmallow's Adoptable Storage. This is a very nice feature that allows you to combine the SD card and the built-in storage into a single storage area. Unfortunately, Samsung and LG intentionally disable this feature so they can upsell you to a more expensive phone with more built-in storage.
 * In other words, if you root your phone and install a stock version of Android, you have two useful options that are built into the operating system. If you don't root your phone, you can't use those options because most manufacturers have gone out of their way to disable them. ApLundell (talk) 14:22, 16 May 2017 (UTC)

App permission
What is it, that is found inside “Settings -- Apps – ‘click the app of your choice’ – Permissions ‘‘feature’’? – there is a switch that goes ‘left’ to ‘right’, if activated, does it mean it will automatically store downloads of the particular app of choice in the Micro SD? 103.67.159.240 (talk) 22:50, 15 May 2017 (UTC)


 * Does it look something like this? Thanks ツ Jenova   20  (email) 10:33, 16 May 2017 (UTC)


 * The "storage" permission does not move the app to the sd card. It simply gives the app permission to save files there. For example, a camera app would probably request that permission so it could save photographs. ApLundell (talk) 14:46, 16 May 2017 (UTC)