Wikipedia:Reference desk/Archives/Computing/2016 July 24

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.

July 24

Open source and Microsoft/Apple

Have Microsoft or Apple released open source software/tools/standards? --Hofhof (talk) 13:33, 24 July 2016 (UTC)

"Open source software is at the heart of Apple platforms and developer tools, and Apple continues to contribute and release significant quantities of open source code." You can find code at: http://opensource.apple.com and you can read more about Unix in Apple's core system software at the System Technology Overview website.

Nimur (talk) 14:09, 24 July 2016 (UTC)

And normally, can this Apple's software be run on non-Apple computers?--Hofhof (talk) 17:12, 24 July 2016 (UTC)

Some Apple software is provided for non-Apple platforms, a well-known example being iTunes. It's up to Apple. Darwin is the free/open-source base for Apple's operating systems; since it's free software, you can do anything you want with it. However, most of the bells and whistles that distinguish Apple's operating systems like OS X are proprietary components layered on top of Darwin. Apple's licensing terms for these only allow you to run them on Apple hardware. --71.110.8.102 (talk) 19:20, 24 July 2016 (UTC)

The last time I checked, iTunes is not open-source software. Apple certainly does not try to portray iTunes as open-source software in any way, so it seems as though the topic is drifting.

To clarify this confusion: the OP asked about open source software. This terminology is frequently confused and mis-used; and maybe our OP wasn't even sure what they were asking for. The follow-up questions seem to suggest that the OP isn't actually looking for source code, and/or does not know what one does when they have source code - in other words, it sounds like they're mis-interpreting the meaning of "open source software."

For clarity: our original questioner might be seeking:

• software that is available at zero cost ("free software" in the sense of zero-cost); or,
• software that can be studied, modified, and used without restriction ("free software" in the sense of liberal licensing); or,
• source code for software (open source software, which may be either freely or restrictively licensed)

As it turns out, Apple provides instances of all of these sorts of software: some Apple software is totally non-free (restrictively licensed); some is free (zero cost) but is not open-source software, and that may or may not be restrictively licensed - an example might be some of the built-in software on your Mac; some Apple software is completely free (for example, licensed under the GPL, a BSD licsense, or similar), like the version of bash that Apple distributes; some is free and open-source under a slightly more restrictive license (including the Apple Public Source License). Apple's software falls everywhere along the entire spectrum.

It is certainly possible to build and run free, open-source software released by Apple, even on other platforms. In some cases, this requires great technical proficiency; in other cases, such as Swift for Linux, Apple provides pre-built executable programs that will run on nearly any hardware that can run a recent Ubuntu variant of Linux. Once again, Apple's software falls everywhere along the entire spectrum.

Nimur (talk) 04:45, 25 July 2016 (UTC)

Just one correction: "Source code available software" is not necessarily open source software by either popular understanding or by The Open Source Definition. To be open source, software has to be under a free (not in the sense of "extremely cheap", but in the 35 words sense ;-) license. Also see FLOSS. --Stephan Schulz (talk) 17:30, 29 July 2016 (UTC)

opensource.microsoft.com lists various Microsoft open-source projects. Microsoft Research has its own list, which includes the Glasgow Haskell Compiler. -- BenRG (talk) 19:12, 24 July 2016 (UTC)

Thunderbird is sending HTML code in the body of the message

Thunderbird has started sending HTML code in the body of the message. I posted a message on TB support a week ago, but haven't gotten a reply. Someone else has the same problem.

I went to Tools/Options/Send options and "send messages as plain test if possible" is checked and I changed "send the message as both plain text and HTML" to "convert the message to plain text", but that didn't fix the problem.

It doesn't do it all of the time, I don't think.

Does anyone have ideas on how to fix this? Bubba73 ^{You talkin' to me?} 16:08, 24 July 2016 (UTC)

It is doing it all of the time, unless I attach a text file. And I've tried other options for the plain/html setting. Bubba73 ^{You talkin' to me?} 16:16, 24 July 2016 (UTC)

And I have the same version of Thunderbird on an old computer, but it does not have the problem. Bubba73 ^{You talkin' to me?} 16:29, 24 July 2016 (UTC)

I would uninstall (even purge) it and install it again. --Hofhof (talk) 17:17, 24 July 2016 (UTC)

The settings do not change when reinstalling it. Settings are stored in the user profile. Send options can be modified in the settings. --Hans Haase (有问题吗) 18:16, 24 July 2016 (UTC)

That's why I said 'even purge.' Hofhof (talk) 18:48, 24 July 2016 (UTC)

What do you mean by purge - delete settings too? Maybe I should do a MozBackup from the old computer where it is working? Bubba73 ^{You talkin' to me?} 23:12, 24 July 2016 (UTC)

Neither Firebird or IE would download any file - it said "failed". I found an unwanted program called ByteFence. After I uninstalled that, downloads started working again and the problem with Thunderbird went away.

I think that Freemake Video Downloader did it. I've been using it for months - maybe over a year. The last time I downloaded an update, it installed a browser hijacker. When I searched for ByteFence, it is mentioned in the article. Bubba73 ^{You talkin' to me?} 23:22, 24 July 2016 (UTC)

What to do in Linux when nothing appears to work?

What to do when even the keyboard is unresponsive, and not even ctrl+alt+F3 would work? I've got this suggestion: Holding ALT, tap [sys rq], r, e, i, s, u, b. But this does not appear to do anything for me. Could it be that different distros need different combinations of keys? --Hofhof (talk) 17:50, 24 July 2016 (UTC)

You're going to need to provide more details. What is happening leading up to what you're describing? Are you booting the system and finding it unresponsive, or is this happening after using it for a while, or what? The SysRq thing is magic SysRq. Most distro kernels I think have it enabled, but it's not guaranteed that every distro does. Also if the system totally locks up—for instance, there's a kernel panic—magic SysRq won't do anything. It's not actually magic! It's more for, say, when X freezes. --71.110.8.102 (talk) 19:15, 24 July 2016 (UTC)

Then again, there is always this technique...  :(   -Guy Macon (talk) 02:38, 25 July 2016 (UTC)

• @Hofhof: Sysrq is deactivated by default, or at least restricted, on some distros. I learnt it the hard way on Ubuntu 14; see the first answer of this for how to fix it on Ubuntu (of course, once the computer froze, it is too late). What is your distribution?

Notice that there is an excellent security reason for such a restriction - but if it is your personal machine which none else has physical access to, it is not relevant. And as the IP said, sometimes, there is just nothing that will work. Tigraan^{Click here to contact me} 09:37, 25 July 2016 (UTC)

I am running Debian. I am the only user of this computer and don't fear people who could have access to my computer. Hofhof (talk) 12:24, 25 July 2016 (UTC)

If only the UI has failed, a common method for debugging is to enable the ssh server, and monitor logs from a remote machine that has used ssh to connect to the local machine, or to restart your graphical desktop environment using a shell command. If you're unfamiliar with using ssh, start with our article on SSH, and next, here's a tutorial specifically for Debian. Here is a list of system monitoring and logging tools you can use to perform basic triage. In many cases, you can simply kill and restart the graphic environment. If you're using GNOME3, here are instructions for managing gdm on debian, including the "correct" way to stop and restart its UI services.

When things go really wrong, like a kernel panic, it means the bug is much more severe, and the system has crashed beyond just the graphical interface. In those cases, you need a lot more expertise: you'll have to set up the machine for kernel debugging, and you'll need a second machine; and sometimes you will require specific computer hardware that makes such debugging possible. If you aren't familiar with this type of workflow, your best recourse as a user is to check if the release notes describe any known bugs and solutions.

Nimur (talk) 14:12, 25 July 2016 (UTC)

• "Debian" (whatever version that is) is a lot like Ubuntu (purists will say it is rather the other way around); going by this forum thread, in particular post #3, the spirit of the Ubuntu thread I linked likely applies. Check if your "magic key" variable is set to accept the "REISUB" sequence.

Additional note: you have to hold down both alt and sysrq while typing in the sequence. On some keyboards configurations it is even Fn+Alt+Sysrq held down during the sequence (so 4 keys at a given time). Tigraan^{Click here to contact me} 15:54, 25 July 2016 (UTC)

If the "keyboard is unresponsive" wouldn't the first thing to try is another keyboard? CambridgeBayWeather, Uqaqtuq (talk), Sunasuttuq 01:00, 26 July 2016 (UTC)

Forgive the user for a little bit of imprecise language in describing the symptom: it's a near-certainty that the keyboard hardware is not a causal factor in this case. The user meant to say something to the effect that the computer system fails to respond to keypress events. This is most probably due to a software bug in the desktop environment management software, or to a kernel oops, which (for most end-users) would be caused by an undiagnosed software bug in some device driver. There is a near-infinite list of other possible causes, and strictly speaking, we can't rule them out until we disprove every possibility, but in a typical workflow diagnosing such a problem, the first action is to start the investigation by collecting detailed information about the symptoms, and chasing down the most probable causes. Nimur (talk) 17:48, 26 July 2016 (UTC)

Has the keyboard ever worked with this OS installed or is this a new OS install? Is this by any chance a Debian install on the Raspberry Pi? Akld guy (talk) 20:34, 26 July 2016 (UTC)

On my Linux machine, the keyboard is plugged into a USB port but the USB ports are not all identical. The motherboard has a chip called a VIA VL800 that allows it to support more USB ports, but somehow this chip apparently gets locked up. If I have the keyboard on one of the ports connected to it, the system either stops hearing keypresses or thinks it is continuously hearing the same keypress. Rebooting the machine fixes things, but if I just want the keyboard to work, moving it onto one of the ports not connected through the VL800 chip does the job. (I normally don't use those ports very much now.) This may be a rare quirk of my machine but I pass it along for what it's worth. --69.159.9.219 (talk) 21:54, 28 July 2016 (UTC)

Why is everywhere so desperate for you to use their app instead of their site?

i.e. Pinterest but especially obscure sites that are commodities (i.e. a funny picture site) and couldn't be helped by an app (unlike i.e. Google Maps). They probably had to pay someone to write that stupid app for them. Why? If you actually got all the apps pushed in your face the phone would be clogged with apps that were only used once. Do the apps have a $.99 to $3.99 or so purchase or something that's far, far, far more valuable than advertising income? Do app ads pay more than site ads? Is it like this?: Say app and suddenly everyone wants it Sagittarian Milky Way (talk) 19:39, 24 July 2016 (UTC)

One reason is that apps can get access to personal information, such as a unique phone ID and your contacts, that in-browser Javascript is typically blocked from accessing. -- BenRG (talk) 01:44, 25 July 2016 (UTC)

Another reason is that they can be more in your field of view, doing alerts and popups, so even if you don't visit their web site or fire up their app they can get a message on your phone screen. Graeme Bartlett (talk) 06:34, 25 July 2016 (UTC)

While apps may be able to do a bit more, websites can now send notifications via the browser if it supports them and the user allows it. Most major browsers on mobile platforms do, with the notable exception of Safari on iOS [1]. (If you've ever agreed to let Facebook send notifications via the browser, you'll probably see it isn't that different from the app.) Nil Einne (talk) 11:43, 25 July 2016 (UTC)

I have some experience with this regarding restaurants. They give their apps away for free, often with an incentive, like a free app (appetizer) when you get their free app (application). I can think of two reasons they hope to eventually implement:

1) GPS related offers. So, when they see you are on "restaurant row", but just walked past their restaurant, they may send you an offer to get you to turn around and come in.

2) Inventory related offers. When they see they aren't going to sell all their shrimp before they close, they might search for anyone in the area and offer them a discount on shrimp to get rid of the excess, before they close and have to toss it out. Or they might send the offer out to everyone, but those who get notifications on their cell phones will presumably see them sooner than those who may not view their email until the next morning, and, once home, people are more reluctant to go out again.

Now these ideas might not be implemented yet, but meanwhile they want to increase their user base so it will be there when they are ready. StuRat (talk) 15:10, 25 July 2016 (UTC)

• Previous answers boil down to "collect user data for profiling future advertisement", which is likely true, but I see another reason here in the case of apps that include advertisements.

At least 99% of web advertisement today, either in terms of money or data, is delegated by websites to third parties (Google, Criteo, etc.). In-app advertisement is probably not much harder to block than in-browser, technically (I think? putting Google et. al.'s advertising domains on the IP blacklist) - but few know how to fiddle with hosts (file), while a significant proportion of internet users can hit the "download Adblock plus plugin" button on their browser. Attack ad: ABP runs an extorsion racket - use Ghostery instead. As a result, the proportion of users that will use your service but not generate ad revenues will be much lower with an app.

Slight off-topic: Note that this delegation to third parties makes it (relatively) easy to filter out the ad traffic, but before ad blockers the delegation was a win-win. Website admins do not have the technical headache to set up the advertisements as it is taken care by specialists; they do not need to vet advertisers, make them compete for spots, etc. On the other side, big ad companies can collect user data across multiple sites, use big data voodoo to profile the users, and serve them tailored advertisements. This increases by a large factor the click-per-view and buy-per-view statistics, which means an advertiser will be ready to pay a lot more per view. Tigraan^{Click here to contact me} 16:16, 25 July 2016 (UTC)