Wikipedia:Reference desk/Archives/Computing/2015 August 9

= August 9 =

Fuzzy logic, when you are using a digital computer
Moved to Reference_desk/Mathematics. --YX-1000A (talk) 21:19, 9 August 2015 (UTC)

A question in computer science with an emphasis in communication & Internet
First I will say I am looking an answer which is very much simple and general, with minimal and basic usage of technical terms, just for primal understanding and first steps:

I understand that any Internet address is built from DNS records which actually comprise it... My question is what is the "A" (Address) DNS record? I understood that is some kind of a mapping mechanism that connects the Domain name (I.e Google.com) to it's original I.P address...

Well, I ask mainly what is this a mechanism? where are the maps it creates via I.P <> Domain-name associations are situated? in my DNS register? in my Hosting service? In what programming language this mechanism is written in? Thank you very much for your answers. Ben-Yeudith (talk) 21:37, 9 August 2015 (UTC)


 * "I understand that any Internet address is built from DNS records which actually comprise it". That is incorrect. An Internet address is not built from the DNS record that comprise it, anymore than a road name is built from a Road Sign that defines it. The road name is created first then the road sign is physically constructed to show the whole world which road it is. The internet address exists first (called IP address) then the DNS server is instructed to convert a "string address (aka Domain Name)" into a numeric internet address (IP address). Aka blog.dnsimple.com maps to 72.32.231.8 according to the DNS server.
 * The DNS Server contains records. Each record is belongs to one of three types. These are A, NS and MX.
 * A type is Address. The record contains the numeric IP addr for a Domain name ("string address")
 * NS type is Name Server. The record contains the address of the server that can return results for names in a DNS zone.
 * MX type is a Mail Server. The record contains the address of a mail server, where email should be send to. 175.45.116.59 (talk) 03:24, 10 August 2015 (UTC)


 * you might want to consult this website for terminology. 175.45.116.59 (talk) 03:45, 10 August 2015 (UTC)

Removing metadata from images: why?
In discussing various ways of compressing files, the Lossy compression article notes that "Metadata, such as ID3 tags, Vorbis comments, or Exif information, can usually be modified or removed without modifying the underlying data." Not surprising. But why in the world would you want to do this as a method of compression? You save literally just a few bytes per file, not at all a significant total unless you're working with a vast number of little files, in which case it will take a massive amount of time to remove the metadata from all the files. There's no source for this statement in the article, or I'd check it. Nyttend (talk) 22:28, 9 August 2015 (UTC)


 * It's not a value judgement, it's just stating a fact. Metadata is an overhead that is not part of the actual image therefore it can be removed to reduce the file size. Whether it's worth doing or not is a different question. Vespine (talk) 23:30, 9 August 2015 (UTC)


 * On a modern computer with an SSD, it shouldn't take a massive amount of time to remove metadata even from 1 million files. In any case, most computers have a lot of idle time, so unless you desperately need the files, there's no reason why you couldn't have the metadata remove in the background. I'm only talking about computing time here, since there's rarely going to be a good reason to remove metadata from a lot of files completely manually and in any case that's a moot point as no one said in the article metadata needs to be removed manually. BTW, it isn't true metadata is only a few bytes. An ICC profile is often at least 4k and can I think be a lot larger than that (see  . Thumbnails could also be a few k and I believe some file formats may allow more than one. Note that for a 40k image, removing 4k or more isn't a small amount, if you're serving that image to millions of browsers daily. Consider the efforts to unnecessary remove whitespace from HTML   to speed up load times for example. Nil Einne (talk) 00:21, 10 August 2015 (UTC)


 * At one file per second - it would take 11 days to do a million files. Probably it could be done significantly faster than that - a few hours maybe.  But in most cases, the metadata for an image is only likely to be maybe a few hundred bytes - so the savings would probably be no more than a few hundred megabytes.  Definitely not worth the effort in a multi-gigabyte data volume. SteveBaker (talk) 03:56, 10 August 2015 (UTC)


 * Also, another reason to remove metadata is that you want to remain anonymous. Let's say you have a pic of cops beating an unarmed civilian, do you want them to be able to identify you when you post it online ?  The less metadata they have, the harder it will be for them to track you down.  If you supplied your name to your digital camera or when registering the PC you download the pic onto, it might have "helpfully" added your name as part of the metadata.  StuRat (talk) 19:36, 10 August 2015 (UTC)


 * What does that have to do with the question? He's asking about compression savings and you're talking about anonymity. Is your point that EXIF data should be stripped during file compression just in case you might have some pictures of cops in there behaving badly? I know you're in Detroit, but I assume that's not something that happens to you with such frequency that you need it as the default setting on your computer. 64.235.97.146 (talk) 14:51, 11 August 2015 (UTC)


 * I may very well want that as the default setting. And, while the tiny space savings alone may not be enough reason to offer the option to strip off metadata, that space savings plus anonymity might well be. StuRat (talk) 15:18, 11 August 2015 (UTC)