Computing desk
< June 5	<< May \| June \| Jul >>	June 7 >

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.

June 6

For IT books, how old is too old?

When do publishers stop selling a book, if at all? (given a case where there is no new edition) Some books at oreilly.com are pretty old. For example, [| XPath and XPointer ]is 12 years old. [| sed & awk, 2nd Edition ] is 17 years. Both can still be purchased through Amazon. OsmanRF34 (talk) 16:51, 6 June 2014 (UTC)[reply]

And the Mach Server Guide is 21 years old, but I still reference it. The C Programming Language is even older, and I still reference that. Speaking of technology documentation, the legally-approved operating handbook for my Citabria is a 1975 Bellanca Aircraft Company document. No newer manual is "legal" for the purposes of keeping the aircraft airworthy, and although my E6B computer is not called out in the installed-equipment listing, it too has documentation that is pretty old. When technology documentation is done well, it doesn't necessarily need to be updated. On the other hand, if you try to write Python 3 using Python 2 documentation, you'll find a lot of incompatibilities; and if you are a Swift programmer, you'll have a hard time finding public documentation older than five days (as of this post). Nimur (talk) 17:08, 6 June 2014 (UTC)[reply]

I see stuff from 2011 in Swift (parallel scripting language). I imagine if you look a bit more you'll find stuff from 2007. Nil Einne (talk) 06:31, 7 June 2014 (UTC)[reply]

For clarity, I ought to have linked directly to Swift (Apple programming language), which was announced to the world only on Monday. Nimur (talk) 15:31, 7 June 2014 (UTC)[reply]

The linkage wouldn't change what the text says so wouldn't have changed my response. (I saw the Apple programming language listed there, but it was irrelevant as the text simply mentioned a Swift programmer, not a programmer of the Apple Swift language. If someone, be it Apple or someone else, is going to call their stuff the same as something similar that already exists, they can't resonably expect their usage to become the only correct usage of the word such that it be automatically accepted their usage is meant with only easter egg links providing any clarity that it's not the older, existing usage of the term.) Nil Einne (talk) 13:39, 8 June 2014 (UTC)[reply]

You must have been too late to catch my other link typo, which I fixed before your response. I would love to own an E-6B. Tragically, all I own is an E6B. I concur with your sentiments; "if someone is going to call their stuff the same as something that already exists..." but as you may surmise, namespace aliasing is the norm in technology. Inevitably, the "most important" incarnation is the one the name sticks to, irrespective of chronological order. Nimur (talk) 17:07, 8 June 2014 (UTC)[reply]

Many computer programming languages try to stay backward compatible, so you can continue to use "deprecated" functions, etc., even in newer versions. Therefore, you could use very old documentation, although then you won't take advantage of any new additions to the language. In Fortran, for example, I don't need to use much that was added after Fortran 77, on a regular basis. So, a 35 year old book would still be of use to me. StuRat (talk) 16:21, 7 June 2014 (UTC)[reply]

I'd guess publishers stop selling a book when customers stop buying it. For paper books there may be some minimum size for a print run. This may lead to books being out-of-print despite the persistence of some long tail demand. For Ebooks this would not be an issue, since the publisher's computers could continue to make individual copies on demand indefinitely. Older book titles may pre-date the rise of Ebooks and thus not have Ebook editions, unless the publisher created them more recently. If you have an old IT book, you might want to look for newer books or articles on the same topic(s), and see if the publisher or author maintain an erratum. The Wikipedia articles about IT topics may provide information about their current status and changes that may have occurred since the dates of your books. Perhaps as a general rule for IT topics, the more popular and widely used it is, the more likely it is to change rapidly. For example, whole shelves of books about Microsoft Windows become obsolete each time another Windows version hits end of life, but that is rarely an issue because the huge market size insures there will always be plenty of books about the current versions. Libraries may have stockpiles of old IT books, thus functioning as a kind of time machine in which you can explore the forgotten history of computing. --Teratornis (talk) 21:46, 10 June 2014 (UTC)[reply]

I'm not sure if they fall in your "IT" category, but references for computer science and many specialized sub-fileds such as disrete control theory can still be useful decades after they were first published. Even if new techniques have been developed, many of the older methods are used as standards because they are simple to understand and implement, and fancier methods aren't used unless something makes them necessary. I'm referencing work by Bernard Widrow that was published in the 1960's for my current project, but also combining it with an much more complicated derivative of it published in a book from 2010. The Art of Computer Programming was first published in 1963, and it can still be a great reference. I once had a professor assign a project to implement a certain algorithm in a parallel machine, and it was intended to be a demonstration of how some algorithms simply can't be parallelized. I looked up the problem in The Art, and read the analysis of the algorithm along with how it was originally derived. Knuth hinted that by raising the order of a polynomial in the derivation a parallel version could be found, so I was able to successfully parallelize the "unparallelizable" problem using a 50 year old reference. It hasn't been as useful in my professional work, but I do find an excuse to look something up in it every once in a while. Katie R (talk) 14:33, 11 June 2014 (UTC)[reply]

How Windows Updates on a computer finds the Windows Updates servers

How does Windows Updates find home to phone? Say I had an XP box that hadn't been connected to the internet since late 2001 and I connected it shortly before support for XP had recently ceased. I know the source code isn't available, but maybe there's a source from someone who knows explaining it. Given that the IP of the Windows Updates server(s) quite possibly has changed over time, it's probably not a hard-coded constant in their source. Is it a URL to a domain name that Microsoft maintains that DNS servers can route to and instead of replying as a webserver to http requests from browsers, only listens for and responds to Windows Updates-format requests? 20.137.2.50 (talk) 17:01, 6 June 2014 (UTC)[reply]

It looks up update.microsoft.com, and most of the files it needs are hosted at update.microsoft.com/redist. If you're a developer, [[1]] explains how to use the Windows Update Agent API, and a lot of the details of how the whole process works can be worked out from that documentation. A lot of the files can be directly downloaded if you know the URL because there isn't really a need to restrict access to specific clients. Katie R (talk) 18:32, 6 June 2014 (UTC)[reply]

Question involving highly compressing MP4s with Format Factory

What's the smallest format an MP4 could be converted to using Format Factory that would still be readab;e as a music video by a Samsung Smartphone? The desire is to send a 6.9MB video (less than 3 minutes) in a reasonably clear format. The target phone has a very low download speed network that charges ridiculous per-minute prices for downloads. If FormatFactory has a more compressed video format than MP4 it will be of great help, but I am an old person who owns a flip phone in case of medical emergencies only. Thanks. μηδείς (talk) 23:32, 6 June 2014 (UTC)[reply]

I don't use Form Factory so I can't help answer your exact question, but if you don't have any means to compress the file to less than 1MB then is there no way you could send it by e-mail to be downloaded over a cheaper link and moved onto the phone? Dbfirs 16:27, 7 June 2014 (UTC)[reply]

I believe the charges anywhere in Cuba are quite high, at least for the person concerned. As to the question, some more info on the phone model etc would help (for example, there's no point sending any video with a resolution higher than the phone screen although in a case like this I wonder if you'll have to reduce a bit more, perhaps the frame rate too) and also the video. I'm not sure if there's really any better choice than H264 for the video and AAC for the audio which I suspect the file already if it's MP4 (although you'll probably want HE-AAC which it may not be) but you could probably get a watchable 3 minute video for less than 6.9MB. Whether you can with 'Format Factory' (whatever that is) I have no idea. P.S. For clarity Opus (audio codec) would probably be a better audio codec and decoding it should be well within the means of most smart phones, but I don't think support is great in standard applications so I wouldn't recommend it) Nil Einne (talk) 13:37, 8 June 2014 (UTC)[reply]

Reformatted title from "What's the smallest format an MP4 could be converted to using Format Factory that would still be readab;e as a music video by a Samsung Smartphone" to "Question involving highly compressing MP4s with Format Factory" due to ToC stretching. -- 140.202.10.134 (talk) 16:12, 12 June 2014 (UTC)[reply]