Wikipedia:Reference desk/Archives/Computing/2011 August 31

src: upload.wikimedia.org

August 31

Video Wikipedia:Reference desk/Archives/Computing/2011 August 31

Do I have Bitcoin straight in my head?

I'm trying to wrap my head around Bitcoin. I get the part about having a wallet with coins in it and then spending the coins on real world merchandise. What I'm not sure about is the "mining". Do you get more bitcoins by basically solving various math problems (simplification, I know) and then presenting the result? So basically, you're being reimbursed for your processor churning away at problems? Sort of like a distrbuted computing effort like Folding@Home. Right? So then, following that logic, basically the millionaires of the bitcoin economy will be whoever can devote the most computing power to their wallet? Dismas|^(talk) 07:32, 31 August 2011 (UTC)

No. If I am not mistaken, those Bit-millionaires purchased (or mined) large amount of Bitcoins early on before their prices skyrocketed. According to bitcoincharts.com, in 31 August 2010, you can purchase one million Bitcoins for a mere $60,000. Moreover, back then, the Bitcoin "mining difficulty" parameter is far lower than today, allowing those Bit-millionaires to mine large amount of Bitcoin without expensive hardware (GPGPU, FPGA, ASIC, ASCI, etc.) 118.96.155.202 (talk) 11:44, 31 August 2011 (UTC)

I don't see how that answers my question. I'm not asking about Bitcoin millionaires. I'm asking about people earning Bitcoins now. Today. What is involved in this mining thing? I've read our article and it makes no sense to me. Dismas|^(talk) 14:01, 31 August 2011 (UTC)

In all fairness, you do ask about Bitcoin millionaires of the future in the last part of your question. What 118.96 was saying is, "even though you can 'mine' coins, the people who are actually going to be rich are going to be the ones who have probably not spent their time mining coins, or certainly won't be in the future." But yes, essentially your understanding of the "mining" process is correct. Except that the problems aren't, to my knowledge, actually useful ones. The question is whether anyone who spends their time mining will really end up with much -- I suspect that mining is not a terribly lucrative affair, from reading the article. It's sort of the minimum wage of Bitcoin, which will generate resources, but probably not at a pace that makes it actually very useful. --Mr.98 (talk) 14:28, 31 August 2011 (UTC)

The purpose of "mining" is because the bitcoin economic theory is predicated on a steadily increasing quantity of money in circulation. In order to control the rate of introduction of new money, the bitcoins must be "mined" - i.e., must use a certain amount of computational effort. It is very closely related to the concept of the proof-of-work system.

The real question you should ask is not why the mining takes place in the form of solving-a-useless-math-problem. The kernel of your question is, "what economic sense does it make to intentionally increase the quantity of money in circulation?" I don't know that it makes any sense. It will result in inflation, i.e. the same tangible good will cost more bitcoins, several years from now. In order to counter that inflation, the bitcoin designer intentionally prints money - resulting in more inflation. I have a suspicion that the innovators of the bitcoin economy spent more time on cryptographic implementation details than they did learning economic theory. Money creation is the relevant article. Rather than having a central bank who mints new money by printing paper fiat-currency, the bitcoin scheme is distributed, and the "mining" represents proof of work completed, in order to regulate the amount of money you may print per unit of time. Nimur (talk) 18:08, 31 August 2011 (UTC)

For a long time I was curious whether I was just daft and missing something, but I'm pleased to see that others think it is an economically problematical thing as well... --Mr.98 (talk) 18:25, 31 August 2011 (UTC)

In my understanding (and I'm not an economist), the idea is to solve the problem of initially distributing the currency. The "natural" thing to do is to have the originator of the system declare at the beginning "There are 1 trillion bitcoins. You can buy some of them from me, by sending me US dollars/Euro/Zimbabwe dollars." But that sounds a lot like a scam. So, they made it possible to get bitcoins without enriching the organizers of the system. Paul (Stansifer) 22:03, 31 August 2011 (UTC)

Currency economies benefit from inflation because they encourage the participants to actually use their money instead of saving it and hoping the value will go up. If everyone horded the currency, and demand grew, the value of the currency would indeed go up and there would be no encouragement to use it as a currency (it would be in deflation as it's value compared to the value of other things grew), it would thus instead be an investment mechanism but in doing so it would fall under the "Greater Fool" category of investment, meaning the only benefit in buying into the system is hoping that some time later someone else will want to buy in for a higher price. This is a horrible way to run a currency (see: any economy with a deflation risk.) Instead, by creating a system that has a certain degree of self-inflation (as any good Fiat Currency economy does) the participants are encouraged to use their currency as *currency* and not wait around for something to happen to it. If you want to put your money somewhere it will increase in value, you should be investing it in something tangible, like a software company ;-) --144.191.148.3 (talk) 15:18, 2 September 2011 (UTC)

Maps Wikipedia:Reference desk/Archives/Computing/2011 August 31

Does the browser or the server look for "index.html"?

In "Fluency with information technology" Chapter 3 page 98-99, it says that when I point my web browser to http://www.example.com/example/ the browser will automatically look for http://www.example.com/example/index.html and display that if it exists. However, is it really the browser that does this for me, is it not the server that serves this page because it assumes that is what I'm looking for? 83.250.156.56 (talk) 12:14, 31 August 2011 (UTC)

It is the server that does the job of picking the file to show. In IIS you set a priority list of names to look for and it will pick the first one that matches. Another way to behave is to give a directory list instead of a file. Or it could give an error message to say no file. Graeme Bartlett (talk) 12:33, 31 August 2011 (UTC)

Depending on the configuration of the server, the paths may not even correspond filesystem paths (for example, pretty much all of Wikipedia's website is like this), in which case looking for an "index.html" is meaningless. Paul (Stansifer) 12:52, 31 August 2011 (UTC)

As mentioned above, the server is responsible for choosing which file to serve. In fact, many times you may have multiple index files (index.htm, index.html, index.php), in which case the server has a (configurable) order of prioritization for which type to serve.
For some extra reading, as mentioned at by Paul, there are server extensions such as mod_rewrite which appear to be serve files/folders, when in fact don't exist on the server at all! TheGrimme (talk) 14:54, 31 August 2011 (UTC)

Thank you all for your answers, so I think we can agree that this expensive book is flat-out wrong. I will read the rest of it with caution. 83.250.156.56 (talk) 18:42, 31 August 2011 (UTC)

src: upload.wikimedia.org

Digitise my life: step one - acquire a scanner

Hello fellow ref desk people. I hang out mainly on Language, Miscellaneous and Humanities, where I've just contributed an answer with an academic quotation, personal knowledge, and internal links to our articles. While composing that response, it occurred to me that I should ask for help as well as giving it. So, if this question is too over-arching to be properly answered here, if I am transgressing the boundaries of this desk, my apologies; responses at my talk page would be just as welcome. Here goes:

I need to digitise my life, and the first and largest task is getting rid of paper, both my own documents and books. To that end, I have determined that I need to acquire a scanner. I've never had one, and looking at consumer review websites such as Reevoo has proven a bit overwhelming, as I can't match up what I think I need (which is in any case open to discussion) with the 121 items listed. In some cases it isn't even clear whether the scanner is a flat-bed, which wouldn't really help when I have thousands of pages that need to be processed.

1. What key words do I need to be looking for, or key features for that matter, in the scanner itself?
2. What kind of software should I be looking for? Is it easy to save a scanned document both as an image (pdf, or are there better alternatives?) and as a searchable (and alterable) text file? (Or can image files be searched for text these days? I am no expert on what software is capable of.)
3. How can I organise this huge amount of material so I can find the documents I need, while only half-remembering phrases within them? My preference is to store the material online, but I am aware that services such as Dropbox can change their T&Cs and attempt to "own" my material. What are my alternatives for storage?
4. Is it necessary to buy a new scanner, or can I lease one for a couple of months? (I am in the UK.) Are second-hand ones notoriously unreliable?
5. There must be people who've already digitised their lives (not just paper, but old music and photos too). Are there blogs or websites you can recommend, or advice elsewhere on tackling a big personal project like this?

I am not going to succumb to the temptation of doing a Princess Leia ("Help me, Computing Refdesk; you're my only hope") -- what scanner should I get, what is the best software, etc. But at least I can ask, how can I determine what scanner to get, how can I choose intelligently among the software available, etc.

OK guys, do your worst. Many thanks! BrainyBabe (talk) 15:21, 31 August 2011 (UTC)

Consider a ScanSnap S1300 from Fujitsu. Google 'scansnap s1300' for lots of reviews and one or two websites that are dedicated to the Fujitsu scanners. It has an automatic sheet feeder (cannot do bound books) and handles 8 ppm. It is delivered with OCR software for both PC and Mac, scans both sides of the page at once, can produce searchable PDFs using OCR, and costs between $200 and $300 US. I first became aware of these scanners per a Dave Winer review. EdJohnston (talk) 15:40, 31 August 2011 (UTC)

Scanning books yourself is, for the most part, bonkers. A consumer grade flatbed scanner takes ~30 seconds to scan each A4 exposure. So, with time for manual pageflipping and breaks, it'd take you most of a day to scan a paperback you paid a fiver for - that's an astonishingly poor use of your time. Automated book scanners will do it in a couple of minutes, but are very expensive (they're intended for libraries and archives). Office-grade scanners (which are essentially digital photocopiers) can scan much quicker (maybe 1 or 2 seconds per side) but are still large and expensive, and to scan a book with them you have to guillotine the spine off (and render it a loose collection of sheets) so you can put it through the sheet-feeder. Commercial scanning companies can do this kind of thing, but for reasons of copyright they won't do your book collection, and it would still not be at all cost effective. If you must have electronic books, it's be easier and cheaper to just buy them new as ebooks for an ebook reader like an iPad or Kindle. And have fun reading their terms and conditions, to see if you're really buying the book, or somehow acquiring a licence to read it under certain circumstances. -- Finlay McWalter ? Talk 15:42, 31 August 2011 (UTC)

What are these automated book scanners that can scan a book in a couple of minutes? Even with destructive scanning (where you just gut the pagers and run them through a feeder) it's going to take longer than that. I'm not sure you're going to find a scanner that operates faster than a current photocopier does. --Mr.98 (talk) 18:19, 31 August 2011 (UTC)

Kirtas Technologies claims up to 2900 pages per hour with their fanciest automatic scanner; that would polish off my copy of The Catcher in the Rye in under 4 minutes. Video of a slower Kirtas scanner (running at about one page every 5 seconds) is here. This experimental scanner runs at 200 pages/min (which would be 12,000 pages/hr if you could sustain it). These high-speed scanners all rely on high-resolution high-speed photographs rather than dragging a scan element over the page (or the page over the scan element), and use a known mapping pattern to reconstruct the page geometry so they can reverse the distortion caused by the page not being flat. Google's "maze" pattern for their scanner is discussed in this patent. I don't know how fast Google's scan is (I believe they use their own equipment, at least for non-special books). Some people have experimented with blown air to separate and flip pages faster than mechanical arms can safely do alone (patent) but I don't know if anyone has commercialised that (getting it 100% reliable, where it doesn't occasionally skip stuck pages, sounds hard). -- Finlay McWalter ? Talk 19:42, 31 August 2011 (UTC)

If you have a large collection of papers (say legal and financial stuff) it's best to approach your local document imaging company (in the yellow pages under "documents" or "photocopying") and they can shove that stuff through their high-quality sheet feeders, give you a DVD with PDFs on it, and charge you for labour. If you need to retain this stuff for legal (e.g. tax) reasons you probably still should keep the paper documents somewhere, and (as with all digital stuff of any value) you need multiple redundant backups in different physical locations. -- Finlay McWalter ? Talk 15:49, 31 August 2011 (UTC)

You'll also need some sort of OCR software for "reading" the images.Smallman12q (talk) 17:35, 31 August 2011 (UTC)

To clarify: OCR software makes PDFs searchable. You can read PDFs without them being OCR'd. --Mr.98 (talk) 18:19, 31 August 2011 (UTC)

It seems very unlikely to me that DropBox either would claim ownership of uploaded material, or could get away with it. Most likely is they will claim you give them a limited license to copy it for the purposes of uploading (just so nobody can charge them with copyright infringement for holding copies of your stuff for you), but that's probably it. I don't think Dropbox is a bad option if you want to keep a lot of files synced between computers. --Mr.98 (talk) 18:22, 31 August 2011 (UTC)

Hi everyone, OP here, thanks so much for the answers so far, and please keep them coming. Thank you EdJohnston for the ScanSnap recommendation; I've been reading about it, and looking on YouTube, and it seems to sync well to EverNote, which I've used a little (but only a little). That would be one answer to question 3 re storage. To clarify, Finlay McWalter, yes, I was considering guillotining some of the out of print books, to make them scannable quickly, with a machine that accepts chapter-long chunks. So, a bit bonkers, but not entirely: I don't (yet) own an e-reader, and it leaves a bad taste to be asked to pay again for an electronic version of a book I already have forked out for. (I am very aware that this issue will look completely different in ten years' time, but I am facing it now and need to deal with it with today's tools and work-arounds.) I have a related question re OCRing: once a document has been scanned in, how easy is it for me to extract the data in order to change it? I have quite a few print-outs of documents I wrote, but I've lost track of the digital file. I want to scan them in to re-capture them, manipulate them, bring them up to date and re-use them. Is this what OCR is about, or have I got the wrong end of the stick? BrainyBabe (talk) 19:40, 31 August 2011 (UTC)

It is not easy to manipulate or change scanned in data. Basically the computer sees it as an image. OCR helps the computer understand that the image contains text, which make it searchable. Now in theory if it is searchable you could also copy and paste it into, say, MS Word or something like that. But the problems are that 1. the OCR isn't perfect, so you spend a huge amount of time picking out typos whenever you do something like that, and 2. OCR is just awful at figuring out formatting, which means your pasted in text is usually a huge mess. So in my experience it's rarely time efficient to scan, OCR, paste, and clean up -- it's usually not any better (and in some ways more difficult and frustrating) than just retyping (or paying someone else to retype) the originals. (The difficulty and frustration comes from typos and errors, some of which are very hard to spot, like when your lower case L's render as 1's or vice versa.) I think of PDF scans as being convenient, searchable photocopies -- not anything like editable documents. --Mr.98 (talk) 20:31, 31 August 2011 (UTC)

(ec)Regarding editing: I've most experience with ReadIris PRO (but I don't especially recommend it, as I haven't used others and maybe they're better). In addition to annotating a scanned PDF with invisible text in the background (allowing the PDF to be searchable) this will generate Word documents using one of two strategies. One is to try its hardest to reproduce the actual physical format of the document - to do this it uses lots of awkward little text frames positioned just so on the page - you get the page pretty much verbatim, but it's a bit painful to make changes. The other strategy eschews formatting acuity for editability, producing a regular word document with normal paragraph separators - this doesn't physically resemble the scanned document much, but the text is in a straightforward format. OCR is pretty reliable when dealing with reasonable resolution scans of decently printed material (reasonable, but not perfect), and its good at handling text and basic numbers. But feed it stuff where format is meaningful (like tables) and it tends to fall apart. Part of the reason for its success is that you tell it (or sometimes it tries to infer) which natural language the text is written in, and it runs all the text its recogniser emits through a dictionary - as with any such automatically-fix-errors strategy, it takes badly to unfamiliar words and particularly words or names in foreign languages. So OCR is certainly preferable to typing the text in again, it can't be completely relied upon to reproduce many documents absolutely verbatim. -- Finlay McWalter ? Talk 20:36, 31 August 2011 (UTC)

One thing to consider when buying scanners is the duplex mode. The cheapest scanners have no document feed at all: you have to place each document on there by hand - that's obviously a lot of work if you're doing more than a few pages. Next up are half-duplex feeders, which feed one side of each sheet past the scan element (like the way paper feeds though a cheap home printer); so if you had an 8 page book, the scan would just have pages [1,3,5,7]. If you flipped the stack over and rescanned you'd get [8,6,4,2], and hopefully the scanner would come with some software that reordered and zipped these together to make [1-8]. Next up is a mechanical duplexer (which a lot of office photocopier-type equipment has). This feeds one sheet through, scans it, then sucks the sheet back through a mechanical thingy that does a 3-point-turn and delivers the reverse surface for scanning - this is very impressive when it's going (and the expensive machines that professional document companies have can run at some scary speeds) but it's a bit Heath Robinson-ish, and may not be suitable for the cheap paper on which some books are printed. Lastly there's the kind that has a simple once-through paper path but has two scan elements (one on each surface) and the software brains to scan both sides concurrently and then emit the pages in the right order. If you're guillotining books, and don't consider clearing paperjams to be a hobby you're interested in pursuing, this last approach seems the simplest for you. The Fujitsu machine EdJohnston recommends works this way. -- Finlay McWalter ? Talk 21:04, 31 August 2011 (UTC)

Sorry if it was said, I couldn't bring myself to read all that. Scanners are expensive (even small ones that all media won't fit on), get yourself a digital camera (or two, you might even have it/them already!). Hacks for more efficiently using them to rapidly digitize dead tree media abound online. ¦ Reisio (talk) 09:04, 2 September 2011 (UTC)

Thanks everyone. This has been very educational, both for concepts and words ("duplex" not just being a kind of house, for example). I looked at the first suggestion, did my "savvy consumer" research, watched a good number of how-I-use-it videos, and decided to go for it. None in the shops yesterday, so I ordered today, and am now awaiting delivery of a ScanSnap S1300. I'll let you know how it goes, and am likely to come back with another question, or series of questions, then. (I'm still open to suggestions for good "digitise your life" blogs.) Many thanks again. BrainyBabe (talk) 14:09, 2 September 2011 (UTC)

$300, rough. High res digital cameras are a fraction of that. Don't get me wrong, I have access to a feed scanner that does duplex myself and it's great for stuff that fits into it that I'm willing to debind. ¦ Reisio (talk) 03:08, 4 September 2011 (UTC)

src: upload.wikimedia.org

Net Access

This was originally posted to Wikipedia talk:AutoWikiBrowser, but it seems like it belongs here instead.

My computer says I am connected to the internet...yet when I try to load a page it won't load. I can use skype but nothing else loads. Safari starts to load the page but I only get a spinning wheel. I have tried on multiple computers and my PS3 and no pages will load. It started today 30 Aug 2011. Is anyone else experiencing this issue? And how do I fix it? I am using Talk Talk and in the UK.. Thank you -- Preceding unsigned comment added by 68.56.111.114 (talk) 17:10, 30 August 2011 (UTC)

If you connect through a router, it may have mismanaged the local IP addresses it assigns to your computers. Try turning the router off and on. Certes (talk) 20:56, 31 August 2011 (UTC)

Your computer may also be set up to use a proxy you cannot access. Check the proxy settings. Also try if you can ping the real net (open a terminal and type e.g. 'ping www.google.org'). Finally, you may have trouble with DNS. Try 'nslookup www.google.org'. --Stephan Schulz (talk) 22:23, 31 August 2011 (UTC)

src: blogs.bodleian.ox.ac.uk

pcie half card with Wimax and WiFI

I've found several out of date cards that offer this dual capability. What is the latest card that has this dual capability? --DeeperQA (talk) 21:54, 31 August 2011 (UTC)

src: upload.wikimedia.org

Telecommuting programming job boards?

What are the best web sites for hiring programmers for remote project-based work? I know about http://odesk.com and I remember reading about some others but can't find them now. 76.254.20.205 (talk) 22:53, 31 August 2011 (UTC)

I'm afraid I can't say which are best, but per the article on oDesk, there is also Elance, Freelancer.com, Guru.com and VWorker. --Kateshortforbob _talk 13:28, 1 September 2011 (UTC)

Source of the article : Wikipedia