Wikipedia:Reference desk/Archives/Computing/2011 October 10

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.

October 10

NFT device and online electronic receipts with item list

Some stores that offer credit cards reproduce the paper receipt's listing of items online in the customer's credit card account. If you use a bank debit or credit card to make a purchase this feature is unavailable. Will any of the Near Field Technologies offer an electronic receipt on the NFT device and online with a item list? --DeeperQA (talk) 02:38, 10 October 2011 (UTC)

Files won't attach to email

I have a an imac. From system profiler:

 Model Name:	iMac
 Model Identifier:	iMac11,3
 Processor Name:	Intel Core i5
 Processor Speed:	2.8 GHz
 Number Of Processors:	1
 Total Number Of Cores:	4
 L2 Cache (per core):	256 KB
 L3 Cache:	8 MB
 Memory:	8 GB
 Processor Interconnect Speed:	4.8 GT/s
 Boot ROM Version:	IM112.0057.B00

I'm trying to attach a Word document to an email (yahoo mail). I have done this thousands of times from my near identical computer at work. This is not a problem with Yahoo mail; I have attached files multiple times in the past week using the same email account. But when I try from my home computer all this week (all files, I've tried pdfs and others) the attachment bar just spins, never uploading to the email. Anyone have any ideas?--108.54.26.7 (talk) 05:43, 10 October 2011 (UTC)

Have you tried a different browser? Dismas|^(talk) 05:49, 10 October 2011 (UTC)

I have experienced this recently when Adobe Flash is used as the file uploading tool. I upgraded my Flash client and the problem went away. TheGrimme (talk) 16:21, 10 October 2011 (UTC)

Well I don't know what was causing it, but taking a cue from that, I just did an upgrade of everything I could think of (including flash) and then restarted (though I had restarted before) and it's working now!--108.54.26.7 (talk) 22:49, 10 October 2011 (UTC)

I have a secondary question:

Spam then delete, or just delete

Spam, spam, sausage, spam, spam, bacon, spam, tomato and spam (sorry Monty Python on the brain). Does Yahoo Mail learn spam if I send things to the spam folder rather than deleting them outright? I mean, when I get some Viagra ad (I get about 12 of those a day, even though I hopefully won't need it for a few decades) it is helpful to first click on the spam folder, and then empty the spam folderm, rather than just deleting, or am I according the software more intelligence than it has? And oh, yes, does anyone know what the airspeed velocity of an unladen swallow is?--108.54.26.7 (talk) 22:49, 10 October 2011 (UTC)

Services with a dedicated Spam folder almost certainly have one because they do Bayesian spam filtering, so yes. Also, it's probable that spam messages are deleted after a certain period of time, and they might not count against your space limit in the first place, so you can probably remove the "empty spam folder" step from your process. Paul (Stansifer) 01:42, 11 October 2011 (UTC)

Obligatory Monty Python reply to the last question: An African or European one? -- 188.105.123.207 (talk) 20:07, 12 October 2011 (UTC)

ask.com

Hello I need a way to remove the ask.com toolbar completely from my computer whenever I try to uninstall it a window pops up and says close all Windows that there are no windows open all this program is malware and says this to prevent you from installing it. There is a removal to the tool somewhere on the web but I don't think it works with the current version. Please help me — Preceding unsigned comment added by 66.209.177.15 (talk) 10:49, 10 October 2011 (UTC)

Either go to "Add and Remove programs" in the control panel, or use a third party tool like CCleaner. KyuubiSeal (talk) 02:13, 11 October 2011 (UTC)

Windows Movie Maker

Is there any way to get the old timeline format in the version of Windows Movie Maker that comes with Windows 7? The series of thumbnails stacked in rows and columns on the right pane is an inferior user interface in my opinion. 20.137.18.53 (talk) 16:19, 10 October 2011 (UTC)

The articles on Windows Movie Maker (the version with XP and Vista) and Windows Live Movie Maker (the version with Windows 7) say that the new version doesn't have the timeline, is missing many other features, and many people regard it as significantly inferior. As to why Microsoft did this, I don't know. --Colapeninsula (talk) 16:31, 10 October 2011 (UTC)

Microsoft have achieved the feat of making a dumbed-down video editor even dumber which may give a clue to which way the inevitable next improved version of Windows will lean. In Vista I find no problem with having both WMM and WLMM installed and leave it to the reader to guess which I actually use. Cuddlyable3 (talk) 16:27, 12 October 2011 (UTC)

None?20.137.18.53 (talk) 14:42, 13 October 2011 (UTC)

Are you okay without HD? There's a download here: http://windows-movie-maker.en.softonic.com/ This article says that it is compatible with 7. -- Zanimum (talk) 00:59, 14 October 2011 (UTC)

Ew, just reading down below: it can't even capture video. So you'd have to capture with the program native to 7, then import it to 2.6, the old version they've made available. -- Zanimum (talk) 01:00, 14 October 2011 (UTC)

nVidia Graphics Card on Amazon.com - 2¼ Times as many CUDA cores at a Third of the Price?

Hi.

  I've been shopping for a new graphics card, and came across this on Amazon.com - I need advice on whether this is too good to be true.

  The nVidia Tesla C2050 Graphics Card has 448 CUDA cores and is being sold at $2,200 while the nVidia GeForce GTX 590 Graphics Card which has 1024 CUDA cores is being sold at just $749! How can a graphics-card with over 2¼ times as many CUDA cores be priced at about one-third the price of its less-powerful counterpart? Is this a pricing mistake on Amazon's part?

Many thanks, Vickreman.Chettiar 17:43, 10 October 2011 (UTC)

Number of cores is only one benchmark to compare these two hardware units . Have a look at C2050 and GTX 590 technical briefs from Nvidia. Two things pop out at me: the GTX series does not support double precision (64-bit) math, and it is not Fermi-architecture (among other things, this means multitasking support is limited). Depending on your CUDA program/algorithm needs, this might mean that the cheaper GTX-series GPUs outperform the more expensive C20xx series Fermi chips - but you really need a design review to determine how the specific technical specs map on to your particular program. Nimur (talk) 17:57, 10 October 2011 (UTC)

Minor correction - GTX supports 64-bit math, but at a performance-cost overhead, compared to the Fermi C20xx chips. I also noticed that the chips have different clockrates and memory speeds, different on-chip cache sizes, and other technical specs that will affect performance. Nimur (talk) 18:09, 10 October 2011 (UTC)

In addition the Tesla has features which support its use in a massively parallel supercomputing architecture - twin independent DMA controllers, Infiniband, and ECC. On raw single-precision FPU performance the GTX is indeed about double the performance of the Tesla, but as Nimur notes memory and cache throughput and coherence is a major factor when considering the performance of an actual algorithm on a given piece of hardware. One-for-one a single GTX will probably outperform a single Tesla in many circumstances, but they're selling the GTX for use as a graphics device, to be used in combination with 0 to 3 of its kind. They're selling the Tesla as a compute platform, to be used in combination with dozens of its kind. If you were building a large compute resource, you might find it more cost effective to have regular PCs with two or four GTXes and handle the interchange of data between them with a PC interconnect like 10Gb ethernet (particularly if there's lots of computation and lots of data coherence with relatively little inter-node data exchange) but for many compute jobs the Tesla can build systems which will scale to do the job and the GTX won't. TD;DR: what Nimur said. -- Finlay McWalterჷTalk 18:30, 10 October 2011 (UTC)

  I read the two technical briefs, and the GTX590 still looks more powerful to me than the C2050; as the GTX590's memory is 1.48x faster with 2.27x greater bandwidth. Here's comparing their major parameters.

Parameter			GTX590		C2050
CUDA Cores		1024		448
Memory Speed		1707MHz		1150MHz
Memory			3072MB		3072MB
Memory Interface	GDDR5		GDDR5
Memory Interface Width	384-bit		384-bit
Memory Bandwidth	327GB/s		144GB/s
Maximum Power Draw	365W		238W

  In simple terms, what does the C2050 have that the GTX590 does not? Regards, Vickreman.Chettiar 21:59, 10 October 2011 (UTC)

Here are some line-items only present on the C20xx series:

• ECC memory support
• Fermi architecture
• Support for 6 GB of attached RAM

You can probably pull a few more if you stare hard at the specs comparatively. Are these line-items worth the incredible extra cost? Well, that depends; and it's sort of subject to your own needs, budget, and opinions. (The reference desk is not really the place for opinions, nor for original research, but I feel like I am qualified to comment; hopefully this self-disclaimer will preempt any WP:OR chastisement). Some time ago, I gave an invited presentation at University of Nevada on exactly this topic: building a low-cost CUDA computer using "gamer" graphics cards, instead of the expensive Tesla series. Depending on your specific needs, you may actually outperform the Tesla-branded GPU by using a GTX-branded GPU. As a general trend, GTX (they used to call them GT or G9x) chips are faster, in clock rate; they sometimes do contain more cores than the Tesla models; and while they lack some extensibility and some advanced capabilities, these G9x and newer GPUs do support Cuda 2.0+ (your GTX590 has a GF110 which supports CUDA 2 / Compute Capability 2.0 - with the newer PTX extensions, I believe). There are definitely some advanced features of the Teslas, as described in Nvidia's product marketing literature, that are not present on the lower-end chips; and in edge-cases, the Teslas will outperform the G9x/G10x/G11x cores. Nvidia also packages Tesla branded GPUs on better boards - meaning better thermal/cooling, better memory, more memory, ECC memory, and so forth. If you are investing in thousands of rack-units, you probably want the Tesla chip. If you're a small-scale researcher or an enthusiast programmer, you can often do better with the "commercial off-the-shelf" GTX line, aimed at gamers, at greatly-reduced cost compared to the Quadro and Tesla line. Nimur (talk) 16:01, 11 October 2011 (UTC)

Thanks, Nimur, now I have a much clearer idea of which card to get. My application is AutoCAD, which primarily needs good graphics rendering, so I'll go with the GTX590. Vickreman.Chettiar 19:28, 11 October 2011 (UTC)

Combine many html files into one complete version

I have about 200 copies of the same html file, but each copy is missing different lines that others have. There is no complete version of the file. I want to somehow combine all the 200 different copies into one complete file. Doing it manually is possible but obviously very time consuming so an automatic way to do it would be great. I'm thinking some kind of diff program that can analyze the files like Wikipedias edit history thing and output a complete version. In case I haven't been clear in explaining the problem, the files are something like this (but obviously more complicated);

Lines in file 1
1 The
2 And
7 Whatever
9 Computer 

Lines in file 2
1 The
6 Internet
7 Whatever
8 Hello
9 Computer 

Lines in file 3
2 And
3 Save
4 Example
5 Time
9 Computer 

etc

and I want to combine them into one file which looks like

1 The
2 And
3 Save
4 Example
5 Time
6 Internet
7 Whatever
8 Hello
9 Computer 

How can I achieve this? What programs could combine the files? I'm on Windows 7 and have limited knowledge of linux programs, but can use them if there's a Win32 port available. 82.43.90.142 (talk) 23:56, 10 October 2011 (UTC)

It sounds like you want to do a merge. Unfortunately, the only standalone merge programs that I know of don't run on Windows (except if you're going to do a lot of work to get them working). You might want to consider using revision control software if you have files that regularly drift out of sync, which will automate the process of synchronizing and merging. Unfortunately, the learning curve on those tools is steep. Paul (Stansifer) 01:37, 11 October 2011 (UTC)

On Linux it would be easy. You could combine all the files into one, sort the lines alphabetically, then cull any repeated lines. Dismas|^(talk) 02:19, 11 October 2011 (UTC)

I suspect the line numbers are just there for illustration, and not part of the files themselves. --Mr.98 (talk) 11:25, 11 October 2011 (UTC)

Yes 82.43.90.142 (talk) 12:20, 11 October 2011 (UTC)

Just GnuWin32's coreutils package should cover it. Doesn't look like it adds itself to the user's path, though, so you'll probably want to do that. Then hit Windows_key+r, and type cmd to get a prompt. You'll probably start at a path like C:\Users\youruser\, from which you could, for example, run cd Desktop to get to your desktop directory. Then assuming the files were there, you could run something like cat *.html | sort | uniq > newFileWithNoDupes. Of course that assumes your HTML doesn't have any sort of header or footer code. The easiest way for a Windows user to remove headers and footers would probably be to install a proper text editor, like Geany, open all the files in it (select them, right click, Open in Geany), then select the header (and later the footer) and do a find & replace for all documents (CTRL+h). It might seem like a lot but it isn't really. Visit http://webchat.freenode.net/?nick=catHTML&channels=##windows for real time help if you can't get it figured out. ¦ Reisio (talk) 02:33, 11 October 2011 (UTC)

I would recommend Cygwin over GnuWin32. However, this may not satisfy the question-asker's problem, because the command line you just gave would sort the list, before it removed the duplicates. He might need to preserve the order. That said, I'm not sure of an easy way to reliably merge that many files without sorting them. APL (talk) 12:22, 11 October 2011 (UTC)

Merging can't always be done automatically. Consider the files "A X B" and "A Y B": human intervention is needed to decide whether "A X Y B" or "A Y X B" is right. Paul (Stansifer) 14:16, 11 October 2011 (UTC)

Nah. ¦ Reisio (talk) 16:20, 11 October 2011 (UTC)

Yes, I think there's a shortage of information here: does the content of these files follow some sort of explicit precedence-rules, such that it can be easily determined whether line 1 from file 1 belongs before line 1 from file 2? If there's some explicit rule, you can design a regular expression, or a script, or a full-fledged text-parsing program, to implement the merge. If no such rule exists, you can simply concatenate the files together. If a rule exists to order the lines, but you are only able to state it in a vague way, you must first address that problem - formalize how you want to merge lines, and we can help direct you toward a best-practice solution that implements your needs. Nimur (talk) 15:34, 11 October 2011 (UTC)