Wikipedia:Reference desk/Archives/Computing/2011 February 25

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Computing desk
< February 24 << Jan | February | Mar >> February 26 >
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.

February 25[edit]

Free software definition (FSF) vs Open source software definition(OSI)[edit]

Hi all,
According to the definition of Free software by FSF and Open source software by OSI, I realize that:

  • 6th criteria of OSI meets freedom 0 of FSF
  • 2nd and 3rd criteria of OSI meet freedom 1 and 3 of FSF
  • 1st criteria of OSI meets freedom 2 of FSF

So, IMHO, every open source software is free software, too. However, FSF states "nearly all open source software is free", it means "exist open source software is not free".
Any idea? If such software actually exists, please give examples and explain what is wrong with my idea.
-- (talk) 06:31, 25 February 2011 (UTC)

Take a look at this and this list. The intersection of both lists is the list of software licenses which are open source but are non-free. (talk) 09:49, 25 February 2011 (UTC)
Examples: the Artistic License (which the FSF says is too vague), and the NASA Open Source Agreement, version 1.3 (which the FSF says doesn't allow you to incorporate other people's work when you modify the software). Possibly also the RPL, but I think the FSF may be talking about an old version. Paul (Stansifer) 14:24, 25 February 2011 (UTC)

Google Android[edit]

If Android is a free software (and open source) mobile operating system, why won't its competitors, say, start reverse engineering Android's API from its source code and start adding "Android application compatibility" to their own operating systems?

Second question, can any handset manufacturer actually use Android as their handsets' operating system without paying anything to Google?

Thanks for any answer. (talk) 09:30, 25 February 2011 (UTC)

Per Android (operating system), the software is released under open source licences, and so yes, any manufacturer can use it without making payment to Google (or anyone else). There does not seems to be any licence-baed reason why a manufacturer of a non-android phone could not provide android emulation for the purposes of being able to use android apps. However I suspect there might be some interestingn engineering issues involved. --Tagishsimon (talk) 11:25, 25 February 2011 (UTC)
For the first, they can and have: see Dalvik Turbo virtual machine. For the second, yes, of course, just as I can run Wikipedia's software on my own computer without paying the Foundation anything. Marnanel (talk) 12:10, 25 February 2011 (UTC)
From what I heard, phone mfgrs can ship Android on their phones at no charge, but if they want the phones to have access to the Android Market, they have to pay Google. I dunno who cares about Android Market, which sounds almost like MS Windows download sites in terms of virus infestation. (talk) 12:48, 25 February 2011 (UTC)
As for allowing users to run free Android apps on their phone... Phone companies in the United States do not appear to be interested in that. They want to charge customers to download and run similar apps. Allowing them to do something free is simply a loss of income. -- kainaw 13:27, 25 February 2011 (UTC)

There is risk in even acknowledging you have a competitor, particularly when you (can easily argue, as Microsoft or Apple) are already ahead — you may inadvertently inform the consumer of an option they weren't previously aware of. Keep in mind that your average consumer is not the sharpest tool in the shed. Once competition becomes closer, of course, other options surface. ¦ Reisio (talk) 21:41, 25 February 2011 (UTC)

(OP here) Thanks for all the responses. To tell the truth, I actually asked this question because I really can't believe Google is actually releasing Android for free, with no strings attached, even for its "corporate customers" (competitors and handset makers) which obviously have a lot of money (and do not mind spending them). Even "free software" Linux distributions are sometimes not free. (talk) 03:27, 1 March 2011 (UTC)

Another factor is that Google are updating Android with great frequency, a new version with exclusive new features every few months (e.g. Flash in 2.2). Trying to keep up with that is very challenging - other projects like Wine and Mono often lag far behind the software they are mimicking and seldom support every application. --Colapeninsula (talk) 15:09, 2 March 2011 (UTC)

Lock users in directory[edit]

Image of tree structure [1]

I have a vps running debian OS and would like to create user accounts on it.

I want it so that when the user logs in with sftp, everything in var appears to be their home directory and they cannot cd out of it.

For example, when user3 logs in, they have access to everything in var (read, write, execute) but cannot view (cd) user1 or user2's personal stuff.

How would I go about doing this?

I think I have to do this in chroot, but I have no idea how this would work.

Thanks -- penubag  (talk) 11:20, 25 February 2011 (UTC)

chmod 700 /home/duke/aa/servers/user? (talk) 11:40, 25 February 2011 (UTC)
No. --Sean 18:16, 25 February 2011 (UTC)
To expand, you could do this with chroot, but you'd need to extensively hardlink /var and so forth back in. chroot is useful for specific daemon processes that only require a small amount of explicit access to system directories, but it's a pain for general user accounts. (talk) 11:40, 25 February 2011 (UTC)
chmod 700 just gives them file permissions no? I would like to jail them into var/ which would only hold a few text files. I don't want to open up my vps anymore than this. Could you please explain (in detail) how I could do this? Thanks for your help so far. -- penubag  (talk) 11:45, 25 February 2011 (UTC)

Attach to Email in Acrobat Reader X[edit]

Hello. When I attach files to my email via Adobe Reader X, it cannot launch my mail application. Is it possible to upload attachments to Windows Live Mail through Reader X? Thanks in advance. --Mayfare (talk) 13:06, 25 February 2011 (UTC)

University crest disappeared on Wikipedia Facebook page[edit]

am presuming my university's Facebook Wikipedia page and the original Wikipedia article it is based on are linked. In the past, any changes to the Wikipedia article (such as the logo) have been reflected automatically on the university's Facebook Wikipedia page.

However, the original Wikipedia article has the logo (as the university crest) and it has been like that for some time, but the Facebook Wikipedia page has suddenly gone blank - the logo has disappeared. I don't believe this is the result of a delay in updating as it has been like this for several days now.

Does anyone know why this has happened and the logo has disappeared from Facebook, even though it has been visible in the Wikipedia article?

Here is the university's Wikipedia page:

Here is the Facebook version:!/pages/Lancaster-University/108262075861948?sk=wiki

Here is a reply from the Help Desk (which referred me here):

"The crest information you mentioned is contained within a Wikipedia template -

Reference desk/Archives/Computing/2011 February 25

. If you look at the bottom of the Lancaster University page below the External links section, you'll see more templates (Navigational box templates). These Navigational box templates do not appear in the Facebook page. For what ever reason, it appears that Facebook is transcluding the entire Wikipedia article except for the material contained within a Wikipedia template. As for why this might be, you may want to post at Wikipedia:Reference desk/Computing. (Perhaps the templates were using up too much of Facebooks computer resources.) However, as Rehevkor points out, you'll be best off contacting the admins of the Facebook page. -- Uzma Gamal (talk) 12:08, 25 February 2011 (UTC)"

It might just be that Facebook removes infoboxes because there isn't enough horizontal space, what with the right sidebar of ads. Paul (Stansifer) 16:15, 25 February 2011 (UTC)

It's now back on Facebook. I am presuming someone saw this and fixed it. Thank you. —Preceding unsigned comment added by (talk) 18:01, 25 February 2011 (UTC)

Honeywell 200[edit]

How does the computing power of the Honeywell 200, shown as the futuristic state-of-the-art in the film Billion Dollar Brain compare with your average desktop home computer nowadays? Thanks (talk) 16:18, 25 February 2011 (UTC)

Being Turing-complete, it can complete any operation a modern computer is capable of. However, the tricky part is computing speed. The ENIAC, at least according to its Wikipedia article, was bested in the 1990's by a small microchip. Honeywell is a lot more powerful than ENIAC, and two-three times more powerful than IBM 1401. This site says the 1401 could perform 193,300 8-digit additions or 25,000 six-by-four-digit multiplications. The modern computers are a lot more effective, performing, I'd approximate 0,5-4 GFLOPS. You'll probably find this section interesting as well: Flops#Hardware_costs. I sure did. Zakhalesh (talk) 16:52, 25 February 2011 (UTC)
Instructions_per_second might also be of interest! Zakhalesh (talk) 16:54, 25 February 2011 (UTC)
Furthermore, besides just "computing power," the Honeywell used interfaces that are now considered obsolete: so you couldn't connect it to a modern color VGA monitor, or printer, or to the internet, for example. (Nor even to a keyboard! It had its own proprietary hardware console board, or could connect to a card-reader) So even if running a web browser were theoretically possible, the software support for such an operation is non-existent. A simple day-to-day task like performing a web search requires very little in the way of "FLOPS", but does require a compliant TCP/IP stack, a video output device, and software to support a graphic web browser. A tiny mobile-phone processor may pale in comparison to even a 1950s-era "supercomputer" if you benchmarked FORTRAN on it (... or maybe not, depending on the phone's CPU) - but the mobile-phone is still a more capable computer in a lot of respects. Nimur (talk) 19:53, 25 February 2011 (UTC)
I think most modern phones would beat ENIAC, for example, in raw computing power as well as usability (supercomputers aren't too good in this generally, as they're meant for professional use in any case). Zakhalesh (talk) 20:20, 25 February 2011 (UTC)

If you benchmarked them using Fortan or Assembler, how would they compare? Never mind the graphics. Thanks (talk) 00:00, 26 February 2011 (UTC)

It more or less depends on what you're doing but I'm still on the smart phone's side. I'm pretty confident that it has more processing power, especially if we strip all the unnecessary phone functionality and leave it as a basic general purpose computer. Every machine should be equally good at running their own assembly code (of course bad instruction sets can ruin this), in comparison to FORTRAN which may cause computation speed differences depending on the implementation of the compiler, so when running assembly code only the raw processor power and implementation should matter. Zakhalesh (talk) 07:28, 26 February 2011 (UTC)
When you say smart phone are you thinking something like an Android phone/iPhone/Blackberry/Windows phone or a more ordinary/average phone perhaps with MMS and WAP and J2ME at best which can be thought of as a smart phone but usually aren't included in that category? Nil Einne (talk) 09:09, 26 February 2011 (UTC)
Smart phones as in iPhone, for example. I still do think that a common phone could also compare quite well against ENIAC, but can't be sure. Zakhalesh (talk) 10:23, 26 February 2011 (UTC)

So am I right in thinking that a modern desktop would benchmark at about 50,000 times faster than the Honeywell 200? I estimated that from the CPU speeds. I don't know how to calculate the comparative flop speeds. Thanks (talk) 15:08, 26 February 2011 (UTC)

I can't be sure about the exact number, but you can be sure that a modern consumer grade computer is way (or perhaps even WAAAAAY) faster than the supercomputers from 1970 in terms of hardware. However, just a random thought, but software back then was much more resource-effiecent, as there were no extra computing resources to spare. This might even it the difference out a bit, but not enough for Honeywell to come even close to modern computers. Zakhalesh (talk) 15:15, 26 February 2011 (UTC)

This clip shows more of the machine: I wonder what the point of all the various labels and buttons was? (talk) 14:44, 27 February 2011 (UTC)

My computer has twelve CPU cores each running at 3.3GHz, and I've been able to overclock up to 3.5GHz, so I'd bet my computer can beat the Honeywell 200 anyday. Rocketshiporion 06:54, 1 March 2011 (UTC)

Its rather a disapointment that all the great computing power in a PC is wasted on playing SuperMario rather than taking over the world, Billion Dollar Brain style. Its like having a sleeping genie that is never woken. (talk) 01:31, 2 March 2011 (UTC)


When I try to use --ignore-length in wget to make it ignore the content-Length headers, --convert-links stops working and none of the downloaded pages are converted for offline viewing. Is this a known problem or am I doing something wrong here? I'm using wget 1.12 on Windows 7 (talk) 16:42, 25 February 2011 (UTC)

Browser cannot navigate to any page[edit]

I'm right now at an internet café. Strangely, I can connect to wikipedia and some pages, but not to others like or What's wrong? I am also able to download torrents. The waitress is useless to solve the problem. What could be the cause of this? I get the error Error 105 (net::ERR_NAME_NOT_RESOLVED): The server could not be found. (talk) 18:33, 25 February 2011 (UTC)

Problem solved, without any interference. But, if anyone knows why it happened, I'm still curious. (talk) 19:08, 25 February 2011 (UTC)
That might have been a DNS error. I was going to ask whether the appropriate IP address got you your NYTimes, but if the problem's gone away by itself then it's too late to experiment. (talk) 19:21, 25 February 2011 (UTC)
Well, but can a DNS error occur locally? — Preceding unsigned comment added by Quest09 (talkcontribs) 19:51, 25 February 2011 (UTC)
DNS can be cached by the local operating system, or by the network gateway or router. The DHCP server who assigns you an IP address will usually tell you its preferred DNS server, and if your computer isn't configured otherwise, you will use whatever server the router suggests. If that server is failing, or is untrustworthy, a DNS resolution error can occur (either failure to resolve, or redirection to spoofed websites, or so on). That is why it is so very critical to use HTTPS and other secure technology and to verify presented certificates against known certificates. It is this last step which is often overlooked - you can have a secure connection to a machine, but unless you can guarantee that it is the server you think it is, by comparing its current certificate to a known, trusted certificate, all bets are off. Nimur (talk) 19:58, 25 February 2011 (UTC)
and to verify presented certificates against known certificates. And how do you do that? — Preceding unsigned comment added by Quest09 (talkcontribs) 20:57, 25 February 2011 (UTC)
Your web-browser (or your operating system) uses a chain of trust approach that accepts signed certificates traced back to an acceptable root certificate authority. These RCAs are the "built-in" organizations that your computer or browser is pre-installed to trust. You should verify that you trust every root certificate authority in the list: here is, for example, Firefox's list of trusted certificate authorities. If you don't trust any of those agencies, then your Firefox is not secure. Most people trust these agencies because they have been vetted by the Mozilla Foundation; but that process is not cryptographically secure; nor is it really a solid, safe practice to "outsource" your trust to another agency (even one as benevolent as Mozilla). If you have special security needs, you should clear your root certificate authority list and only add root certificates from agencies you know.
Alternately, you can eliminate chain of trust and only accept specific certificates for specific servers (such as those servers you own, or servers operated by agencies you trust). You would do this by manually obtaining and installing the public certificate from each server - via a secure channel such as receiving the certificate on disk, in person, from the server operator (or via a trusted network) - and then comparing the certificate provided via the untrusted network connection.
For the average user, who does not work with extremely sensitive data, the most practical thing that they need to do is to understand these Firefox error messages (or the equivalent for their web-browser or other internet browser / software). When one of these certificate errors shows up, your browser is warning you that something (either the network, or the server you are accessing) is awry - and this can mean a serious security vulnerability. Nimur (talk) 21:59, 25 February 2011 (UTC)
Addendum - you will, occasionally, see some security warnings to the effect of "... a highly capable attacker could ..." - which is, of course, a euphemism to describe the fact that a malicious government could overtake a root certificate agency and issue arbitrary security certificates that would be accepted by anyone with their default security settings - in other words, the implicitly-trusted list of organizations included with most operating systems and browsers. If, simultaneously, a "highly capable attacker" might overtake the root DNS servers, they would be able to redirect any of your web-browser queries to a completely different server than you expect - while providing you with a completely valid SSL certificate and secure connection. Such cryptographic attacks by "highly capable organizations" might seem very unlikely in this era; it is easier to shut off the internet than to eavesdrop on it. Nimur (talk) 22:08, 25 February 2011 (UTC)

CCleaner & Passwords[edit]


(EC?!) Is there any way to stop CCleaner from deleting all my passwords? I will admit it's a bit of a pain having to login again everywhere, especially when I actually don't know the password and have to request a password reset email (which sometimes doesn't arrive for several hours). Actually, I don't think it is actually deleting the passowrds, because on some sites (notably here) I just have to press the 'login' button and I am taken to a page with my username and password already filled in for me, so I am not sure what CCleaner is doing. In any case, is there a way to prevent this? --KägeTorä - (影虎) (TALK) 19:33, 25 February 2011 (UTC)

Cleaner → Windows → Internet Explorer → uncheck Saved Passwords. ---— Gadget850 (Ed) talk 19:54, 25 February 2011 (UTC)
I will give that a go, but does it work for all browsers? Is 'Internet Explorer' a generic name for all browsers? I don't even use Internet Explorer.... (by the way, it is unchecked anyway). --KägeTorä - (影虎) (TALK) 22:51, 25 February 2011 (UTC)

(EDIT) - 'Session' appears to be the correct option to uncheck. Cheers. --KägeTorä - (影虎) (TALK) 22:57, 25 February 2011 (UTC)

Firefox rendering of small tag (in OS X)[edit]


My browser doesn't currently display any difference in text marked with the small tag (e.g. in this post right there). No choices in 'Preferences' -> 'Fonts and Colors' seem to fix it. I'm using Firefox 3.6.13 in OS X. How do I get it to work properly? Thanks, SemanticMantis (talk) 22:39, 25 February 2011 (UTC)

Not sure about Firefox on OS X, but if it is similar to Firefox on Windows, you may want to check the advanced settings under "Fonts & Colors"— specifically the setting for minimum font size. It may be imposing a minimum font size that Firefox won't go below. -- (talk) 03:19, 26 February 2011 (UTC)
Of course, it's obvious now. Thanks! SemanticMantis (talk) 03:50, 26 February 2011 (UTC)