Wikipedia:Reference desk/Archives/Computing/2009 June 3

= June 3 =

Cheap computer parts
Okay well here is my dilemma. I want to build a cheap computer that I can use just to surf the internet and I was wondering if anyone had any sites to get some cheap parts. I just want to spend like 100 bucks to build it. If you have anything I would really appreciate it. Thanks guys RandomAccessDawg (talk) 03:27, 3 June 2009 (UTC)
 * If you are buying new (or even if you go for second hand) $100 will probably not even be enough for just a case and motherboard - two of your most basic components. http://www.newegg.com/ is a good site for the US (although I am from the UK so others might have better suggestions). --JoeTalkWork 03:58, 3 June 2009 (UTC)

Well I am currently in the process of getting a job so when I do I will probably be able to spend more. What would you think would be a reasonable price for a cheap computer? I am planning on putting the latest Ubuntu on it so the OS will be free. But ya I didn't look at newegg yet but I was planning on it. Thanks man! —Preceding unsigned comment added by RandomAccessDawg (talk • contribs) 04:07, 3 June 2009 (UTC)


 * If I were you, I would look on Craigslist in the "computer" category, where sub-$100 computers are common. They will be early Pentium 4 systems with monitors and not a lot of RAM.  Be sure to try it out for a few minutes on your favorite websites to see if performance is OK before you pay the money and walk away with it.  Tempshill (talk) 04:23, 3 June 2009 (UTC)


 * Yeah, to be sure, you can find computers that will easily surf the web and do lots of stuff for < $100. But finding someone parting with those can be tough. Most non-savy people think their computers are worth a lot more than they are, and most savvy people either keep them around, or don't even bother selling them. Even eBay tends to have really overpriced low-end machines. That said, I think eBay might be a good place to start so long as you're somewhat discriminating about pricing. Same thing for craigslist. I'd say garage sales too, but I think that might be too much time invested for the payoff.


 * Hopefully you can find working peripherals. Just make sure you know how a monitor and keyboard attach, and a network card too. USB is a definite plus for the latter two. Shadowjams (talk) 04:52, 3 June 2009 (UTC)


 * Also, the entry price point on some low end machines (those new non-windows laptops) is < $300. Shadowjams (talk) 04:55, 3 June 2009 (UTC)


 * As with a used computer, be sure to test out such a netbook on your favorite sites before buying it. "Sluggish" is the term everyone seems to use when describing them.  Tempshill (talk) 18:41, 3 June 2009 (UTC)


 * In rural parts of the USA, residents often bring trash to a town transfer station. There is often a place to leave potentially reusable items that anyone can take for free. I've seen computers in such areas (but never saw one I wanted). --Jc3s5h (talk) 00:26, 4 June 2009 (UTC)

Well I am looking for something along the lines of 2+ gb of ram. I would like 2.5+ ghz processor. Basically I want something that I can play Assassin's Creed or Crisis with no delay or lag or anything. A perfect computer. This is just a cheap test computer that I would like to build to see if I can actually put one together and make it work. I might put this one downstairs for my little brother to play and leave that older XP for my sister to have or something. On the other hand I might build this one and sell it but I havent decided yet. I would like new parts for this just so I can become familiar. RandomAccessDawg (talk) 03:12, 4 June 2009 (UTC)


 * Check out tigerdirect and get a barebones kit. Normally you'll need to get a CPU fan, and an optical drive, but sometimes not even those, and they're sub-500, and quite nice.  The reality is that you won't really be able to purchase anything from any store online for $100, as parts worth that little aren't worth the cost to sell online.  Craigslist is your best bet, but I imagine cheap parts don't exist there, but cheap computers will. Chris M. (talk) 13:33, 5 June 2009 (UTC)

What is Kerberos used for?
Hi. I understand that Kerberos is an authentication protocol. But where have I seen it in action? I searched Google and even read your entry on the subject, but haven't found an answer. Your entry says that it is the default authentication protocol for Windows 2000 and later, but it is a network protocol. I log on to all kinds of servers through FTP, HTML forms, RDP, SMB, etc. Are those all using Kerberos to authenticate me?--76.120.121.128 (talk) 06:04, 3 June 2009 (UTC)


 * At the risk of over-simplifying with an analogy, Kerberos is like a government agency that issues an ID card such as a driver's license that is generally trusted by others as reliable proof of identity. When you applied for your ID card, you likely needed to provide a primary evidence of identity, such as a birth certificate.  (Well, lets assume you did.)  The ID card is then used as a reliable secondary evidence of your identify for the purposes of bank transactions and other business activity.  I may not know you, and I may not even have the expertise to verify the authenticity of your birth certificate, but I have full faith in the reliability of that government ID.  The point is that you do not need to carry your birth certificate with you everywhere, and I only need to be able to recognize and accept one reliable form of identification.


 * The birth certificate would be analogous to your username and password that you use to log into the network. The government agency would be the Authentication Server.  The ID card would be the credentials (Ticket-Granting Ticket) issued by Kerberos that presumably cannot be forged.  Banks and other businesses would be the other computers on the network that provide protected services such as file servers, database servers, and internet proxies.  These other computers are not going to trust your computer directly to verify your identify, and may not even have the ability to verify your password if you supplied it.  Instead, they trust the credentials issued by the authenticating server, and then only under well defined conditions.  The point is, you enter your password once, and the entire network recognizes who you are.


 * And, by the way, all the sensitive Kerberos communications are performed using encryption methods that are designed to ensure the integrity of your credentials and to prevent non-authorized persons from hijacking your identity (i.e. forging or copying your ID card). -- Tcncv (talk) 07:18, 3 June 2009 (UTC)


 * ok. Thanks for that. But, where have I used it? I have used browsed things like network printers and network shares inside Windows. I noticed that I am asked for a password, but am logged in automatically on subsequent visits. Is that done using Kerberos?--76.120.121.128 (talk) 08:18, 3 June 2009 (UTC)


 * Not necessarily. Are you using Active Directory?  --  JSBillings  11:29, 3 June 2009 (UTC)


 * A lot of incarnations of OpenAFS use Kerberos authentication. I use Kerberos to mount my university-wide network drives with single-login authentication.  I can also use it to check my email from the command prompt with Alpine - again, I don't have to reauthenticate to the web server.  I feel that this is slightly more secure than saving passwords in a web browser's password-manager, because the Kerberos authentication never actually transmits the password over the network (unlike a browser's password manager).  Nimur (talk) 14:44, 3 June 2009 (UTC)


 * What Jsbillings is asking is, are you using a Windows 2000 (or later) operating system? Those use Active Directory which has Kerberos-based authentication. Jay (talk) 03:41, 4 June 2009 (UTC)


 * ok. Thanks guys. Yes. We are using Windows Server 2003 with active directory. I use Windows XP at home.--76.120.121.128 (talk) 06:30, 4 June 2009 (UTC)

Junk Mail Sent by Me
Over the past year or so, my email inbox has received several spam emails sent "by me." I realize that the send address is spoofed, but... Is my email address being used as a 'from' address for your spam? 96.227.82.128 (talk) 09:07, 3 June 2009 (UTC) I really want to get these fuckers!

The from section can be easily forged and chances are the program the spammer is using is just set to make the from section the same as the to one. Your email address probably isn't being used by spammers. edit: If it was you might be getting lots of failed delivery messages or something, from the spammers sending mail to places that don't exist or outright refuse it.Gunrun (talk) 09:13, 3 June 2009 (UTC)

See Email "..Also note that the 'From' field does not have to be the real sender of the e-mail message. It is very easy to fake the 'From' field and let a message seem to be from any mail address.." You can assume that the spam email recieved from other senders has the 'from' faked too. Why email/internet service providers allow this to continue is a mystery to me. Put me down on your lynch mob - they definately deserve need a hard smack.77.86.10.194 (talk) 11:36, 3 June 2009 (UTC)


 * If your email address is used to send spam to others, you will get tons of "could not send" errors. I know this because I get hundreds of them each day (the problem of having a publicly available email address since 1994).  As far as stopping it, it is not possible.  Until someone figures out how to legally run an international bounty system to catch the spammers dead or alive, there is no punishment for being a spammer. --  k a i n a w &trade; 15:31, 3 June 2009 (UTC)

Z80 question
In Zilog_Z80 it says that "..except for the mnemonic “jp (hl)”, which is inconsistent.."

I've tagged it with 'clarify' in the text - is this true or just creative vandalism? Thanks77.86.10.194 (talk) 09:36, 3 June 2009 (UTC)


 * There is an inconsistency in the asembluy language notation for some jumps. In most Z80 assembly language statements, operands not contained in parentheses represent immediate values contained in the instruction (if a number) or to the contents of a register (if a register name).  Parentheses around a value or register name changes the meaning so the operand is the memory addressed by that value or by the contents of the specified register.  All of the load, arithmetic, and other data manipulation instructions follow this convention.  For example:

LD BC,2130H ; Load the value 2130H into register BC    LD SP,HL     ; Load the value from register HL into register SP    LD A,(2130H) ; Load the contents of memory at address 2130H into register A    LD A,(HL)    ; Load the contents of memory at the address contained in register HL into register A


 * Jump instructions are kind of in a class by themselves in that they refer to a memory location, but do not actually access that memory location (the next instruction after a completed jump will be retrieved from that location). Jump instructions are essentially special case load instructions whose target is the program counter.  Consider the following jump instructions.

JP 2130H    ; Jump to location 2130H (load the value 2130H into the program counter) JP $-28H    ; Jump relative to location $-28H (subtract the value 28H from the program counter) JP HL       ; *** Invalid *** (Intended to load the value from register HL into the program counter) JP (HL)     ; Jump to the location contained in register HL (Accepted, but inconsistent syntax)


 * If you think of a jump as simply a register load, the first two instructions above follows the same convention as other instructions. The third instruction would be the natural continuation of this convention, yet this form is not valid.  The fourth instruction is the accepted syntax, even though it would appear to say "load the contents of memory at the address contained in register HL into the program counter.  -- Tcncv (talk) 01:49, 4 June 2009 (UTC)
 * Thanks, So there never was any 'dereferencing' or 'indirection' instruction where hl acts as a pointer.
 * Is there any explanation for this? (I mean the inconsistent syntax?)77.86.10.194 (talk) 09:34, 4 June 2009 (UTC)

Funny floats?
I'm experimenting with a program that works with numbers - and I was trying to find out what sort of system it uses to store the numbers. I found that it could store 2^n+1 without losing the "1" up to about n=53, so I guess 53bits for mantissa. I also found that it could store up to 2^1023 as a number and that it also handles higher values as "infinity" - so I thought 64bit floats right??

Then I tried 2^-n - it gives answers for up to 2^-1074 (7 is not a typo) it evaluates
 * 2^-1073 as 9.88131E-324
 * 2^-1074 as 4.94066E-324 (corrected twice)
 * 2^-1075 as 0
 * 2^-1024 as 5.56268E-309
 * 2^-1023 as 1.1125E-308
 * 2^1023 as 9.98847E+307
 * 2^1024 as "infinity"

the actual values seem correct? What is going on here? Any ideas?77.86.10.194 (talk) 11:12, 3 June 2009 (UTC)


 * This looks like double precision with denormal numbers near 0. PrimeHunter (talk) 12:34, 3 June 2009 (UTC)
 * The smallest denormal number should be 2^-52 × 2^-1022 = 2^-1074 so it matches your data. PrimeHunter (talk) 12:51, 3 June 2009 (UTC)
 * Of course. thanks.

Grand Celebration
i am interested in buying the ship Grand Celebration from the owner when it is to be scraped —Preceding unsigned comment added by 12myoung (talk • contribs) 12:48, 3 June 2009 (UTC)


 * You are at Wikipedia, an online free encyclopedia with millions of articles. One of them is Grand Celebration but Wikipedia has no relationship to the ship or its owner. PrimeHunter (talk) 13:09, 3 June 2009 (UTC)
 * This is also the COMPUTER help desk.77.86.10.194 (talk) 14:22, 3 June 2009 (UTC)

PRC Firewall Circumvention Tools
I will be taking a vacation to the People's Republic of China this summer and would like to be able to fully access YouTube, Wikipedia and Facebook while over there. It is my understanding that these sites have certain restrictions and certain pages are blocked?

What would be a good, free proxy that will allow me to circumvent the "Great Firewall of China?" Would Psiphon be a good one?

Acceptable (talk) 19:54, 3 June 2009 (UTC)


 * Internet censorship in the People's Republic of China has a lot of information about what is blocked and how, and about tools that are often used for circumvention. Psiphon is specifically mentioned, though the article isn't a ratings guide, so it doesn't state how effective it is.  By the way, if I were vacationing in China, I would be careful to obey the laws of China.  Tempshill (talk) 20:22, 3 June 2009 (UTC)


 * This constitutes original research; but last summer, a few friends of mine and I experimented with the so-called "censorship" and also with secure encrypted remote access. My colleague(s), visiting China, tested internet access at public (internet-cafe) and private (university and private residence) internet terminals, located in the People's Republic of China.  We also managed to set up SSL tunnels to my server located in California as a further proof-of-concept.  There was no interference with the SSL tunnel, which was also used for proxying web browsers and email.  Also worth noting is that there was no noticeable disruption to "normal web browsing activities" even without the proxy.  My conclusion is that if any filtering is taking place, it is largely ineffective - we could not identify any websites which we could not access.  And, if any were hypothetically found, we had an uninterrupted, high-grade, secure, private tunnel to an outside proxy.  All you need is an SSH server running anywhere else in the world.  I'm reluctant to call the numerous cited sources of internet censorship "false" - but I think that they maybe overstate the impact of any such censorship.  From a strictly technical sense, it is very very difficult to implement the sorts of things listed in the technical section - I suspect that packet-level keyword sniffing is "outright impossible" on a national level.  I also think that the List of websites blocked in the People's Republic of China is grossly inaccurate (again, though, this is original research and so I am reluctant to modify that article - but we accessed all of those websites even without a proxy).  My real suspicion is that these widespread rumors of censorship are the result of a huge, technically-incompetent user-base in China who are new to internet technology are willing to readily blame every sluggish server, every slow connection, every 404-error on a (largely nonexistent) government censorship program.  It's also worth mentioning that even progressive nations such as the United States and the United Kingdom also censor certain content (such as child pornography), and have used such arguments to justify their technical capability to block access to other websites.  This notably resulted in nation-wide blocks of Wikipedia in the UK only a few months ago.  Nimur (talk) 04:23, 4 June 2009 (UTC)
 * You're going to have to cite a source when you claim that the US blocks access to any websites. They have certainly arrested people for child porn on the web, but you're really stretching the word "censor" if this is what you mean.  Tempshill (talk) 04:29, 4 June 2009 (UTC)
 * If you read carefully, I said that the U.S. censors such content, not that they block access to its websites with a firewall. Finding an example of a take-down is easy, I just went to the FBI's web page: the FBI.  "On July 25, 2006 the website www.namgla.net was shut down by the SAFE Team" (of the Los Angeles branch of the FBI).  This seems like a perfectly appropriate thing to do (I'm not a fan of child abuse or child exploitation).  However, in the strictest sense of the word, this is censorship, in that the FBI decided that this content was not "free speech" and shut the web site down.  The technical means used were a bit more realistic - they actually had to shut the server off.  (I don't believe that DNS redirects or packet-level filtering would be effective on a nation-wide scale, because I have a vague idea about how the internet works).  If the server is alive and is connected to a network, it'd be really hard to block it with a "firewall" of any type.  Nimur (talk) 04:36, 4 June 2009 (UTC)
 * You talked about the US using child porn censorship arguments to justify a technical capability to block access to websites. I don't think the US has demonstrated any ability to block access to any websites.  They just drove over to the guy's house, arrested him, and turned his computer off.  Tempshill (talk) 16:42, 4 June 2009 (UTC)
 * (Even the Wikipedia block, mentioned above, was grossly ineffective - a DNS bypass and URL-sniffer apparently was used, but it failed to block most of the accesses, and was totally unable to block accesses that used Wikimedia's secure server). Nimur (talk) 04:43, 4 June 2009 (UTC)
 * Before suggesting that "From a strictly technical sense, it is very very difficult to implement the sorts of things listed in the technical section", you should probably consider this tidbit: "The telecommunications industry in China is dominated by three state-run businesses: China Telecom,China Unicom and China Mobile." It is not only possible, but probable that the state has an interest in limiting many facets of the internet, especially the global collaboration venues like Wikipedia. --66.195.232.121 (talk) 17:14, 4 June 2009 (UTC)
 * I don't know what China does but I would imagine packet level keyword checks would not normally be used to stop communication but only to alert a program to capture the whole conversation for a security officer to inspect. Dmcq (talk) 07:51, 4 June 2009 (UTC)
 * According to Blocking_of_Wikipedia_in_mainland_China, Wikipedia have been unblocked since the Olympics. But seriously, unless you are doing some really bad anti-government stuff it's rather unlikely that you'll get anything like the scare stories. --antilivedT 10:53, 4 June 2009 (UTC)
 * And if you want circumvention methods, try a Google translation of this. --antilivedT 11:03, 4 June 2009 (UTC)

Simple line matching in linux / bash
I am trying to compare two lists and display only the matching lines. Ideally I could match with regular expressions, but being limited to literals is acceptable. I have done this with a bash script, but that method is cumbersome and not very efficient.

It seems like there is some obvious program out there that I don't know about. Is there a native utility that does this in most linux/unix distributions, and if so what is it? Also, if I am stuck doing this in a script, what would the most efficient way to do that be? Right now I'm taking each line from file 1 and searching for it in file 2, returning it to the stdout. Is there an improvement over this? Shadowjams (talk) 22:40, 3 June 2009 (UTC)


 * sort, comm, and cut.&mdash;eric 00:10, 4 June 2009 (UTC)


 * Perfect. Thank you. I don't know how I didn't find comm before. Shadowjams (talk) 00:30, 4 June 2009 (UTC)