Jump to content

Wikipedia:Reference desk/Archives/Computing/2010 September 9

From Wikipedia, the free encyclopedia
Computing desk
< September 8 << Aug | September | Oct >> September 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.


September 9

[edit]

ram

[edit]

Does stuff in ram degrade over time? —Preceding unsigned comment added by Taiier4 (talkcontribs) 13:30, 9 September 2010 (UTC)[reply]

All electronic devices degrade over time. RAM is stable enough for much more than the average lifespan of a computer (about 5 years). Abusing RAM by putting it under heat stress or static shock will cause failure before normal degradation. -- kainaw 14:08, 9 September 2010 (UTC)[reply]
I assumed from the wording that the OP meant programs being stored in RAM Quadrupedaldiprotodont (talk) 14:16, 9 September 2010 (UTC)[reply]
The little 1's and 0's do not degrade if the computer is on. They do degrade if the computer is turned off. It is rare to have a computer with RAM that retains data when the computer is turned off. That doesn't mean that there is no data corruption. It is not the job of RAM to protect against corruption due to programming errors, virus activity, etc... -- kainaw 14:22, 9 September 2010 (UTC)[reply]
The 1's and 0's can be actually corrupted spontaneously, which is why ECC RAM exists.—Emil J. 14:31, 9 September 2010 (UTC)[reply]
DRAM, which is the "normal" kind of RAM, stores data in tiny capacitors (since capacitors can be made smaller and simpler than flip-flops, I believe). The capacitors are constantly leaking, so the RAM is constantly reading and re-writing itself. However, none of this is visible to the programmer or the user, so DRAM is effectively totally stable as long as the power is on (and assuming that cosmic rays don't flip any bits).
Note that a 1-bit error would be sufficient to crash a program, so you would notice if your computer's memory weren't stable. Paul (Stansifer) 15:10, 9 September 2010 (UTC)[reply]
Surely Stanisfer meant to write: A 1-bit error could be sufficient to crash a program. Comet Tuttle (talk) 16:42, 9 September 2010 (UTC)[reply]
... No, "sufficient to" doesn't mean "necessarily will". A one-bit error is not necessary to crash a program, but it is certainly sufficient if it occurs at a critical location. (We seem to have a different interpretation of the word "would" in this context.) Dbfirs 14:06, 10 September 2010 (UTC)[reply]
It's quibbling now, but I think it's misleading to tell a layman that a 1-bit error "would be sufficient to crash a program" with no further explanation. Comet Tuttle (talk) 19:58, 13 September 2010 (UTC)[reply]

mac address

[edit]

I have a problem with my internet connection and I need some help. My friend informed me that I need to change the "mac address" of my computer to fix the problem, however I have a Windows PC. W4576 (talk) 13:50, 9 September 2010 (UTC)[reply]

See MAC address. Though how it will help, I don't know. Your MAC is, AFAIK, normally hardcoded in your NIC! --Tagishsimon (talk) 13:53, 9 September 2010 (UTC)[reply]
See MAC spoofing.—Emil J. 14:02, 9 September 2010 (UTC)[reply]
A MAC address is not the same as a Macintosh or 'Mac' computer, hence the capitalisation. Chevymontecarlo - alt 14:38, 9 September 2010 (UTC)[reply]
We really need more information about your problem. What exactly is wrong with your computer? Rojomoke (talk) 16:02, 9 September 2010 (UTC)[reply]
It sounds like the OP's friend has suggested MAC spoofing. But whether this will actually fix the problem depends on what the problem is. Can you elaborate? MAC spoofing will only solve a very limited, specific type of network miscongifuration - i.e., when the DHCP server is set to manual-mode IP assignments indexed by MAC address. It is pretty rare that ordinary users encounter this condition; and you will need to know a valid new address that you can change to. Furthermore, not all network cards or drivers permit changing the MAC address (for many, this address is "burned" into the card). Nimur (talk) 16:10, 9 September 2010 (UTC)[reply]
What might be happening is that the OP has changed internet service provider, and kept the same cable modem, whereupon it becomes necessary to phone up the ISP and "ask for a new MAC address". I think that's how it goes, anyway. I'm very vague about this myself, but this is the kind of situation in which "changing your MAC address" is legitimate and sensible (and possible). (Or is it that you keep the same MAC, and arrange for it to be handed from one ISP to the other?) 81.131.4.100 (talk) 17:29, 9 September 2010 (UTC)[reply]
No. It is extraordinarily unlikely that the ISP would request you to do anything with your MAC address. I think you are probably confusing IP address with MAC address. Nimur (talk) 18:19, 9 September 2010 (UTC)[reply]
Another suggestion: perhaps the OP is asking about the Migration Authorisation Code which is required (here in the UK, and maybe other countries as well) so you can change your internet service provider. With some exceptions, the MAC number effectively orders BT to change the connection to your new ISP at the telephone exchange. Astronaut (talk) 20:36, 9 September 2010 (UTC)[reply]
That must be what I was vaguely remembering, above (no, I wasn't mixing it up with IP addresses). Does it happen with cable ISPs, too? I don't remember ever having DSL. Evidently I'm very bad at remembering things, though, so maybe I did once. 213.122.53.167 (talk) 15:39, 10 September 2010 (UTC)[reply]
Change the section header to distinguish this question from others with the same/similar header. Astronaut (talk) 00:18, 11 September 2010 (UTC)[reply]

Saved pages

[edit]

A few years ago I could save a page, go back to the saved copy and click "submit" or whatever, and it would connect to the server it was saved from and work as though it was a live page. Now, when I try to click "submit" Firefox does nothing. It doesn't load, give a error, or anything. What changed? Has firefox been updated at some point to remove this behavior? Photoplants (talk) 14:02, 9 September 2010 (UTC)[reply]

This could be a change with FireFox, but I think it has to with what web page you were saving. When you submit a form on a HTML page, the markup contains information that tells the web browser where to send the submitted form. For example, as I'm writing this response this is the HTML code for the form I'm submitting:
<form id="editform" name="editform" method="post" action="/w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit" enctype="multipart/form-data">
The action element in the markup tells the browser when I hit "Save Page" it will send the information to /w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit. This URL references the root of the site it's currently on. It only makes sense in the context of the website that I'm currently on; because I'm on the site en.wikipedia.org, the browser knows that it is submitting the form to http://en.wikipedia.org/w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit. When you save a page then open it up on your machine, the context is gone; the browser doesn't understand that /w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit refers to a page on en.wikipedia.org since the page has been opened up from a file on your computer and not the en.wikipedia.org domain.
However, depending on the website's HTML code, this action element could contain the whole URL, including the domain. For example if the code above looked like:
<form id="editform" name="editform" method="post" action="http://en.wikipedia.org/w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit" enctype="multipart/form-data">
The browser would know to submit the form to http://en.wikipedia.org/w/index.php?title=Wikipedia:Reference_desk/Computing&action=submit regardless if it's on en.wikipedia.org, a HTML file on your computer, or any other location. --—Mitaphane Contribs | Talk 17:29, 9 September 2010 (UTC)[reply]

Alarm

[edit]

I've been trying to make a little in-browser alarm clock with greasemoney. So far I have the following:

var time=String((new Date()).getTime()).replace(/\D/gi,'')
if (time > 1284055200)
window.alert("It's 6pm")

This pops up a little box when you refresh a page after 6pm today. However, if you don't refresh a page nothing happens. I want it to work without any user interaction. Is that possible? 82.44.55.25 (talk) 18:00, 9 September 2010 (UTC)[reply]

I thought this question was answered earlier this week by Kainaw et al., and again here. What are you still having trouble with? Nimur (talk) 18:28, 9 September 2010 (UTC)[reply]
I thought my question here was pretty clear, but I'll try to clarify. Those questions asked how to make javascript wait / execute at a certain time. I figured that out (see above code). This question is asking how to make the script execute without loading a page. Currently it only works when loading / refreshing a page. 82.44.55.25 (talk) 18:54, 9 September 2010 (UTC)[reply]
It appears that you are asking how to make it go off every day at 6pm without reloading the page. Is that correct?
If so, two options:
The obvious one: have the alarm automatically reload the page when the user clears the alert box.
The less obvious one: You should have a timeout function that waits until 6pm to call the alarm function. Have the alarm function do the same timeout so it will show an alert and then timeout until 6pm. -- kainaw 19:05, 9 September 2010 (UTC)[reply]
Yes. I don't understand the first suggestion, the alert box only shows up when the current time is greater than 1284055200 (which needs to be changed to whatever future time I want to alarm to go off at, since 1284055200 is now past). Using a timeout thing isn't possible, because whenever the tab is refreshed / reloaded, greasemonkey reloads the javascript and the timer resets. This is why I ended up doing a "if greater than current time" thing, because the other way wasn't practical. 82.44.55.25 (talk) 19:24, 9 September 2010 (UTC)[reply]
Untested, has drift problems, but how about:
var MS_IN_A_DAY = 24 * 60 * 60 * 1000;
function advance_by_a_day(d) { d.setTime(d.getTime() + MS_IN_A_DAY) }
function now() { return (new Date()).getTime(); }

var next_deadline = new Date();
next_deadline.setHours(6);
next_deadline.setMinutes(0);
next_deadline.setSeconds(0);
if (next_deadline.getTime() < now())
    advance_by_a_day(next_deadline);
setTimeout("alarm()", next_deadline.getTime() - now());

function alarm() {
    advance_by_a_day(next_deadline);
    setTimeout("alarm()", next_deadline.getTime() - now());
    alert("wake up");
}
--Sean 22:47, 9 September 2010 (UTC)[reply]

Vintage Optical Computer/Mainframe

[edit]
Resolved

Hello Wikipedia,

I am looking for the name of a vintage computer or mainframe from the 1960's to 1980's era. It was unique because of its massive storage capacities for the time, in the realm of gigabyte or maybe even terabyte storage. It used an optical storage method similar to microfiche with a robotic arm that stored the imagages in what looks like a card catalog. The machine was used up until recently, in the 1980s or 1990s when the lead technician retired or passed away.

Thanks for your help. aszymanik speak! 20:23, 9 September 2010 (UTC)[reply]

Probably, you mean the IBM 1360. You can see more information at the Computer History Museum's website, and this YouTube video. Here's a nice inside-view photo of IBM 1369, which was used at Lawrence Livermore National Laboratory. Here's an original operator's manual and much more in-depth discussion, hosted by an enthusiast website operated by a bunch of old Large-Scale Computing junkies from LLNL - notably, George Michael, who passed away in 2008. Nimur (talk) 21:16, 9 September 2010 (UTC)[reply]
Thank you. That was exactly what I was looking for. aszymanik speak! 01:26, 10 September 2010 (UTC)[reply]

RAID block size

[edit]

I'm setting up a RAID for my wife's photography files. She uses raw format on a Nikon D700, so the files are large. If I set the RAID block size to the maximum (256K) will that have any consequences when I back up the RAID to another disk? Dismas|(talk) 23:53, 9 September 2010 (UTC)[reply]

No. The RAID block size has no effect on anything except read/write speed. (It's unlikely to have a noticeable effect on that either—I'd just use the default size.) -- BenRG (talk) 00:15, 10 September 2010 (UTC)[reply]
The RAID volume's block size has no effect on replication of the data to any other data storage medium. However, unless all of the files exceed 256KB in size and/or are slightly less than a multiple of 256KB in size, it is may be more advisable to use a smaller block size, in order to avoid wasting large amounts of diskspace. AFAIK, the degradation in read/write speed incurred from using a smaller block size is usually acceptable, in order to avoid excessive diskspace wastage, but the choice between optimizing for diskspace usage and optimizing for throughput ultimately depend on your particular applications. Rocketshiporion 02:15, 11 September 2010 (UTC)[reply]