Computing desk
< December 3	<< Nov \| December \| Jan >>	December 5 >

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.

December 4

Eight Nehalems in 2U: Electrical & Thermal Considerations

I have been wondering; is it electrically and thermally possible to put eight Intel Xeon 7500 Series processors and a terabyte of RAM into a single 2U server? I mean a single system server, whether it has eight sockets on a singe baseboard or it consists of two quad-socket boards glued together. Amax's ServMax X1401 has half of that (four Nehalems and 512GB of RAM) in half the space, and the Cray CX1000-SC achieves eight Intel Xeon 7500 Series processors and a terabyte of RAM in 3U of rackspace but lacks redundant PSUs and multiple PCIe slots... Rocketshiporion ♫ 09:17, 4 December 2010 (UTC)[reply]

Gulftown Xeon and Desktop Motherboard

If I installed a Gulftown Xeon processor on a (LGA 1366, X58) desktop motherboard, can I also install non-ECC, unbuffered ("regular") memory modules on the motherboard? The motherboard is known to support the processor.

More generally, given a motherboard and a supported processor, how do I determine what kind of memory to install? I am not asking about memory type, clock, size, and the number and placement of modules, but the ECC-, buffered-, and registered-ness of the memory.

Thanks for the answer. 118.96.155.129 (talk) 01:01, 4 December 2010 (UTC)[reply]

AFAIK, RDIMMs and UDIMMs cannot be mixed in the same system. As for how to determine what kind of memory to install, supported memory types and configurations will be stated in the processor's technical specifications. And the technical datasheets for most Intel processors are available from the Intel website. You may find the technical documentation for the Xeon 3000 Series and the Xeon 5000 Series useful. Rocketshiporion ♫ 09:30, 4 December 2010 (UTC)[reply]

Embedded objects in HTML and changing their size

Is there a way in HTML to embed an object, like a flash movie, but instead of a height and width value in the embed tag, have them both be variables and use some sort of input (on the page or otherwise) to change these variables "live" on the page (without needing to refresh)?

137.81.118.126 (talk) 01:27, 4 December 2010 (UTC)[reply]

They should allow percentage values, such as "width:50%". Then, it will have a width 50% of the container it is in. The advantage is that it will be resized to the page size. If you grow/shrink the window, it will change size. The problem is that this all depends on the ability of the object to resize. I seriously doubt flash is capable of resizing without completely reloading the page. However, you can resize images, tables, divs, etc... easily. I've even set up text to get larger/smaller as the window changes size. -- kainaw ™ 01:38, 4 December 2010 (UTC)[reply]

(edit conflict) Yes, this is easy with Javascript. For example:

Extended content

<html>
<body>
<!-- below: an image on the Wikipedia front page, sized to its default height and width. Note that we have assigned it a unique id so we can reference it easily. -->
<img id="obj" src="http://upload.wikimedia.org/wikipedia/en/thumb/b/b0/William_Wilson_Talcott.png/84px-William_Wilson_Talcott.png" width="84" height="100"> 
<p>
<!--the width form entry; 84 is the default. We have assigned it a unique id so we can easily get its value. -->
width: <input id="obj_width" type="text" value="84"><br> 
<!-- the height form entry; 100 is the default. Ditto on the id. -->
height: <input id="obj_height" type="text" value="100"><br> 
<!-- this is the "make it happen" button. It assigns the width of the element with the id of "obj" using the value in the "obj_width" form element, and then does it to the height as well. "return false" just means, "and don't reload the page". -->
<input type="submit" value="change" onclick="document.getElementById('obj').width=document.getElementById('obj_width').value; document.getElementById('obj').height=document.getElementById('obj_height').value; return false;">
</p>
</body>
</html>

In this case, the element with the id "obj" is an image tag, but it could be an object tag as well. Now whether this will produce the right results with the object will depend on how it handles having its canvas resized. I don't know how well, say, Flash would deal with that. My feeling is that you will need some kind of Actionscript in the Flash file itself to dynamically accommodate its changing size if you don't want it to just crudely re-size it the way that an image is. The way you'd do this is have a function within the Flash file that would recalculate its own canvas depending on variable input, and then trigger that with Javascript. It's considerably more complicated than what I've shown above. --Mr.98 (talk) 01:40, 4 December 2010 (UTC)[reply]

You can load flash movies directly into a window without any container, and then resize the window (while playing the movie), and they're perfectly happy to adapt to the new size. So I think it would work. 81.131.39.230 (talk) 02:13, 4 December 2010 (UTC)[reply]

(edit) I'll play around with this. The flash video in question DOES allow different sizes and seems to work with various sizes, i just dont know how it would respond to being told to change its size on the fly as opposed to editing the tag and reloading the page. :o 137.81.118.126 (talk) 01:44, 4 December 2010 (UTC)[reply]

It'll probably depend on the Actionscript that is used in the video in question. Alternatively you could probably reset the object's embed src on the fly with Javascript as well, which might cause just the Flash element to reload (and not the whole page). But I haven't tried that and it might not work. --Mr.98 (talk) 01:47, 4 December 2010 (UTC)[reply]

I tried it with a YouTube embed tag, using the "object" as "obj" above, and it worked 100% fine. The interface is clever enough to resize on the fly without any difficulty. So it can be done. I don't know how much cleverness is behind that, though. --Mr.98 (talk) 01:50, 4 December 2010 (UTC)[reply]

I tried this with my video file as well, and also got the desired result, thanks! 137.81.118.126 (talk) 01:54, 4 December 2010 (UTC)[reply]

Loss of Net Neutrality in US vs Internet Censorship in China

Let us suppose a hypothetical (yet possible) scenario where the United States passes a bill that destroys net neutrality and allows ISP's to selective throttle internet traffic or implement multi-tiered service. While it is likely that certain sites unable to pay a premium will be slower than before, suppose that the ISP really dislikes a specific site, say WikiLeaks for example, are they able to completely block the site or slow the traffic to a point where it takes like an hour just to load one page?

If so, how does this compare to the "Great Firewall" internet censorship in China? In China, you are still able to access blocked sites using proxies such as Tor, but can you do the same if ISPs in America decide to censor certain websites? Are there any technical solutions around this in the US?

Acceptable (talk) 02:23, 4 December 2010 (UTC)[reply]

The difference: In China, the blocking of a specific site is mandated by the government, while in the United States, it is mandated by the market. 118.96.155.129 (talk) 02:45, 4 December 2010 (UTC)[reply]

Where "Market" means that one single company with no competitors that I could get broadband from. We call that Freedom here. APL (talk) 04:01, 4 December 2010 (UTC)[reply]

I am looking more towards the technical implications of both scenarios. What is the relative difficulty between the two systems in accessing restricted websites? Acceptable (talk) 04:24, 4 December 2010 (UTC)[reply]

It's pretty clear that there has been an unreasonable push by the current administration to influence private companies, both in the Wikileaks case and the early International Burn a Koran Day suppression. Even so, the ability of private Internet broadband companies to act as censors must be limited. Techniques like using HTTP tunneling and other tunneling protocols, done in China, might be legally punished as subversive; but in the U.S. they would seem to be legitimate business practices. The cable company would have to make a very clear policy of forbidding, with some kind of fine, any kind of content that they couldn't intercept and successfully interpret - but besides extraordinary customer outrage, there's the little point that they don't have the employees to implement something like that. They're making their profit providing content, not rummaging through it to punish customers. Besides, they'd face the prospect that angry customers would help get new competitors off the ground, or engage in other forms of terrorism. Think of where a cable company would be if one angry person spent an 8-hour night driving around with a pair of long pruning shears.

So I think you'll see any such censorship being imposed essentially at gunpoint, only when companies are intensively targeted by a coordinated campaign of legislative and administrative threats, high-profile boycotts, and denial of service attacks. I'm not saying they won't be pushed into it, but I don't think they'll be willing partners without a carrot-and-stick the size of CALEA - in other words, formal government censorship. Wnt (talk) 16:41, 4 December 2010 (UTC)[reply]

Is wikileaks down?

Both [1] and [2] seem to be down for me. [3] and [4] says "it's just me." What's wrong here? —Preceding unsigned comment added by 212.169.185.150 (talk) 17:57, 4 December 2010 (UTC)[reply]

The .ch address was predicted [5] to go the same way as the .org address, since both are administered by EveryDNS.net. I suppose that downforeveryoneorjustme has slow reactions (something to do with Domain propagation?) and that both domains (but not the IP address) are in fact down by now. 81.131.27.188 (talk) 18:41, 4 December 2010 (UTC)[reply]

Thee are a number of URLs in the WikiLeaks article; at least some are active. ---— Gadget850 (Ed) ^talk 18:51, 4 December 2010 (UTC)[reply]

Wikileaks supposedly has new DNS providers [6] [7]. However changes to the name servers can take up to 48 hours or so to propogate depending on the caching settings. On the other hand a whois on the wikileaks.org domain at [8] still shows their name servers as everydns. Nil Einne (talk) 19:08, 4 December 2010 (UTC)[reply]

Should not any reference here to downforeveryoneorjustme be deleted as just more spam?--Aspro (talk) 19:52, 4 December 2010 (UTC)[reply]

Huh? 82.44.55.25 (talk) 20:56, 4 December 2010 (UTC)[reply]

We need to define "down." At present, (20:19, 4 December 2010 (UTC)), the Wikileaks server is active, the domain name has been de-activated, and the PayPal account has been de-activated. Nimur (talk) 20:19, 4 December 2010 (UTC)[reply]

Wikileaks being down could be for a variety of reasons, DDOS attacks, DNS support dropped, its blocked in your country, etc.Smallman12q (talk) 20:23, 4 December 2010 (UTC)[reply]

Note that as I was told above, http://searchdns.netcraft.com/ provides retroactive name service. With governments returning to using DNS as a censorship gateway, this type of service is now a basic net essential. Wnt (talk) 21:57, 4 December 2010 (UTC)[reply]

Wikileaks servers are still up at http://213.251.145.96/ according to https://twitter.com/#!/wikileaks/status/10621245489938433 --h2g2bob (talk) 00:41, 5 December 2010 (UTC)[reply]

As I noted above, AFAIK wikileaks domain names were never deactived (recently). Their free DNS provider stopped providing service because they kept getting DDOSed to such an extent it was affecting their DNS provider (and therefore other customers), as most free providers would. They still have/d full control of their domain name but need/ed to find some other DNS provider. (Evidentally they were given 24 hours before the service was pulled, the info above suggests they found one but so it's not clear to me why they still seem to be pointing to everyDNS with their wikileaks.org domain name. One would have to wonder who was in charge of the domain name if they thought it a good idea to use a free service without (at least I haven't seen any info to suggest they did) at least making sure they would be okay with you getting repeatedly DDOSed when the AUP said "If we find that any user is utilizing a certain amount of resources which we deem to be excessive we will contact that user and allow him or her to either begin paying us or we will recommend a premium DNS provider such as our friends at Dyn Inc", which may explain why the domain name is still pointing to everyDNS as its name servers. IIRC for some reason some of their domain names were reported as belonged to either wikia or the WMF for a while, given the common confusion and the fact wikileaks is controversial I don't think Jimbo in whatever role was particularly happy with it but it took a while to get someone in wikileaks to fix it which perhaps somewhat supports the theory.) See also [9] which mentions their domain name registrar saying wikileaks still have control although they are facing pressure from both sides. Nil Einne (talk) 16:51, 6 December 2010 (UTC)[reply]

java object reference.

Hi, I have a quick question.

Suppose I have some nice code like so:

public void foobar(thing a)

{

thing b = a // this is normal, of course.

}

But what if foobar is a method within class thing... can I say

thing c = this

Basically, I need a method within a class where two object references point one to the object passed and the other to the object that called the method, or vice versa according to a random() coin toss.

Thanks!

Duomillia (talk) 23:02, 4 December 2010 (UTC)[reply]

I don't really understand what you want, but perhaps you want to be using the this keyword? It's always valid, where foobar is a nonstatic method. Beyond that, it's difficult to see what problem the code you propose is trying to solve, so it's difficult to give you a constructive answer about what code you should write to achieve your goal. 84.51.189.0 (talk) 23:47, 4 December 2010 (UTC)[reply]

Hmm, let me see if I can rephrase the problem.

There is a class called nCraft. It has the member data nCraft.health and nCraft.fire and a method nCraft.attack(nCraft victim) So for example, I could say alice.attack(bob) and in the method, based on a random() outcome winner = bob and loser = alice or vice versa and health is changed accordingly.

Duomillia (talk) 00:25, 5 December 2010 (UTC)[reply]

Maybe something like this?

public void foobar(Thing t) {
	Thing winner, loser;
	if(Math.random()) {
		winner = this;
		loser = t;
	} else {
		winner = t;
		loser = this;
	}
}

Also, if other methods depended on who won and who lost, it'd be better to make winner and loser be instance variables, or better yet, have the method return a Thing[]. And, something neat I learned just yesterday, (the super long question about jdialogs) are some code tags. Use <syntaxhighlight lang="java"> and </syntaxhighlight> around your code. KyuubiSeal (talk) 04:32, 5 December 2010 (UTC)[reply]

Thanks! That's exactly what I needed! All I need it for is something like this:

winner.gold += looser.gold;
looser.gold = 0;

But I wasn't sure if I could use 'this' all by it self as a reference to self. Well, it looks like I can. Thanks! Duomillia (talk) 04:49, 5 December 2010 (UTC)[reply]

One further off-topic comment, because this isn't the Language Desk: if anyone besides you is ever going to read this code, they'll think that YOU are the "looser" if you don't learn the difference between LOSE and LOOSE. DaHorsesMouth (talk) 21:13, 5 December 2010 (UTC)[reply]