Wikipedia:Reference desk/Archives/Computing/2010 April 28

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.

April 28

Program updates

Why is it that not everyone gets a particular update at the same time? Surely once it's released it's on the Internet for everyone to get. Chevymontecarlo. 05:32, 28 April 2010 (UTC)

I notice that, too. I have two computers, both Dell Inspirons (6000 and 6400) running Windows XP SP2 and SP3 respectively. SP3 always updates itself sooner (usually a day in advance) than SP2. Rarely does it happen at the same time. Of course, once the updates are released, everyone can go to the official website and get it, but not eveyone's computers that are set for automatic updates receive them at the same time. 24.189.90.68 (talk) 08:07, 28 April 2010 (UTC)

There are five cases where computers do not update:

1. Many people don't connect to the Internet long enough to download updates, or their connection is so slow (and unreliable) that the download cannot finish.
2. People who pirate their copies of Windows generally turn off automatic updates in order to avoid downloading Windows Genuine Advantage.
3. Many viruses turn off automatic updates in order to avoid downloading the Windows Malicious Software Removal Tool.
4. Many people (including me) prefer to download updates manually for greater control. Updates can create bugs in certain cases. So, if you administer a large number of computers for an organization, you often test the updates on one computer, and then allow the computers under your control to download some (but not all) of the new updates.
5. Other Windows installations are just screwed up in general. For example, the Automatic Update service may not start (or it may start but crash immediately).--Best Dog Ever (talk) 08:36, 28 April 2010 (UTC)

Umm... this answer is completely irrelevant. The question is about computers being notified of updates at different times, not about computers not updating. --169.232.246.169 (talk) 08:41, 28 April 2010 (UTC)

No. He said, "Why is it that not everyone gets a particular update at the same time?" Where did he even write the word notified? You are putting words in his mouth.--Best Dog Ever (talk) 08:48, 28 April 2010 (UTC)

Still, it is completely irrelevant. His computers do get updates. --169.232.246.169 (talk) 09:10, 28 April 2010 (UTC)

I don't know how it is with a Windows machine but I know that with my Macs, they check once a week or some such thing. So if the two systems of the first repsondent are like mine, then their check-for-updates days might be just a bit off from one another. Additionally, unless the update is critical, there may not be much of a point of getting it out to everyone RIGHT NOW! The software companies may want to keep their own servers from being flooded with too many downloads. Dismas|^(talk) 09:31, 28 April 2010 (UTC)

Yeah, only during critical updates do my computers update at the same time, and even then it doesn't always happen. A lot of the time, I still have to download it myself on the slower computer. Windows users typically receive their updates on the second Tuesday of every month. However, Best Dog Ever's point about manually downloading them to make sure the update isn't affected with a troublesome bug reminds me of the whole debacle with McAfee that just happened. Perhaps I should disable Automatic Updates and wait a few days after the release of a new update to see if everyone will bitch about some problem with it. 24.189.90.68 (talk) 10:12, 28 April 2010 (UTC)

Thanks guys. By the way, I am using a Mac :) Chevymontecarlo. 12:12, 28 April 2010 (UTC)

I've no reference for this, but I seem to remember toward the beginning of the time that Windows Update was installed as a service on Windows machines, to poll Microsoft's servers to see if there was anything new, Microsoft intentionally set up the Windows Update service so that the computers would check for updates over the space of about a week, to try to spread out the bandwidth over a week and avoid a single gigantic spike. Comet Tuttle (talk) 20:38, 29 April 2010 (UTC)

I've read before [1] that updates my be made available to certain geographical regions only (I presume by IP address but don't really know). Microsoft could have additional ways of managing updates, e.g. perhaps each installation has a random number from 0-15 or something and non essential updates are rolled out to a random number every day or whatever. Nil Einne (talk) 02:57, 4 May 2010 (UTC)

Adaptive Huffman coding with variable symbol length

Are there any variants on adaptive Huffman coding that allow multiple symbols to be turned into one at runtime, so that repeated substrings can be compressed without having to transmit their length (which has already been observed) and even if they occurred far earlier in the file (unlike with the "sliding window" of LZW or Gzip)? NeonMerlin 10:50, 28 April 2010 (UTC)

Maybe you want dynamic Markov compression. 69.228.170.24 (talk) 20:59, 28 April 2010 (UTC)

+99

I was just wondering. Which country has +99 area code? Simply south (talk) 11:37, 28 April 2010 (UTC)

According to List of country calling codes neither +99 nor +999 are assigned. --antilived^{T | C | G} 11:46, 28 April 2010 (UTC)

And for a very good reason. In the UK, the way calls were routed from outlying areas to the main exchange (central office in US-speak) was by dialling an initial 9. So, to call the emergency services (999), the intial 9 just routed the call to the central exchange, and the 2 final 9s got you the emergency operator. If you lived in a location served by the central exchange, calling 99 was enough to get the emergency operator. I can guarantee this from personal experience. So if 99 was a country code, people in the UK would forever have been bothering the emergency operator. --Phil Holmes (talk) 14:50, 28 April 2010 (UTC)

That's not true. Because they must dial the international call prefix (in the case of the UK, 00) before dialing the country code. --169.232.246.212 (talk) 19:30, 30 April 2010 (UTC)

Well it's possible some people may forget to dial the international call prefix, however if that was a concern other numbers would be unassigned but +91 is assigned to India for example and they do have STD codes beginning with 1 like New Delhi. Besides that I would say it's inaccurate to say +99 isn't assigned. +99 isn't one country code but is instead split up into 3 digit codes so while some of those (+990, +997, +999) aren't assigned, others are. Nil Einne (talk) 02:46, 4 May 2010 (UTC)

The weakest hyperlink...

I've defined the link, hyperlink, as:

[[Hyperlink#Anchor|hyperlink]].

But, when I roll the mouse over it, it just says "Hyperlink", not "Hyperlink#Anchor". How can I fix this, under Firefox 2.0.0.20 on Windows 98 ? StuRat (talk) 14:23, 28 April 2010 (UTC)

I don't think you can fix it. It is how Mediawiki functions, not how your browser functions. Mediawiki takes that link and gives the following HTML from it:

<a href="/wiki/Hyperlink#Anchor" title="Hyperlink">hyperlink</a>

Which as you can see defines the "title" (which the mouse over text) to be "Hyperlink." It strips out the hash tag automatically (and probably purposefully). --Mr.98 (talk) 14:27, 28 April 2010 (UTC)

That sucks. I can see the full web address "http://en.wikipedia.org/wiki/Hyperlink#Anchor" down in the Status Bar, I guess I'll need to squint down there, then. StuRat (talk) 14:31, 28 April 2010 (UTC)

You could probably rig a Greasemonkey script to automatically rewrite Wikipedia link "title" fields for you to include the hash. But I don't know how to do that. --Mr.98 (talk) 14:34, 28 April 2010 (UTC)

Advice in meta:Help:Link gives us this: hyperlink -- Finlay McWalter • Talk 14:48, 28 April 2010 (UTC)

I use WP:Popups, and when I mouseover the link, the "headline" is "Hyperlink#Anchor" and the preview begins at the "Anchor" section. -- Coneslayer (talk) 16:48, 28 April 2010 (UTC)

You must have some different setting in your custom css or js file, under My Preferences, then, because that's not the behavior I get. That's just what I'm looking for. StuRat (talk) 15:39, 29 April 2010 (UTC)

User:Coneslayer/monobook.js appears to be the only non-redlink customization file on my account. -- Coneslayer (talk) 19:43, 30 April 2010 (UTC)

I copied it over to mine, but see no difference. It's a mystery. What's your browser and O/S ? Perhaps that does make a diff after all. StuRat (talk) 12:39, 1 May 2010 (UTC)

Firefox 3.6 on Windows XP and Windows 7. Everything I've done is by copying-and-pasting magic incantations, so I'm sorry I can't be of much help. -- Coneslayer (talk) 13:43, 3 May 2010 (UTC)

It's worth noteing that the mouseover text is set server side. If you feel WP should behave otherwise it's probably best to take the discussion to Wikipedia:Village_pump_(technical). I don't know how much changes would be required to change the default behavior in this case. Taemyr (talk) 09:11, 29 April 2010 (UTC)

Mozilla Firefox unable to use an external XSL stylesheet?

I have run into an interesting problem at work. There is an XML file located on one server and an XML stylesheet transforming it to HTML on another server. The XML file has an embedded link to the XSL stylesheet. Mozilla Firefox refuses to display the file, giving the error message: "Error loading stylesheet: An unknown error has occurred (805303f4)". But if both files are loaded on the same server, and the link is changed to point to the local directory, everything works fine. Is there something in Mozilla Firefox preventing XML files from using external XSL stylesheets? JIP | Talk 14:49, 28 April 2010 (UTC)

A quick Google search suggests (could someone confirm this?) that this is a deliberate security feature of Firefox. Some kind of XSS-type attack they're trying to prevent? « Aaron Rotenberg « Talk « 17:26, 28 April 2010 (UTC)

If that is true, then is there some kind of option in Mozilla Firefox to disable it? There's nothing fundamentally preventing the source XML document and its XSL transformation stylesheet from coming from different places, so the problem here must then be that Mozilla Firefox explicitly refuses to handle such a situation. JIP | Talk 20:00, 28 April 2010 (UTC)

This (rather ancient) Mozilla document says "Mozilla won't load XSLT stylesheets from a different domain for security reasons." You may be able to get around this with access-control, which it seems FF3.5+ supports for XMLHTTPrequest. So I think (but haven't tried) that you can load the XSLT that way and then apply it with transformNode -- Finlay McWalter • Talk 21:13, 28 April 2010 (UTC)

Oops, I forgot to link, the thing about FF3.5 supporting access-control is here. -- Finlay McWalter • Talk 21:22, 28 April 2010 (UTC)

I'd be surprised if this isn't either a file permissions problem (look in error.log or equivalent) or the XSLT being sent with the wrong content-type, which should (I think) be application/xhtml+xml or text/xml. Viewing what actually transpires between browser and server can often be enlightening in this circumstances - use livehttpheaders. -- Finlay McWalter • Talk 20:36, 28 April 2010 (UTC)

No, I can do a wget on the server hosting the XSL file all OK, and it gives the content type as text/xml. As the XSL file itself is publically accessible (although direct access to the server isn't), I can do the same tests both at work (in a Windows-only environment) and at home (in a Linux-only environment). The exact same situation occurs at both places. JIP | Talk 20:47, 28 April 2010 (UTC)

Having a different from address in an email

Hi, Want to send emails from an address but have it show another address in the From section.

I bought a domain that allows me to create email addresses on it; emails sent to those addresses will be forwarded to any addresses of my choice for free. But I want my responses to appear from the same address emails are sent to. AOL, my usual email provider doesn't seem to do it. Hotmail says the hotmail address "on behalf of" my domain's address (seems silly!). GMail does it I think, but I don't use Google services - especially not emial - due to them keeping, reading, and anylising everything. Outlook does it too i think, but I don't know how, and I want a website (such as Hotmail) not software. Can you help? Thanks 86.177.7.170 (talk) 15:36, 28 April 2010 (UTC)

In order for webmail to work, the webmail provider's servers need to keep and read all your email (otherwise, they wouldn't be able to display the content of your mail to you. Gmail has the nice feature of letting you turn on HTTPS so that it's impossible to in-between servers to read your mail; I don't know if the other major providers do that.). The only unusual thing about Gmail is that the ads are automatically generated based on keywords in the mail messages. You might find them creepy (I just find the ads amusing), but it doesn't pose any danger to your privacy. Any webmail provider has the ability to let a human read your mail, and (I assume) all of them promise not to do this, unless required to by law.

If you're interested in email privacy in webmail, there's Hushmail, but read the "controversy" section: unless you use desktop mail software (which defeats the point of webmail) and encrypt all your email (which requires effort from everyone you email), the provider can (and probably will) obey a court order to open your mail. Paul (Stansifer) 16:29, 28 April 2010 (UTC)

Oh, GMail does seem quite qood with the HTTPS setting. Google are good like that, they have great features that other providers don't use for some reason. BUT they do have software that reads all your information, even if it is just for advartising. Other providers merely transfer the information and keep it until a little while after you choose to delete it. Google keeps everyhing indefinately at great cost to them - this type of action is awfully suspect. I will never use Google. Ever. Thanks though.

And that Hushmail seems good too. I wouldn't use software because, as you say, that defeats the point of webmail. And i'm not bothered about court orders or anything; I'm not using it for anything bad, just regular mail. And it hides IP addresses in the header which rocks; practically what AOL does too. But, most importantly, does Hushmail let me change the From section in an email? Thanks. 86.177.7.170 (talk) 17:42, 28 April 2010 (UTC)

GMail will let you do that, the latest I could find on hotmail was that only if the domain is hosted with Microsoft, is that supported[2]Cander0000 (talk) 03:50, 3 May 2010 (UTC)

Apple iWeb problem

When I'm using Apple's iWeb application, I get this issue where the screen goes blank when I scroll down the page. If I'm creating a page that is longer than the screen and I attempt to scroll down at all, the whole page just goes blank, until I scroll back up. I've updated to the most recent version that my OS will allow (v.2.0.3) but it still happens. Anyone else seen this?91.109.225.48 (talk) 21:24, 28 April 2010 (UTC)

Hi, I've said this before to someone else with a problem with their Apple stuff. I would advise you, if no one gives you a good answer here, to go check out the Apple Discussion forums - [3] and post your question there. You need an account though, however you just need an Apple ID, which you will already have if you use iTunes. On there are loads of other Apple users who will be happy to help you. I'm on there too! You can search for 'R94N'...! Chevymontecarlo. 12:15, 29 April 2010 (UTC)