Wikipedia:Reference desk/Archives/Computing/2008 June 11

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.

June 11

Ubuntu programming

What programming tools does Ubuntu have?--71.185.140.19 (talk) 01:46, 11 June 2008 (UTC)

There's tons... See here. --Wirbelwindヴィルヴェルヴィント (talk) 03:02, 11 June 2008 (UTC)

Yeah, then click on "Development". If you want any more details, you'll have to be a little more specific. Are you interested in a particular language? Do you want an integrated development environment, or are you just interested in finding out what gui toolkits or libraries are available? Maybe just a rich text editor? Indeterminate (talk) 04:49, 11 June 2008 (UTC)

Opera Master Password

I have set a Opera Master Password and have forgotten it. How can I get rid of it? It doesn't matter if I also get rid of all my other passwords. GoingOnTracks (talk) 03:50, 11 June 2008 (UTC)

Delete your wand.dat and opcert6.dat files from your profile directory. You should probably do it with opera closed, then re-open it afterwards. See here: [1] Indeterminate (talk) 05:01, 11 June 2008 (UTC)

Program

I remember seeing an old computer program called "Goldie". It has an animation of a naked lady dancing. But I can't find any mention of it anywhere on the internet. Anyone have a link? Interactive Fiction Expert/Talk to me 06:33, 11 June 2008 (UTC)

Internet porn... :p HardDisk (talk) 17:58, 11 June 2008 (UTC)

Keyboard protector

Yesterday I purchased a new Toshiba laptop and I was flipping through a catalog of accessories and noticed that they offered a "keyboard protector" which resembles a plastic transparent sheet. While I am interested in it, do you recommend I buy it? Is it worth the money doing so? I don't know anyone with a laptop having one. Also, do I remove the plastic strip over the integrated camera? I think I should. --Blue387 (talk) 07:44, 11 June 2008 (UTC)

It depends on whether you plan on spilling your tea on the laptop's keyboard ;-). Seriously, if your environment is dirty or dusty or you do have a proclivity for spilling your drink, the protector may help. But it does impede keyboard action somewhat. And after several years accumulation of crumbs (which really do accumulate, even for the most fastidious user), most keyboards can be vacuumed clean with good results.

Meanwhile, regarding the plastic strip: If it is simply a piece of almost-clear vinyl or the like, surrounds the lens and the surrounding plastic/metal bezel, and it has an obvious pull-off feature, then yes, remove it. Such strips are used to prevent scratches/crate burn during shipment and in the case of your camera, will definitely impede the functioning of the camera if left on.

Atlant (talk) 11:36, 11 June 2008 (UTC)

Ipod Touch/WinSCP

When I delete applications off my ipod touch through WinSCP it doesn't seem to actually delete the application because I can still see it on the spring board. Is there something I need to do to delete the application in another way? Possibly delete it from another location? And if you guys are wondering why I am using WinSCP to delete applications it is because my installer keeps crashing on me. Yer so I'm wondering can you fix that through WinSCP also? Thanks. 220.233.83.26 (talk) 07:54, 11 June 2008 (UTC)

Dynamic IP for WebSite

Hi, I am thinking about making my own website, and I have the option to choose a static IP, my question is: Would it be less safe to use a static IP than a dynamic IP? Thanks in advance. SF007 (talk) 13:27, 11 June 2008 (UTC)

• You would want to use a static IP. A dynamic IP changes periodically. Which means that every so often the URL will change. A static IP is fixed, so it doesn't change. As far as security goes, it isn't going to make a big difference, except in the sense of Security through obscurity - if people can't find your website (because the IP keeps changing) they won't be able to attack it - but if people can't find your website, you might want to wonder why you have one anyway. :) Plus the value of security through obscurity is debatable, (it certainly shouldn't be seen as sufficient on its own, even under the best circumstances). - Bilby (talk) 14:39, 11 June 2008 (UTC)

• • URLs can use domain names, and domain names can change IP (e.g. DynDNS), so URLs can be constant with dynamic IP. MTM (talk) 18:53, 11 June 2008 (UTC)

• Technically that is possible. However, domain names need to be registered through a DNS, and it takes time for changes propagate - I normally quote about 24 hours for a change in the IP address to start working for all users. So while you could do it, there would be a period during each changeover during which you would get inconsistency, with some people going to the correct IP, and some going to the old one. This would also loose most of the advantages gained via security through obscurity, as the domain name would be consistent for attacks, and could be used to resolve the IP if needed. - Bilby (talk) 22:42, 11 June 2008 (UTC)

• The only reason why I use a dynamic IP is that it is cheaper in Poland. Since my IP changes once per several weeks or months, I did not think about security implications of this. MTM (talk) 21:19, 12 June 2008 (UTC)

yeah, I think a static IP will do fine, better to focus on trying to keep the software up to date and stuff like that. Thanks. SF007 (talk) 23:44, 11 June 2008 (UTC)

Most advanced file compression?

What's the most advanced file compression method available to an end user? My preliminary searches point to 7z, but I just wanted to check in with you experts. --70.167.58.6 (talk) 14:27, 11 June 2008 (UTC)

Define "advanced". Fastest? Most compression? Both? --LarryMac | Talk 15:16, 11 June 2008 (UTC)

maximumcompression.com has benchmarks of a lot of different lossless compression programs. PAQ8O10 and WinRK win most if not all of the compression ratio tests. PAQ is free (GPL), but it's a research compressor with no GUI or fancy features or support, and it's very very slow. WinRK is commercial and has a GUI etc., and I think it's faster, but I've never used it. Unless you have unusual needs you should go with 7-Zip (or WinRar). -- BenRG (talk) 17:38, 11 June 2008 (UTC)

Thanks! Yeah, I should have specified I need cross-platform and no CLI! :) --70.167.58.6 (talk) 22:24, 11 June 2008 (UTC)

Maybe one of the PAQ derivatives or front-ends such as PeaZip might fit your needs, then? —Ilmari Karonen (talk) 11:55, 14 June 2008 (UTC)

By cross-platform, I mean Mac and Windows. Thanks for your suggestion, though! --70.167.58.6 (talk) 13:54, 16 June 2008 (UTC)

Most advanced video codec?

It seems that h.264 seems to be the most advanced, most widely available codec (that also isn't under the control of one company). What is next? Is there a more efficient h.264 successor in the works? --70.167.58.6 (talk) 14:30, 11 June 2008 (UTC)

• VC-1 - better known as WMV3/WMV9
• H.265 - not really started yet

--tcsetattr (talk / contribs) 18:23, 11 June 2008 (UTC)

And eventually maybe Dirac. -- Q Chris (talk) 10:13, 12 June 2008 (UTC)

I don't think VC-1 is intended to be a successor to H.264. It's more of a contempory (and in fact performance worse in most visual comparison tests) with the advantage of being easier to decode. Dirac may be a next generation codec. Or it may simply compete with H.264 (i.e. MPEG-4 AVC). One of the key advantages is it's (believed to be) patent free. Theora the current primary patent free codec is not particularly good. It doesn't even compete with MPEG-4 ASP currently. One of the key problems I've heard is that the developers have been severely restricted in what they can do given how heavily patented the field is. It's obviously going to get better, but I personally have my doubts it will achieve quality/bitrate comparable to H.264 for a long while. Dirac may have an advantage there as wavelet video compression is still relatively new. Nil Einne (talk) 18:31, 17 June 2008 (UTC)

language support for internet explorer of windows mobile

Can anybody please tell me how to enable indian language support for internet explorer of windows mobile ?59.93.196.42 (talk) 17:00, 11 June 2008 (UTC)

iphone unlocking

Now that iphone comes with a 2 year contract, will iphone unlocking continue? I think such two year contract was there last year also. Then how did many American citizens unlock and sell phones in other countries when there was a contract? was there a contract last year?

On the previous version of the iphone you could buy it without a contract, and purchase a contract later. The new pricing and availability strategy will reduce the motivation, as the initial cost for the phone will be lower, but you will be locked into paying, and also the phone will be available in many countries. Nevertheless you may want to change the SIM when traveling to avoid roaming charges. For example in Hong Kong local mobile phone charges are only 1% of the cost when using a foreign SIM. Graeme Bartlett (talk) 21:35, 11 June 2008 (UTC)

I read in some blog that previously it had a no commitment price and now it does not. Any idea what is that?59.92.107.231 (talk) 03:46, 12 June 2008 (UTC)

fragmentation

Can fragmentation of a hard drive be good in that it ensures not the same part of the drive is always used and its use is spread out randomly across the drive? JessHalie (talk) 19:14, 11 June 2008 (UTC)

I doubt it. If anything, the increased need for seeks will accelerate degradation. -- Meni Rosenfeld (talk) 20:15, 11 June 2008 (UTC)

Fragmentation really has no benefits. It uses more disk space than a nonfragmented disk. It takes longer to read/write files. Both of those put the drive to more use than it would in a nonfragmented system. As for using the same spot of the disk over and over - there is no harm there. It doesn't degrade the disk (unless the drive is broken). It merely alters the magnetization of the block. I know of no risk for disk damage through magnetization. However, I should point out that the popular flash-media USB sticks are not the same. They do degrade and have a limited number of read/write operations before the fail. At that point, they truly fail and are unusable. Therefore, they have a sort of built-in fragmentation to spread the work around the internals of the disk in an attempt to get the maximum amount of use out of the whole drive. -- kainaw™ 00:01, 12 June 2008 (UTC)

Fragmented files do not use more disk space. A defragmented file has all of its parts next to each other in order on the disk. A fragmented file has one or more parts spread out on the disk. Either way, the file will use the same amount of space on the disk. See file system fragmentation. --Bavi H (talk) 02:04, 12 June 2008 (UTC)

So, how does the OS know it hit the end of the file or the end of a fragment and the rest of the file is elsewhere? It has to store the information about the fragment on the disk. That takes space that wouldn't be used if the file was not fragmented. -- kainaw™ 12:41, 12 June 2008 (UTC)

I'm not sure to what degree it varies across filesystems, but in the old FAT filesystem, the actual FAT was a table of cluster numbers, with as many entries as there were clusters. If j were stored in slot i, that meant that i was followed by j in a (hopefully the) file that contained both. There was a special cluster number that meant EOF — perhaps the cluster number for the FAT itself, since nothing could be continued there. (The OS knew where within the cluster to stop because it knew the total length of the file.) So the space really was constant; an unfragmented file just had entries that looked like (@4b00d) 4b00e, 4b00f, 4b010, …. In general, I suspect that this pattern is common; if you can either store a fragment pointer or something else in a given space (thus saving space when you don't need the pointer), you have to have some way of "escaping" real disk data that happens to look like your pointers (however you store them). The added complexity and wasted space (for, say, a bit that indicates whether a pointer or data follows) may not be justified. --Tardis (talk) 14:37, 12 June 2008 (UTC)

I vaguely recall from days gone by it being a linked list or such. back then you could use norton edit or similar to just go and look at the FAT. boy i'm oldGzuckier (talk) 20:57, 12 June 2008 (UTC)

Kainaw: The operating system knows how to find all the clusters of a file by using the directory entry to find the first cluster, and the file allocation table to find all the rest. The file allocation table of a particular disk is always a fixed size (it has an entry for every cluster whether it's used or available), and is really a special reserved part of the file system, not a part of any particular file. I explain a little more here, because I couldn't find any brief explainations online. --Bavi H (talk) 03:00, 13 June 2008 (UTC)

A related discussion on file sizes, clusters, sizes on disk, and fragmentation was here. I got so frustrated with the incorrect explainations I wrote quite a bit to try to clarify. --Bavi H (talk) 04:38, 13 June 2008 (UTC)

Isn't it rather misleading to use a defunct filesystem that isn't the default filesystem on any modern operating systems as an example? How about using NTFS instead? I may very well be wrong, but my understanding is that it stores everything in metafiles, which are files in the root directory beginning with $. As fragmentation grows, information about the fragmentation is added to the metafiles - making them larger. -- kainaw™ 11:19, 13 June 2008 (UTC)

but as for the original question, for quite a few versions of windows now the OS has been smart enough to keep moving forward through the unused disk space while writing rather than just rewriting the beginning of the disk over and over as you suggest. although i don't know if it's specifically to reduce wear, it was presented to me as a way to reduce fragmentation, since obviously when it finally comes around to writing at the beginning of the disk again, there will be more continuous open space after more time has passed and more files have been deleted, and less need to cram little pieces of file into little holes. Gzuckier (talk) 21:01, 12 June 2008 (UTC)

Making a symmetric matrix

I have a matrix A and a symmetric matrix B, and I'd like to calculate $A^{\top }BA$, which is obviously symmetric. Using such tools as GSL and/or BLAS, how can I go about doing this efficiently? Neither of the intermediate products $A^{\top }B$ and $BA$ will in general be symmetric, so the normal routines that read or write (slightly more than) half of a symmetric matrix won't work. The general routines will, of course, wastefully calculate every entry in the final result. If it matters, A is typically not square. --Tardis (talk) 19:42, 11 June 2008 (UTC)

Go through unique pairs of columns of A (e.g. do a loop to find i<j), and do the inner product defined by B? In the end, though, I'm not sure whether this much optimisation is worth it.--Fangz (talk) 00:05, 12 June 2008 (UTC)

You can check the matrix chain multiplication first to see if you should do (A^TB)A or A^T(BA). However, it seems to me that many of the operations you perform will be highly repetitious. So, memoization may be use to shortcut the operations and return the answer. -- kainaw™ 00:18, 12 June 2008 (UTC)

The (direct) multiplication has the same expense in either order because of the symmetry. Perhaps the best thing is to simply do the first multiplication normally; then do the final multiplication "by hand" (without calling into a library), skipping the terms above/below the diagonal (as Fangz suggested). I think the "correct" answer for large matrices (when it's worthwhile to recognize the symmetries) is to factor $B=Q^{\top }DQ$ (D will have all positive entries if B is positive-definite, which it is for me) and write $A^{\top }BA=A^{\top }Q^{\top }DQA=C^{\top }C,\;C={\sqrt {D}}QA$. The standard packages know how to deal with $C^{\top }C$, so efficiency is had; the required Schur decomposition is $O(n^{3})$, so it's not necessarily more expensive than the matrix multiplications. --Tardis (talk) 15:34, 12 June 2008 (UTC)

In case anyone's implementing this, the general Schur decomposition is really overkill; the precise function I'll use (if I need it) is gsl_eigen_symmv. Thanks for the input. --Tardis (talk) 15:44, 12 June 2008 (UTC)

Posting a file without exposing myself to spambots

I need to temporarily files for other message board users -- what are some good sites? By good, I mean easy to use, no porn advertising, clean interface. --70.167.58.6 (talk) 22:16, 11 June 2008 (UTC)

Temporarily what? I am ahving difficulty understanding your question. Astronaut (talk) 23:04, 11 June 2008 (UTC)

Temporarily file? You mean upload a file for temporary use? Try Microsoft SkyDrive. Just create an account, upload the files and set them to public or semi-public. GoingOnTracks (talk) 23:15, 11 June 2008 (UTC)

You want a website to place a file? you can use MediaFire, I think it is a good site because:

• No registration required
• upload files up to 100 MB
• Files are not deleted
• you can registrate and organize files in folders
• the person that download the file don't have to wait time, or enter any CAPTCHAs, (unlike rapidshare and it's kittens....)

link: www.mediafire.com SF007 (talk) 23:50, 11 June 2008 (UTC)