Wikipedia:Reference desk/Archives/Computing/2010 November 8

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.

November 8

EXIF, meaning of "Optical Zoom Code"

Does anyone know how to interpret the line

Optical Zoom Code               : 7

in the EXIF data of a photo? The camera is a Canon PowerShot A590 IS. I'm assuming it means 4X optical zoom (the maximum) but I'd like to find a resource for interpreting this stuff more generally. Thanks, --Trovatore (talk) 01:36, 8 November 2010 (UTC)

Considering its widespread use, EXIF is surprisingly nonstandardized. The "official EXIF specification" is not really official - it is defined by a small organization that has loose affiliations with several Japanese camera manufacturers called CIPA (Camera & Imaging Products Association). Our EXIF article links to the EXIF DCF specification version 2.3 - and as you can see, "optical zoom code" is not a standard tag. It is an "internal use only" tag for the manufacturer, by the manufacturer - and any conformity to any spec should be considered "use at your own risk."

Optical Zoom Code usually directly maps to a zoom-state; your camera might use codes (0-7) or (0-127) or so on to denote all possible steps between minimum and maximum zoom. The exact mapping of zoom-code to focal-length will vary from manufacturer to manufacturer (and model to model, even firmware to firmware). You can use a tool like gphoto2 to look up the mappings between zoom-codes and actual focal-lengths for your particular camera (if you trust their tables).

If you really need to know the true optical zoom, the EXIF tag you should use is FocalLengthIn35mmFilm. You can compare this to your camera's FocalLength tag to determine the optical zoom as a "1x" or "2x" number. Nimur (talk) 04:09, 8 November 2010 (UTC)

Thanks much. Can't seem to find that tag. But there's FocalLength and LongFocal and ShortFocal, and FocalLength equals LongFocal which is four times ShortFocal, so I take that to mean the picture was shot at 4x optical zoom. --Trovatore (talk) 04:58, 8 November 2010 (UTC)

Mac OS X version adoption rates

Where could I find a breakdown percentage of Mac users who are running Mac OS X 10.6, 10.5, 10.4, even Mac OS 9? --70.167.58.6 (talk) 08:38, 8 November 2010 (UTC)

They did a poll over at MacOSXHints a few months back, if I remember correctly, which showed that 10.6 adoption rate was over 70%. 10.5 was something like 18% or 19%, 10.4 less than 5%, and a smattering of people with lower systems. Unfortunately I can't find the link. Of course, that's a bit of a geek site which might queer the results a bit. I suspect upgrade rates are lower for people who use their machines for business purposes (why risk the possibility of having to rewrite your business webpage because of software changes?). --Ludwigs2 09:41, 8 November 2010 (UTC)

ah, found it: [1] as of last january. 68% 10.6, 21% 10.5, 8% 10.4, and the remainder of the X systems accounting for roughly 2%. 21,000 votes, though, so it's a decent sample even for a convenience sample. The poll didn't go into os 9 users, thought the previous poll on system version, which was held back in 2006, showed only 0.69% of users using os9 and only 0.21% using os8 or earlier. --Ludwigs2 09:52, 8 November 2010 (UTC)

It may be a 'decent sample' but clearly not a random one, not even close (to state the obvious, people visiting a site like MacOSXHints are likely to be fairly technically minded). I would trust a decent scientific poll with only 2000 people more then this one with 20000 votes. Nil Einne (talk) 12:30, 8 November 2010 (UTC)

oh, no question. but I've never seen any other type of site do such a poll, so I don't know how we'd get around that problem. I mean, I suspect that Apple has that data somewhere in its nefarious databanks (just from phone-homes on installs and updates), but if so they do not seem to be inclined to share. --Ludwigs2 22:10, 8 November 2010 (UTC)

Here is an article from February 2010 which says that 10.5 was actually still the most common version at that point, but 10.6 adoption was rapidly increasing. The numbers look far more realistic than the ones given above, which look like early adopter rates. --Mr.98 (talk) 00:32, 9 November 2010 (UTC)

HELP. How do I remove specific auto-suggestions from my Google Chrome's URL bar?

Anytime I start typing any web address with "d," guess what is the first site Google Chrome suggests???

It's DailyDiapers.com!

I feel haunted by that reappearance; even though I only go there on Incognito mode now, I still get reminders that I used to surf it on normal mode.

A BIG problem would be if a friend borrowed my laptop and decided to type up a website that started with "d." You can imagine what could happen next!

Now how do I remove that offending auto-suggest from ever showing up again?

(I had to register a new name just to ask embarrassing questions; I wasn't even going to post this from my IP because that could trace back to me as well.) --EmbarrassedWikipedian (talk) 09:58, 8 November 2010 (UTC)

Step by step instructions from http://www.google.com/support/forum/p/Chrome/thread?tid=6501481bab1c67eb&hl=en

Turning off Auto-Suggestions (also see Reference 1) 1. Clear your browsing history 2. Click the Tools menu 3. Select Options 4. Click the Under the Hood tab and find the Privacy section 5. Deselect the 'Use a suggestion service to help complete searches and URLs typed in the address bar' checkbox. 6. Click Close.

General Rommel (talk) 10:16, 8 November 2010 (UTC) :Extra note :Under my dev versoin of Chrome, it appears as 'Use a prediction service' not suggestion service General Rommel (talk) 10:18, 8 November 2010 (UTC)

Long-Term Data Archival

Hi Everyone,

   I am seeking to store large amounts (~ 2TB) of data for a very long (minimum 50 25 years) period of time. The media has to be rewritable, and the data will be very infrequently (annually or semi-annually) rewritten. The data must be preserved for atleast 50 25 years. Which of the following types of storage media will (or is atleast most likely to) preserve data for this length of time?

1. Ultrium LTO-5 750GB Rewritable Tape Cartridge
2. Removable Serial ATA Hard Disk Drive
3. Ultra Density Optical 30GB Rewritable Disk
4. Multi-Level Cell Solid State Drive

   Thanks, everyone. Rocketshiporion♫ 12:36, 8 November 2010 (UTC)

One of the big problems you'll have either way is whether you'll be able to edit any of those mediums in 50 years. 50 years is a loonnngg time in terms of information technology — consider that this is what a computer looked like even less than 50 years ago. (Computers from 1960 did not even use integrated circuits!) Will SATA be around in 2060? I wouldn't bet on it. SATA itself dates from only 2000 or so, as far as I can tell. Even parallel ports only go back 40 years.

Rather than trying to find one system that will last for 50 years, might I recommend that a rotating system be used? Pick a system that will last for 10 years. After 10 years, upgrade it to whatever the equivalent is at that time. After another 10 years, repeat. You've both reduced your necessary lifespan for the system significantly, and also basically guaranteed that it'll be compatible with whatever fancy new computers there are in 2060. 10 years is an acceptable tech jump — people have all sorts of common methods of playing 10-year-old software, or using 10-year-old peripherals — whereas 50 is a bit much. And if you forget to upgrade it after a point... well, at least you're 10 years more up to date than you would have been otherwise.

A brief analogy. The Bible did not survive to modern times because people made one very good copy of it and kept it very safe. (There are a couple verrry old copies, but the fact that they have survived is basically attributable to luck.) It survived because people were constantly re-copying it, on modern paper, with modern technology.

All of the above is making the assumption that you'd have said drives in a place accessible to upgrades. If you're putting it on a rocket ship, well, I suppose that would introduce additional variables. --Mr.98 (talk) 14:00, 8 November 2010 (UTC)

As the first Parallel ATA HDDs appeared in 1986, and there are still motherboards being sold with PATA connectors, I am quite confident that SATA (which AFAIK appeared in 2004) will be around atleast as long - until 2028. In addition, most of the SSDs of which I'm aware use SATA, and those that don't use FC. Magneto-Optical disks (which have been superseded by UDO disks) were introduced in 1985, and they are still in use today, 25 years later. Although I agree that 50 years is a very looooong time for data storage media, and so I shall amend my question - see above. Rocketshiporion♫ 19:38, 8 November 2010 (UTC)

I would just also note that the tech difference between 2060 and 2010 is probably a lot greater than 1985 and 2010, thanks to Moore's Law and its analogues. Whether such a thing would impact peripherals and media, one can only speculate. But I would not be surprised if it did (especially since the bandwidth rates on many modern ports leave quite a lot to be desired for, even with present requirements, much less future ones). --Mr.98 (talk) 23:27, 8 November 2010 (UTC)

I'm pretty sure your AFAIK is wrong. I bought a SATA motherboard in 2003. I may have also purchased a SATA HDD then although I'm less sure of that. Our article claims a creation date of 2003 but confusingly describes how many sold between 2001 and 2008. I think 2001 is more likely to be accurate as I find people discussing SATA stuff as if it existed on usenet in 2002 [2]. It's possible it was not finalised until 2003 so the 2002 and 2001? stuff were based on the draft standard. Nil Einne (talk) 00:28, 9 November 2010 (UTC)

I would point you at RAID 6 which is a way of using 4 or more hard drives in a collection such that data integrity can fully survive the simultaneous failure of 2 drives. Rather than assuming that any method will work flawlessly for 25 years, it would generally be better to develop fault-tolerant systems. Whenever you update the data, run data integrity checks and replace any drives that generate errors. As long as the same data block never fails on three drives simultaneously, the data will be safe. For added safety you can remove the drives to different remote storage locations or create additional off-site copies to guard against localized events (such as a fire) destroying multiple drives at once. Dragons flight (talk) 01:32, 9 November 2010 (UTC)

I think the conventional wisdom is that RAID is not meant to insure data integrity, but uptime. Off-site copies aren't just for protection from fires and power problems and theft and hardware failure, they also protect you against accidental erasure, which is an even bigger threat to data than the delicacy of hard drives. No one's ever more than, like, twenty keystrokes away from erasing their hard drive. Whatever damage a user does to their files, a RAID will happily duplicate it across multiple drives. So don't bother trying to make data more secure in one location; back it up. Paul (Stansifer) 04:55, 9 November 2010 (UTC)

To be clear, I mean RAID in the archival system, so that the archives efficiently store multiple redundant copies of the data. But presumably one updates the archive only infrequently and it would not be an online system such that one could accidentally overwrite it. Preferably the archive would also be offsite. If you want to have a backup, then merely making a copy that you keep for six months or a year is fine. If you want to ensure long-term archival integrity then you need multiple redundant copies in your archive system, which is one way that RAID can used (e.g. modes 5, 6, 1+1, etc.). Dragons flight (talk) 15:13, 9 November 2010 (UTC)

RAID only insures the integrity of data which is being kept online (i.e. on a live computer) against the failure of one or more of that computer's HDDs. As for my 'AFAIK', now I agree it is incorrect. The first SATA HDDs (atleast those I'm aware of) appeared in Sim Lim Square in 2004.

In addition, I should rephrase my question for clarity, as the discussion seems to have turned to whether computer bus interfaces which are used today will still be around in 50 years time, and the use of fault-tolerant systems.

   I want to know which of the following types of media have the longest archival life, i.e. which type of media will preserve data integrity for the longest period of time before physical degradation begins?

1. Ultrium LTO-5 750GB Rewritable Tape Cartridge
2. Removable Serial ATA Hard Disk Drive
3. Ultra Density Optical 30GB Rewritable Disk
4. Multi-Level Cell Solid State Drive

Rocketshiporion♫ 06:12, 9 November 2010 (UTC)

I suspect you need to specify more details before there's any possibility for a helpful answer. In particular, you haven't specified under what conditions the media are going to be stored. While all media tend to do best in controlled low humidity, cold environments, some do worse then others in less ideal environments. For example you live in Singapore I believe. From my experience tapes tend to get moldy relatively easy in Malaysia so I would guess the same in Singapore if you aren't controlling the conditions. While I would presume archival tapes do better, I still suspect tapes would perform relatively worse in uncontrolled hot and humid environments compared to controlled archival environments then hard drives or SSDs. Of course it's possible tapes are always worse or always better but narrowing down the conditions should help you find an answer. In particular, for such long archival times, I wonder what sort of figures exist for random uncontrolled environments. I also wonder how well stimulated aging conditions pick up these sort of things. I mean even CDs can get moldy... Nil Einne (talk) 06:36, 9 November 2010 (UTC)

The media will be stored in airtight Pelican 1500NF Hard Case(s) at room temperature, which varies between as low as 20°C and as high as 36°C in Singapore. Rocketshiporion♫ 07:28, 9 November 2010 (UTC)

Sun says LTO-5 is good for 30 years if archived at 16-32 C and 20-80% humidity [3]. RDX hard drives are supposedly safe for 20 years if stored at < 29 C and 5-60% humidity, or 30 years at < 25 C and 5-95% humidity [4]. HP says UDO is good for 50+ years if archived at 5-55 C and 3-90% humidity. [5] I can't find anyone that gives an estimated lifetime for SSDs when used for archival purposes (probably they are just too expensive to be competitive as an archival media). MLC SSDs can survive ~10000 write cycles, which gives a 5-10 year lifetime for normal daily usage, but no idea about the failure mode for archival storage. One should be a bit careful comparing these numbers since it could be apples to oranges unless one can figure out the failure standard being used. For example, a manufacturer could give a mean time before failure (MTBF) of 30 years even if a significant fraction failed in 10-15 years, as long as some are expected to last ~50 years. A more appropriate standard for archival media, and the one more often given by independent testing labs, is the time to a 1% chance of failure. However, without more information you can't be sure what standard is actually being quoted. However, I would reiterate that relying on any single copy media is largely stupid. For archival purposes it is much better to have multiple redundant backups. For example, if you have two low-cost copies of a data set with an independent 5% chance of failure, then the chance both fail at the same time is only 0.25%, which is safer than having a more expensive copy with a 1% chance of failure. With a little thought and engineering, a triply-redundant systems can be both much safer and cheaper than typical single copy archival systems, even if they do require more maintenance on average. Dragons flight (talk) 13:43, 9 November 2010 (UTC)

Let me expand on my comment about redundancy a bit. Suppose you use four standard commercial hard drives of decent (but not exceptional) quality. Every six months you write a copy of your data to all four drives and verify that the image on each drive is correct (which wards against having a drive that is already defective at write time). Under heavy load a typical hard drive has a six month failure rate of 1-3% and far lower failure rates under storage conditions (this does not include a portion of drives that would be fully or partially defective at installation due to manufacturing defects, since these would fail the data verification at write time). Let's assume the failure rate for archiving is 1% per six months (the actual failure rate under storage conditions is probably far lower, with some standard drives demonstrating failure rates less than 0.05% when not used for a year). Each drive contains a complete copy of the data and the archive will be considered intact as long as at least two of the four drives are functioning and agree on what the data should be. Every six months you check the drives for errors and replace any that are no longer accurate. Under these assumptions, the odds of a data integrity failure during any six month window (defined as at least three drives failing) is 0.0004%. On average you will replace 0.04 drives every six months. After 30 years you will have bought about 7 drives and have 99.97% chance your data is intact, a significant improvement over the typical 1% chance of failure used for single copy archival standards. If you are willing to tolerate a personal maintenance hassle, the whole thing could be set up using 4 external 2TB hard drives for under $500. Or a decent archival system could automate the work for a few to several hundred dollars beyond that. Dragons flight (talk) 15:01, 9 November 2010 (UTC)

Assuming you have control over the data format, you can also compute and store a checksum (as a CRC or an SHA hash, or similar), and recognise even a single intact copy with very high probability indeed, i.e. you can recover the data if 3 of the 4 drives fail. --Stephan Schulz (talk) 08:45, 10 November 2010 (UTC)

What does this line of assembly mean?

Hi! I'm working on some assembly code for class and I haven't been able to understand this line:

movzbl (%eax), %eax

For some context, %eax, before calling the operation above was a char *. After this function, the value of %eax became an integer. From the little I picked up, this has something to do with turning of the first 24 bits (?) of %eax, and doing something to the rest (?) Could some one please me out? Thanks! —Preceding unsigned comment added by Legolas52 (talk • contribs) 14:58, 8 November 2010 (UTC)

movzbl means retrieve (move) a byte (a char), add 24 zero bits to it (to make it a long, which is 32 bits in x86 assembly, whatever long is in C), and store it. The parentheses mean "the memory at address" (so they correspond to the * operator in C). So it's equivalent to eax=*(char*)eax; in C, bearing in mind that assembly has no strong type system so things like casts are entirely implied by how you use the values in question. You thought it meant "turning off the first 24 bits", but it actually means "adding 24 'off' bits". --Tardis (talk) 15:25, 8 November 2010 (UTC)

Eye Altitude in Google Earth

What does the eye altitude shown on the Google earth screen mean? —Preceding unsigned comment added by 113.199.204.115 (talk) 15:15, 8 November 2010 (UTC)

I'm assuming it's the distance to where you're pointing from the camera view. General Rommel (talk) 06:43, 9 November 2010 (UTC)

It is not the answer I expected.Please anybody could explain it better. —Preceding unsigned comment added by 113.199.165.180 (talk) 12:09, 10 November 2010 (UTC)

It is how high in the air you would have to be for the view below you to approximate the view in Google Earth. Basically, how high the camera is in the air Zzubnik (talk) 11:44, 11 November 2010 (UTC)

To rephrase what Zzubnik has explained, it's the approximate altitude at which your eyeball must be, in order for your eye (when looking straight down) to see the same view as is shown by the camera. Altitude is basically the 90°-vertical distance from the ground to any point above. Rocketshiporion♫ 04:56, 12 November 2010 (UTC)

Mongolian script support

Hi, I know this is getting a recurring question here, but I checked Multilingual support as well as Multilingual support (East Asian) and did some Google searching, and yet didn't manage to find from where I can download support for the Mongolian script so my browser can get able to view this properly: ᠤᠯᠠᠭᠠᠨᠪᠠᠭᠠᠲᠤᠷ. Could you help me? --Theurgist (talk) 15:40, 8 November 2010 (UTC)

mongolian script should work in Vista. If you are not having success try downloading from one of the pages like School of Computer Science McGill University - the menksoft ones keep giving me a virus warning and attack warnings! Chaosdruid (talk) 02:23, 10 November 2010 (UTC)

I'm using Windows XP. From the link you provided, I downloaded fonts like CMs Huree, CMs Huree dp, CMs Ulaanbaatar, CMs Ulaanbaatar dp, CMs Urga, CMs Urga dp, but I am still seeing the above word or any other Mongolian script text on the Internet as a horizontal string of boxes. Text copied and pasted on MS Word is invisible too, and when I try typing something there by using the "Insert symbol" option, it aligns horizontally instead of vertically. --Theurgist (talk) 15:04, 11 November 2010 (UTC)

This is probably a dumb question but when you say "downloaded" you do mean you also installed the fonts right? And restarted the comp (probably not necessary but always advisable when having unexpected probs particularly with XP. Nil Einne (talk) 15:46, 12 November 2010 (UTC)

If installing the fonts means just copying and pasting the .ttf files to the "Fonts" folder of Control Panel - yes, I did that. I have turned my computer on and off many times since then. But just now I realised I kind of misinformed you with my previous comment. Here's what the situation actually is. I can open MS Word and select the "Symbol" option from the "Insert" menu. There I can specify "CMs Ulaanbaatar" as the "Font". This produces a list of just a hundred or so exotic-looking symbols, but their descriptions at the lower part of the dialogue box are merely DIGIT ZERO, DIGIT ONE, LATIN CAPITAL LETTER A, LATIN SMALL LETTER F, RIGHT PARENTHESIS and so on. When I insert a symbol therefrom by clicking the "Insert" button, and then copy-paste it to, say, the Wikipedia search box, it appears simply as a Latin character, number or some punctuation mark. When I copy-paste ᠤᠯᠠᠭᠠᠨᠪᠠᠭᠠᠲᠤᠷ to MS Word, it keeps looking as a horizontal string of boxes. Note that I have previously downloaded and installed several other types of fonts (Khmer, Ethiopic, Glagolitic, Inuktitut, etc) but I have never had any problems of that kind. (Sorry, I'm not too good at using computer terminology.) --Theurgist (talk) 23:13, 12 November 2010 (UTC)

Now I figured it out. The fonts I installed are fake. They're just ordinary Latin fonts designed to visibly resemble the Mongolian script. So the problem is not with my computer, but I still haven't found the support I wanted. --Theurgist (talk) 07:10, 13 November 2010 (UTC)

Subliminal

I spend all day in front of a computer. I want to flash subliminal messages to myself on the screen at random intervals, for example "Get a girlfriend" or "Improve your life". What programs could do this, easily and at low or zero cost? Thanks. —Preceding unsigned comment added by 178.66.8.154 (talk) 19:40, 8 November 2010 (UTC)

"J", would be the first one. "Y" would be the last. In between, in no particular order, "I", "O", "N", "H", "T", "E", "N", "V", "A".... 85.181.151.31 (talk) 20:31, 8 November 2010 (UTC)

There are a number of software that will do this. Here's one.

Be aware that this type of subliminal message has never been shown to work. The original work by James M. Vicary was almost certainly a hoax designed to promote his own company that would have specialized in subliminal advertising. [6]APL (talk) 21:54, 8 November 2010 (UTC)

You'd do better just to have your machine speak the words or pop up a dialog - there's no evidence that subliminal messages work at all with sentences, and only produce mild generic effects with simple stimuli. If you think it will work and you want to try it, a flashed image of a playboy bunny or a wad of cash will work better than the phrases 'get a girlfriend' or 'improve your life.' though keep in mind that different people have different subliminal ranges, so something that doesn't register consciously on your brain might be perfectly visible to others, which might cause you some consternation if other people are around.

Really, though, if you think you need outside help to get yourself motivated, go spend an evening in the self-help section of your local chain bookstore. there are a number of motivational-type books that are actually reasonably effective at helping you get in gear. they'll be a lot better than toying with subliminal suggestion, anyway. --Ludwigs2 22:05, 8 November 2010 (UTC)

saving a level of game

Is their general method for saving the levels of games. ho can a save the level of old games dave1 ,dave2 —Preceding unsigned comment added by 119.154.120.194 (talk) 21:20, 8 November 2010 (UTC)

Do you mean from the point of view of a game programmer, or from the point of view of game player? Every programmer writes their save-game code a little differently, so no. From the point of view of a player, the answer is normally no; but if you happen to be running the game within an emulator, all emulators that I have seen allow you to pause the virtual machine, and you can usually then save a copy of the virtual machine. For example, you could be using Virtual PC or VirtualBox to play your old copy of Doom 2, and then pause the game when you meet the big boss, save the state of the virtual machine to a file (meaning all of the contents of RAM, the CPU registers, etc are saved), and then load that virtual machine as many times as you want in a row until you win. Poor Doom 2 - there's no way for it to even know you've played it 666 times like this. Comet Tuttle (talk) 21:48, 8 November 2010 (UTC)

Doom and Doom2 has quite usable saving system. This feature would be more necessary for games which cannot be saved normally. -Yyy (talk) 12:26, 10 November 2010 (UTC)

Probably though serialization of the current variables/settings.Smallman12q (talk) 23:35, 10 November 2010 (UTC)