Wikipedia:Reference desk/Archives/Computing/2016 January 9
Computing desk | ||
---|---|---|
< January 8 | << Dec | January | Feb >> | January 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. |
January 9
[edit]Two cable modems.
[edit]Can I connect two cable modems to the cable outlet in different rooms of my house? I have a single Time Warner cable internet account and I can't get a straight answer from their stupid call center people.
SteveBaker (talk) 03:48, 9 January 2016 (UTC)
- No, this would not work, as the modem would be sending data, and two would interfere with each other. Graeme Bartlett (talk) 03:57, 9 January 2016 (UTC)
- Ack. That's what I suspected. OK - many thanks. SteveBaker (talk) 04:02, 9 January 2016 (UTC)
- If you are trying to connect two computers to the internet, you need either a router or a cable modem with a router built in. Bubba73 You talkin' to me? 04:25, 9 January 2016 (UTC)
- Do you mean with a single account? If so, I agree it probably won't work although the reason would be because it's unlikely Time Warner will like you logging in twice.
If you have two accounts (or if Time Warner allowed it on a single account), it seems to me this would work and these sources seem to agree although they don't refer to Time Warner in particular [1] [2] [3] [4]. The fact that both modems will be sending and receiving data would seem to be irrelevant, since that's how Cable Internet access works. It's a shared connection. Multiple people are connected to the same cable (on the last mile). The fact that the cable outlets happen to be in the same house would likely be irrelevant.
For the upstream, the cable modems in collobration with the Cable modem termination system are already sharing the upstream somehow, probably via Time division multiple access according to some standard, probably some variant of DOCSIS. See [5] [6] [7] for example. (This is happening with the downstream too, but it's easier for people to understand since the CMTS is the one sending all the data.)
We aren't talking about DSL or anything using phone lines where the last mile is unshared, so normally only one device is capable of using it at one time. (Well unless you count the DSL and voice components seperately where one device uses the higher frequencies and one device uses the lower ones.) Passive optical network is somewhat similar to cable albeit with fewer people sharing the same last mile.
The only reason I can think of why using two cable modems wouldn't work if the operator allowed it, would be if they'd already reached the limit of devices on that line. Or if there was some filtering from where the cable is split off from the main line before entering your house. (I don't know that much about how splitting the cable before entry in to a premise is done.)
I guess there is a slight possibility that the portion of cable in your house including any splitters is too poor to handle multiple devices, but it seems unlikely for only 2. And I'm not sure if this is really possible or it'll normally mean it won't work with even one. And perhaps if the input for the second modem is actually coming from the first modem, the first modem has some internal filtering. (If the cable is simply split with typical coaxial splitters, then this doesn't apply.)
- Do you mean with a single account? If so, I agree it probably won't work although the reason would be because it's unlikely Time Warner will like you logging in twice.
- →Port trunking. When providers do not block a second login, it might be possible, but needs to be supported by the hardware as well. Many providers are blocking the second login due stolen passwords. As DSL needs a login and password, it is not just aggregating two dedicated ports on a managed switch. When the prodider supports mutible login, You still need to have or get identical WAN IP addresses. --Hans Haase (有问题吗) 20:29, 13 January 2016 (UTC)
What does this mean in C?
[edit]What does
x = (t>c) ? t-c : t;
mean in C? (I don't know what the question mark and colon do). Bubba73 You talkin' to me? 04:24, 9 January 2016 (UTC)
- The ? means test, if it is true the the bit before : is evaluated, and if it is false the bit after. the same as if T>c then x=t-c else x=t; (syntax may be wrong) Graeme Bartlett (talk) 04:34, 9 January 2016 (UTC)
- Our article on this is at ?:. Graeme Bartlett (talk) 04:36, 9 January 2016 (UTC)
- As the article says, this is commonly referred to as the "ternary operator". --71.119.131.184 (talk) 07:02, 9 January 2016 (UTC)
- With shades of the "computed goto" in old Fortran. Bubba73 You talkin' to me? 01:04, 10 January 2016 (UTC)
- As the article says, this is commonly referred to as the "ternary operator". --71.119.131.184 (talk) 07:02, 9 January 2016 (UTC)
- Our article on this is at ?:. Graeme Bartlett (talk) 04:36, 9 January 2016 (UTC)
Thanks
Bubba73 You talkin' to me? 05:32, 9 January 2016 (UTC)
AT&T to MIPS Translation
[edit]Hi, I want to translate some code from AT&T to MIPS, and I encountered the directive ".long". What does this mean? What is its equivalent in MIPS? 213.8.204.41 (talk) 16:17, 9 January 2016 (UTC)
- Two things. One, "AT&T" isn't an assembly language. It's one of the two forms of x86 assembly language in use. Saying you want to "translate...from AT&T to MIPS" is like saying "I want to translate from capital letters to Japanese". Second, .long isn't an assembly language instruction at all. It's an assembler directive. Depending on what assembler you're using you may not need to "translate" anything at all. But, and don't take this personally, your questions indicate you're probably new to assembly language programming, and have some wrong ideas. I suggest reading a good book on assembly language programming before trying to jump right in. --71.119.131.184 (talk) 20:43, 9 January 2016 (UTC)
Protect data
[edit]On Windows, what is the best way to ensure the data integrity of 200GB of assorted files and prevent data corruption? Simply duplicating the files does not protect from data corruption unfortunately.
Thank you
62.37.237.16 (talk) 16:39, 9 January 2016 (UTC)
- Backups. Lots of backups, some offline, stored in different places. -- Finlay McWalterᚠTalk 16:51, 9 January 2016 (UTC)
- But if the original data gets corrupted, all the backups will be corrupted too since they are just copies of the original corrupted data. Simply duplicating the files does not protect from data corruption unfortunately. I'm looking for a way to ensure the original data doesn't get corrupted. 62.37.237.16 (talk) 17:00, 9 January 2016 (UTC)
- So you don't continually re-use all the backup media - you adopt a data archival strategy which involves keeping some of the dated backup images for lengthy periods of time, or forever. The precise strategy that's appropriate for you depends on your rate of changes, the threat models you're seeking to address, any regulatory obligations you may have, and your budget. -- Finlay McWalterᚠTalk 17:19, 9 January 2016 (UTC)
- (EC) The point is to backup before the original data is corrupted. The shorter the time frame between the file creation and a backup, the least time there is for corruption (although with very short time frames people may have bad backuping practices). And it does protect against data corruption since if the original data is corrupted after you backup, you use one of the backups instead. Note that you aren't supposed to delete the backups and take a new backup of the original data unless you're sure it's the same data. You may have a problem determining which copy of the data is correct, particularly if you can't find any two copies which are alike or you have reason to believe both may be corrupted but while you can reduce the risk of this, it's always going to be there.
You can add hashes or some other form of error detection to try and detect corruption. Many backuping tools may do this automatically, and some file systems have this built in. However this doesn't stop corruption, it simply allows you detect it after the fact. You can add a level of error correction to try and recover from when corruption occurs. To some extent a number of forms of RAID have this as do file systems like ReFS and ZFS. But such integrity checking and monitoring will only provide a baseline of protection, it's not going to help of corruption happens in a major way for some reason. You're far better having regular backups than just having such protection. Combining them both may be a good idea if it's really matters although in such a case you may also be better off seeking help outside the RD.
It also depends on what sort of corruption you're referring to. If it's intentional, it's very difficult to prevent, since any form of integrity or error detection you build in could likely be counteracted by whatever is corrupting your data. That's again where backups are an advantage since if it done right, it's far more difficult for the adversary to affect all your backups.
- (EC) The point is to backup before the original data is corrupted. The shorter the time frame between the file creation and a backup, the least time there is for corruption (although with very short time frames people may have bad backuping practices). And it does protect against data corruption since if the original data is corrupted after you backup, you use one of the backups instead. Note that you aren't supposed to delete the backups and take a new backup of the original data unless you're sure it's the same data. You may have a problem determining which copy of the data is correct, particularly if you can't find any two copies which are alike or you have reason to believe both may be corrupted but while you can reduce the risk of this, it's always going to be there.
- Depending on what you're trying to accomplish, other approaches might be (1) simple access controls, or more likely (2) checksums (especially cryptographically secure ones). —Steve Summit (talk) 15:51, 11 January 2016 (UTC)
9 pin DE-9 PC video display port
[edit]I just got an old IBM clone PC, not sure quite how old, but it's running Windows XP. It has a nice graphics card I would like to salvage (two standard 15 pin female (socket) D-sub connectors with high resolution), but that would mean I could no longer use that old PC, because it's only display port on the motherboard is a male DE-9 connector (plug), see the top pic at D-sub. Any idea if this old display port would work with a modern monitor that expects a 15-pin VGA connector or better ? What resolution could I expect if I could find a cable to do the conversion ? (My other option is to find a low res graphics card and put that in this old PC.) StuRat (talk) 19:56, 9 January 2016 (UTC)
Out D-sub article says: "A female 9-pin connector on an IBM compatible personal computer may be a video display output such as MDA, Hercules, CGA, or EGA (rarely VGA or others). Even though these all use the same DE9 connector, the displays cannot all be interchanged and monitors or video interfaces may even be damaged if connected to an incompatible device using the same connector." So, this sort of scares me off, but mine has a male connector. Any difference there ? StuRat (talk) 19:59, 9 January 2016 (UTC)
- Quick question, why do you believe the DE-9 is video out? Did the manual or markings on the motherboard indicate it's video out? Have you ever actually seen the DE-9 on the computer used for video out? Do you even know if the motherboard has video out? (There should be signs of it in the BIOS and probably in Windows.) I find it fairly unlikely a computer capable of running Windows XP would have a DE-9 video out. I wouldn't be surprised if such a computer has a DE-9 RS-232 port and those were IIRC normally male. Nil Einne (talk) 20:28, 9 January 2016 (UTC)
- It has a monitor icon next to it, etched into the sheet metal. It does seem to have a curious mix of very old and fairly new stuff. That graphics card is better than anything I have on my newer PCs, and yet it has this ancient hardware port. I am guessing that this was an outdated port when the PC was new, included just for backward compatibility, and that the graphics card was added far more recently. StuRat (talk) 23:42, 9 January 2016 (UTC)
- I think it's more likely that the case markings don't match the motherboard (probably the motherboard has been replaced at some point in the machine's history). A real EGA connector would be female (not male), and the standard has been obsolete for nearly thirty years (1987); a female DE-9 connector is far more likely to be a joystick/MIDI port. But if your connector is male, it's almost certainly RS232. Tevildo (talk) 00:00, 10 January 2016 (UTC)
- What would it be for then, a joystick ? It also has a game port, so it does seem to have originally been a gaming PC. StuRat (talk) 01:25, 10 January 2016 (UTC)
- A female connector might be for a "traditional" digital joystick, or for some non-standard serial bus (such as a UPS controller), but I wouldn't expect either to be on the motherboard rather than a separate card. If, however, it's male, it's an RS232 serial port. Tevildo (talk) 01:50, 10 January 2016 (UTC)
- What would it be for then, a joystick ? It also has a game port, so it does seem to have originally been a gaming PC. StuRat (talk) 01:25, 10 January 2016 (UTC)
- Regardless, you can just throw a dirt-cheap PCI graphics card in there (or AGP, if it has an AGP slot). You can probably get one for free from discarded computers if you do a little looking around. Note that's PCI, not PCI Express, which is only on modern hardware. --71.119.131.184 (talk) 20:50, 9 January 2016 (UTC)
- Actually, the more I think about this, the more I'm curious. Are you sure that graphics card will work in the PC you want to use it in? PCI graphics cards are ancient, and no recent PCs have AGP. Modern motherboards are all PCI Express, some with one or more PCI slots for compatibility with old hardware. However I don't think modern motherboards support using PCI graphics cards even if they have PCI slots. Also PCI is limited in speed and bandwidth, which is why AGP was developed as a stopgap measure, as graphics cards were butting up against its limits, so I kind of suspect the graphics card is not PCI. Do you know that PCI and PCI Express are totally different things that aren't intercompatible? The only thing in common is the "PCI" name; you can't plug PCI cards into PCI Express slots (and you can't plug AGP cards into either). --71.119.131.184 (talk) 00:54, 10 January 2016 (UTC)
- I hadn't considered that. I was hoping to put it in one of my Windows 7 PCs, but, if it won't fit there, I have an old Windows 98 PC. (You might wonder why that is preferably to leaving it in the Windows XP box. That one seems to have some problem with the web, which prevents it from loading most pages. I suppose I could try to fix that, but just taking out the graphics card, which is about the only good thing in that PC, seemed to make most sense at the time. I was planning to then give it to my Mom to play solitaire on it. She doesn't even have a web connection, and has no need for a high res dual monitor display.) StuRat (talk) 01:22, 10 January 2016 (UTC)
More details
[edit]The old Windows XP PC has a Radeon 7000/Radeon VE graphics card, with dual monitor displays up to 2048×1536 each. It has a Celeron with a motherboard driver date of 7/1/2001.
I have two Windows 7 PCs I might like to put that graphics card into:
- Optiplex 320
- Celeron G460
So, will that graphics card work in the newer PCs ? Any more info I need to get ? StuRat (talk) 01:48, 10 January 2016 (UTC)
- The 7000 VE is a really ancient AGP card - see this site for technical details. It's not compatible with the Dell OptiPlex 320 (which has rather better on-board video). If your Celeron motherboard has an AGP slot, then it'll work, but you should be able to pick up something more modern for (literally) nothing - people throw away cards that are much less than 15 years old. I would recommend leaving it in the XP machine. Tevildo (talk) 02:04, 10 January 2016 (UTC)
- Yeah, I kind of suspected so. You can't just assume that any PC part will work in any system. Other advice: basically all modern PCs ship with onboard video, which is generally just fine for 2D graphics. And if you are going to benefit from a graphics card, you aren't going to want to use some ancient card, leaving aside the compatibility issue. You can get a quite good graphics card for cheap, especially if you avoid the newest models, which are marked-up to profit from early adopters. Newegg often has good deals on computer hardware (though they only sell to some countries). --71.119.131.184 (talk) 04:43, 10 January 2016 (UTC)
- The on-board graphics for my Optiplex 320 and Celeron G460 are each only 1600×1200, single monitor. So how is this better ? Faster response time ? You say I can get better graphics cards for nothing now ? Where ? StuRat (talk) 04:32, 10 January 2016 (UTC)
- Try ebay. Anyway I'm fairly sure that the G460 supports dual displays. So does the lowest end chipset [8]. If your computer doesn't I guess it's a motherboard limitation, most have provided dual outputs for quite a while but if it's an ultra small box perhaps they didn't bother. As for resolution, again the C460 and H61 should support at least 1920x1200 on the VGA (probably 2560x1600 on the DisplayPort). If yours doesn't I suspect either a driver problem or user error. Likewise the Radeon Xpress 1100 supports dual monitors (and probably at least 1900x1200 but I'm lazy to check that. The Optiplex 320 not supporting dual displays is down to the manufacturer now bothering to add the necessary for a second display. I believe the Radeon Xpress 1100 had at least one integrated RAMDAC so I suspect again not getting more than 1600x1200 is down to either a driver issue or user error. You may be able to use an ADD2 card for either, but I'm not completely sure they work with the Radeon Xpress 1100 and I can't remember if they still work with Sandy Bridge. In any case, I'm not sure it's worth it since you can probably buy a GPU at least as good as the onboard for about the same price (definitely for the Xpress 1100). Nil Einne (talk) 10:15, 10 January 2016 (UTC)
Edit: Realised the Intel page for G460 does mention dual display support [9]. I looked a bit more and I think a Sandy Bridge computer (with iGPU) should support an ADD2 (Serial Digital Video Out) card. Looking a bit on eBay, comparing what GPU you will get for the $8 inc shipping it costs for an ADD2 card it's complicated which one will be better. So perhaps an ADD2 is slightly worth considering. (I'm now more sure it won't work on the Radeon Xpress 1100, and confirmed my believe it would be a silly choice anyway.) That said, you will want to check compatibility, my impression is that ADD2 cards are fairly generic but I'm not sure if this extends to the iGPU era. Of course if you just want dual monitor VGA 1920x1200 and really don't care about anything else it probably doesn't matter since any GPU from the past 8 years or so should be fine. (Not that I really understand why you want dual monitor VGA 1920x1200/1080 but I digress.)
- For free cards, there's Freecycling, or, if you have a friend who works in an office, they can ask their IT department for any cards they're going to throw out. Dumpster diving may be illegal in your jurisdiction, so I can't really recommend it, effective though it may be. Tevildo (talk) 12:33, 10 January 2016 (UTC)
- Try ebay. Anyway I'm fairly sure that the G460 supports dual displays. So does the lowest end chipset [8]. If your computer doesn't I guess it's a motherboard limitation, most have provided dual outputs for quite a while but if it's an ultra small box perhaps they didn't bother. As for resolution, again the C460 and H61 should support at least 1920x1200 on the VGA (probably 2560x1600 on the DisplayPort). If yours doesn't I suspect either a driver problem or user error. Likewise the Radeon Xpress 1100 supports dual monitors (and probably at least 1900x1200 but I'm lazy to check that. The Optiplex 320 not supporting dual displays is down to the manufacturer now bothering to add the necessary for a second display. I believe the Radeon Xpress 1100 had at least one integrated RAMDAC so I suspect again not getting more than 1600x1200 is down to either a driver issue or user error. You may be able to use an ADD2 card for either, but I'm not completely sure they work with the Radeon Xpress 1100 and I can't remember if they still work with Sandy Bridge. In any case, I'm not sure it's worth it since you can probably buy a GPU at least as good as the onboard for about the same price (definitely for the Xpress 1100). Nil Einne (talk) 10:15, 10 January 2016 (UTC)
- The on-board graphics for my Optiplex 320 and Celeron G460 are each only 1600×1200, single monitor. So how is this better ? Faster response time ? You say I can get better graphics cards for nothing now ? Where ? StuRat (talk) 04:32, 10 January 2016 (UTC)
- Male D-SUB-9 connector were used a RS-232/V.24 interface. XT and AT computers used a female D-SUB-9 connector for CGA, EGA, HGA graphic adapter standards. If Your computer has an ATI rage installed, it is a much newer generation of PC. I guess somebody just installed a mechanical ATX-I/O-shield which declares the second serial interface as a display interface, located at the same place of the board and connectors have same measures except pins and male/female shielded housing.[10] Else, some ATI cards have much space on the slot beside the VGA connector. Maybe some manufacturers replaced the slot shield using this space for other connectors.[11][12] Beside, I do not recommed to meet the FCC EMI regualtions. --Hans Haase (有问题吗) 20:59, 13 January 2016 (UTC)
- The Dell Optiplex 320 has as passive addon a dual PS/2 Mouse/Keyboard-Jacks and the secord RS-232/V.24 assemled on a slot shield, free the choose the slot where to be installed into.[13] --Hans Haase (有问题吗) 21:06, 13 January 2016 (UTC)
Ntuser.dat
[edit]I have 51 Ntuser.dat{______________}.??? and one ntuser.ini, do I need them? They tend to be growing as files since the last time I checked this folder i.e. C:\User\... Can I delete them all? -- Mr. Zoot Cig Bunner (talk) 20:14, 9 January 2016 (UTC)
- See User profiles in Microsoft Windows. Rather than deleting the files manually, go to Control Panel > Performance and Maintenance > System > Advanced > User Profiles, and delete any that you don't want to use. Copy any files from the individual User directories that you want to keep before doing this. Tevildo (talk) 22:39, 9 January 2016 (UTC)
- The default profile is 1.58MB in size, and another profile (I guess the one I'm using) is of 3.99GB. Type, Status column of both are local. I can only change type of the second profile... Is there any way I could decrease the GB of the second profile? -- Mr. Zoot Cig Bunner (talk) 09:51, 10 January 2016 (UTC)
- Clear your browser cache, and delete any unwanted files from "My Documents". You might also want to clear out your Temp directory (C:\Documents and Settings\<UserName>\Local Settings\Temp). Tevildo (talk) 12:21, 10 January 2016 (UTC)
- Your profile is where your "My Documents" and Desktop files are stored, so whatever you have saved there will be in your "user" directory. Get a program called WinDirStat, this will quickly show you what is taking up the space on your hard disk. Vespine (talk) 04:33, 11 January 2016 (UTC)
- Tevildo, Vespine: They're always kept clean, right before I turn off my PC using CCleaner. I use a RHDD and do not keep anything in the PC HDD.
- I'm using Windows 7 Ultimate. I also can't get into C:\Documents and Settings, its a shortcut icon available and it has a lock drawn in it. I used the 'shift' button and 'right' clicked on the icon, I did not find any run as administrator option...
- I'll try with the software vespine stated, if that doesn't delete anything than I have to let it go... I'm okay, as long as the ntuser.dat files (that are increasing in numbers) don't send any information to third parties via the internet...
- Mr. Zoot Cig Bunner (talk) 08:32, 11 January 2016 (UTC)
- Under Windows 7, the temp directory is C:\Users\User Name\AppData\Local\Temp . You can find out where it is for the account you're currently using by typing %temp at a command prompt. Tevildo (talk) 09:01, 11 January 2016 (UTC)
- Found it. There is not much here, I'll try using prefetch, temp and some other words if I could recall, to use it on the run window. I've also downloaded the WindirStat. I'll go through. Thanks guys. Regards. -- Mr. Zoot Cig Bunner (talk) 20:01, 11 January 2016 (UTC)
- Just for clarification re: can't get into C:\Documents and Settings, its a shortcut icon available and it has a lock drawn in it. That directory is deprecated in windows 7, Microsoft kept a "Junction Point" with that name for legacy software, there's nothing in those folders. Vespine (talk) 21:28, 11 January 2016 (UTC)
- Noted! Thanks. -- Mr. Zoot Cig Bunner (talk) 18:21, 12 January 2016 (UTC)
- Just for clarification re: can't get into C:\Documents and Settings, its a shortcut icon available and it has a lock drawn in it. That directory is deprecated in windows 7, Microsoft kept a "Junction Point" with that name for legacy software, there's nothing in those folders. Vespine (talk) 21:28, 11 January 2016 (UTC)
- Found it. There is not much here, I'll try using prefetch, temp and some other words if I could recall, to use it on the run window. I've also downloaded the WindirStat. I'll go through. Thanks guys. Regards. -- Mr. Zoot Cig Bunner (talk) 20:01, 11 January 2016 (UTC)
- Under Windows 7, the temp directory is C:\Users\User Name\AppData\Local\Temp . You can find out where it is for the account you're currently using by typing %temp at a command prompt. Tevildo (talk) 09:01, 11 January 2016 (UTC)
- Your profile is where your "My Documents" and Desktop files are stored, so whatever you have saved there will be in your "user" directory. Get a program called WinDirStat, this will quickly show you what is taking up the space on your hard disk. Vespine (talk) 04:33, 11 January 2016 (UTC)
- Clear your browser cache, and delete any unwanted files from "My Documents". You might also want to clear out your Temp directory (C:\Documents and Settings\<UserName>\Local Settings\Temp). Tevildo (talk) 12:21, 10 January 2016 (UTC)
- The default profile is 1.58MB in size, and another profile (I guess the one I'm using) is of 3.99GB. Type, Status column of both are local. I can only change type of the second profile... Is there any way I could decrease the GB of the second profile? -- Mr. Zoot Cig Bunner (talk) 09:51, 10 January 2016 (UTC)
Can someone give a basic, primal, general explanation for what are "Data serialization languages"? What are they actually? Are they structure languages? What they use for etc? Someone would maybe want to write a short article about this ? Data serialization languages. Ben-Yeudith (talk) 21:50, 9 January 2016 (UTC)
- It's a language that allows a program to represent complex data structures in linear ("serial") text format. Applications are e.g. storage on disk or transmission over the net. A better-known example is JSON. We have an article at Serialization - maybe we can just add a redirect? --Stephan Schulz (talk) 20:30, 10 January 2016 (UTC)
- User:Stephan Schulz, can I simply put that it's actually a kind of "Intermediate language" That takes a data structure in manner 1, serializing it in linear data pieces in another manner (2) in such a way it could be easily re compiled to manner 1 ? Ben-Yeudith (talk) 20:56, 10 January 2016 (UTC)
- I wouldn't necessarily call it "intermediate", but you basically describe the operation correctly. In modern programming languages and operating systems, you cannot simply do a memory dump, since you only have partial (or no) control of the memory subsystem, and moreover, such a dump would often be wastefully large and very brittle against even small changes in implementation. So you need some way to convert your data from a format that allows arbitrary nesting and pointing to a linear format that you can write to a sequential medium (like a tape, or a network port, or, for realistic performance, a disk file).You can write a dedicated printer and parser, of course, but a language like JSON or YAML allows you to automate quite a bit of the process. --Stephan Schulz (talk) 22:01, 10 January 2016 (UTC)
- Thank you dearly Setphan! Much and much of thanks! Ben-Yeudith (talk) 23:30, 10 January 2016 (UTC)
- If indeed in computing the terms "Serialization" and "Data serialization languages" are generally synonymous than sure I would support creating a redirect...
- Thank you dearly Setphan! Much and much of thanks! Ben-Yeudith (talk) 23:30, 10 January 2016 (UTC)
- I wouldn't necessarily call it "intermediate", but you basically describe the operation correctly. In modern programming languages and operating systems, you cannot simply do a memory dump, since you only have partial (or no) control of the memory subsystem, and moreover, such a dump would often be wastefully large and very brittle against even small changes in implementation. So you need some way to convert your data from a format that allows arbitrary nesting and pointing to a linear format that you can write to a sequential medium (like a tape, or a network port, or, for realistic performance, a disk file).You can write a dedicated printer and parser, of course, but a language like JSON or YAML allows you to automate quite a bit of the process. --Stephan Schulz (talk) 22:01, 10 January 2016 (UTC)
- User:Stephan Schulz, can I simply put that it's actually a kind of "Intermediate language" That takes a data structure in manner 1, serializing it in linear data pieces in another manner (2) in such a way it could be easily re compiled to manner 1 ? Ben-Yeudith (talk) 20:56, 10 January 2016 (UTC)
- The suggestion that JSON is "a better-known example" is a tad dubious. As of YAML 1.2, JSON is a subset of YAML. Previously, JSON worked with every popular YAML editor that I know of because the differences were so minor. Then, in YAML 1.2, the differences were worked out and JSON became YAML. Because YAML is a superset, you can say that JSON is YAML, but YAML is not JSON. 199.15.144.250 (talk) 12:39, 11 January 2016 (UTC)