Jump to content

Wikipedia:Reference desk/Archives/Computing/2015 October 24

From Wikipedia, the free encyclopedia
Computing desk
< October 23 << Sep | October | Nov >> October 25 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


October 24

[edit]

How do websites tell whether a user is a bot or human?

[edit]

How do websites tell whether a user is a bot or human in a technical way? And why is it important? What does it have to do with spam? If a user wants to spam a website and cannot create a bot to automate the process, then can't he just spam manually or get a whole team of human spammers to spam a website? What kind of knowledge does one need to create such an algorithm? 71.79.234.132 (talk) 01:18, 24 October 2015 (UTC)[reply]

A CAPTCHA is one method. (The Chili's restaurants website now have a check box you click that says "I am not a bot". Somehow I doubt if that is bot-proof, though.) StuRat (talk) 02:09, 24 October 2015 (UTC)[reply]
That's not what I meant. I wondered about the logic behind the algorithm. 71.79.234.132 (talk) 04:01, 24 October 2015 (UTC)[reply]
I'm pretty sure the "I am not a bot" thing is part of Google's ReCaptcha, as I've seen it on other sites. ReCaptcha is a fancy JavaScript thing. You only get the "I am not a box" checkbox from ReCaptcha if it thinks you're likely a human, based on a bunch of heuristics. If you visit a page numerous times or do some other things that lower ReCaptcha's confidence, you'll trigger the actual captcha. --71.119.131.184 (talk) 21:08, 24 October 2015 (UTC)[reply]
Thanks, but then why even ask the user to check the box ? StuRat (talk) 19:52, 27 October 2015 (UTC)[reply]
As for why it's important, you just can't get enough humans to do the same thing at once to create a denial of service attack, for example. But if you have bots running from multiple computers, perhaps spread by a virus, then you can.
For another example, lets say they are trying to crack a password on an account. A person could only try maybe a few hundred combos until their fingers would get sore, but a bot could try millions of combos. Of course, there are other ways to limit the number of tries, like a limit on the total count or a time delay between tries.
As far as how to tell a person from a bot, the basic idea is to find something that people do better than bots. Character recognition is a common one, where bots are easily confused by extraneous lines and curves. Speech recognition is another area where bots are easily confused. StuRat (talk) 04:14, 24 October 2015 (UTC)[reply]
Part of the detection will be to see repetative activities, eg the same pattern of text being added, or the same spam link. THis may not tell a bot from a team of spammers, but detecting and preventing it will have the desired effect. Graeme Bartlett (talk) 05:01, 24 October 2015 (UTC)[reply]

what is datarate in bytes for analog crt tv recieving via coaxial cable?

[edit]

OP watch 28 inch crt tv and curiousMahfuzur rahman shourov (talk) 13:43, 24 October 2015 (UTC)[reply]

"data rate in bytes" really only makes sense for digital signals, not analogue ones, which you're asking about.--Phil Holmes (talk) 14:23, 24 October 2015 (UTC)[reply]
See SECAM, PAL and NTSC bandwidth including audio sub carrier which is 3,58 or 4,433 MHz from the main carrier. A TV antenna cable is made to support up to 1 GHz of RF, a sattelite cable is made for frequencies up to 2 GHz. A digital receiver uses a symbol rate in decoding. A digital stream is beeing decoded and buffered. It is based on MPEG or similar video codec. --Hans Haase (有问题吗) 15:10, 24 October 2015 (UTC)[reply]
Firstly, it doesn't matter what size screen you're viewing it on. But I agree with Phil that it's not a very meaningful question. But we can get some sort of a handle on it. In the UK (for example), old-style analog TV using the PAL system had 625 lines of video, with half of them refreshed every 50th of a second, so that's 15625 lines per second. If we imagine that the horizontal resolution is 4/3rds of the vertical resolution (a somewhat rash assumption - but we're trying to get a ballpark figure here) then there would be roughly 830 horizontal "pixels" of information....so very roughly, this is an 800x600 image being displayed at 25Hz frame rate. There is also video compression of a sort because the color information is sent at lower bandwidth than the monochrome. A complicating factor is that (being analog) there is no concept of "bits per pixel" - the precision of the color depends entirely on the quality of everything from the camera to the TV. If I had to come up with a number, I'd say it was close to sending 25 800x600 JPEG-encoded images per second...but the "quality" level is hard to assess. Real digital video uses frame-to-frame compression, which saves a TON of data compared to analog transmission. SteveBaker (talk) 16:54, 24 October 2015 (UTC)[reply]
SteveBaker, in US, most of America and wide parts of Asia 60 Hz were used! --Hans Haase (有问题吗) 19:26, 27 October 2015 (UTC)[reply]
Yes - but the number of lines in an NTSC image is proportionately smaller - I think the bandwidth is pretty comparable. There are at least a half dozen different standards in use around the world, with different frame rates and number of lines. But we're not going to get anything remotely like an accurate number - it's analog, not digital. 19:59, 27 October 2015 (UTC)

Installing Windows 95 From CD

[edit]

I have a CD copy of Windows 95 that I'm trying to install on an old desktop tower. The tower has Windows XP installed on it, so to install Windows 95, I have to boot from a floppy into MS-DOS and run the CD from there (the CD itself isn't bootable). I've done this, but after I boot to DOS, I can't access the CD drive: I get the "Invalid Drive" message when I try and switch to it. I'm guessing this is because I don't have a DOS driver for the CD drive, but I'm not sure. Is that the problem here, and if so, what is the probability that I can find a driver for a CD drive built around 2002? OldTimeNESter (talk) 17:54, 24 October 2015 (UTC)[reply]

I'm not a DOS expert, but I think you might be right. Is installing Windows 95 on the system really something you want to do? If you're interested in retrocomputing, it's way, way more convenient to run old software under a virtual machine. --71.119.131.184 (talk) 21:01, 24 October 2015 (UTC)[reply]
Setting up old PCs with obsolete versions of Windows is a fine way to use an old tower, and a lot of those old PCs use very little power as long as you use an LCD monitor. I wouldn't advise accessing the internet with Win95 unless you have an external firewall, though.
Try going to http://www.answersthatwork.com/Downright_pages/Boot_CD_and_Boot_Disk_Images_and_ISO_free_downloads.htm and downloading "Boot Disk - Windows 95a boot disk with internal CD-ROM support (USA Version)". While you are at it, get " Boot CD – MS-DOS 6.22 Boot CD with internal CD-ROM support (USA Version)" for those times when you don't have a working floppy. --Guy Macon (talk) 22:20, 24 October 2015 (UTC)[reply]