Well, I don't know if you call it spamming or cramming, but how can you keep people from uploading duplicates of videos or cramming your video web site with bad or questionable videos, or uploading the same video over and over? Any ideas? thanks.

probably the easiest way to identify duplicate videos is by creating an MD5 hash for uploaded videos. Compare the hash and file-size to find duplicates.

As for questionable videos...Only way I can think of is through a review process.

I think the only way to do it is a review process. The technology isn't really practical to try to have a program look at the videos and determine if they are appropriate or not.

Can you enlighten on this:
"Creating an MD5 hash for uploaded videos. Compare the hash and file-size to find duplicates"

I'm not familiar with "hash".

Thanks

Most every unix/linux based system has a commandline utility md5sum.
just: md5sum file

Also, just about every language has a library for doing the same.
For example, with java see:
http://java.sun.com/javase/6/docs/ap...ageDigest.html
and
http://java.sun.com/javase/6/docs/ap...putStream.html

with PHP you'd do something like:

$upload = $_POST['file']; 
$file_hash = md5($upload);

Originally Posted by willcode4beer

Most every unix/linux based system has a commandline utility md5sum.
just: md5sum file

Also, just about every language has a library for doing the same.
For example, with java see:
http://java.sun.com/javase/6/docs/ap...ageDigest.html
and
http://java.sun.com/javase/6/docs/ap...putStream.html

with PHP you'd do something like:

$upload = $_POST['file']; 
$file_hash = md5($upload);

So, how will you deal with videos that have a titlte sequence before the copied video starts? The md5's will be completely different and your comparison will return that the videos are not the same when they actually are.

Side note: The PHP command to hash files is "md5_file($file_to_hash);". Calling a md5 of $upload will just hash the filename of the video specified.

The md5 hash will only work if it's the same file.
For some things, you'll either want voice recognition software feeding text to a search index or, you'll just need to have reviews and reporting.

The later will be easier.

I am considering charging a small fee for the first video uploaded by a user. With the theory that a spammer mindset may not want to pay, therefore it automatically seperates the good from the bad.

I'd be interested in feedback on this idea. Thanks.

what about monitoring the usernames etc - the text fields of the members, and filtering out spam that way - most spammers are "professional" spammers, ie spend days spamming tonnes of crap everywhere - thus their presence on your machines is usually large collections of crap with slightly different names but usually all easily linked by some easy "rule" or other. filtering them can be very easy. no need to go all the way into the video itself!

(fee idea will backfire and kill your site - why can't they go to youtube and upload it for free? why pay someone who's infinitely smaller than youtube, when youtube gladly let her/him do it free?)

monitor the email id and IP address these 2 things are sufficient to block spammer ...

ah, i already gave my view on this one. no need to repeat myself.

Originally Posted by chrisj

I am considering charging a small fee for the first video uploaded by a user. With the theory that a spammer mindset may not want to pay, therefore it automatically seperates the good from the bad.

I'd be interested in feedback on this idea. Thanks.

You know, your game plan really depends on what kind of a site you run.

For instance, do you run the kind of site where you can charge users to use your site and not expect them to leave? or, like HairyGunther has suggested, just go to youtube- or any number of sites on the web and do it for free? Would you be chasing away legit people?

I don't think I could provide a sincere answer without knowing much about your site.


so far, I think willcode4beer has suggested about the most feasible answer here so far: as some of these guys can create 100 emails in 5 minutes, and sit on about 500 different IP addresses, if your site is THAT valuable to them, they will continue to find ways to exploit your system every time.

Perhaps hiring a programmer will be in your best interest.

A possibility (if you have the programming inclination and a good server to do the processing required) is to create a video 'fingerprint' when processing an upload (create a strip of pixel thumbnails periodically e.g every 5 or 10 seconds and gather data from this) then use some math to compare against existing fingerprints and flag up videos that match an existing pattern within a certain threshold.

Brilliant!

very gud web & very nice Work...;-)