that smart junk filter
 

that smart junk filter should, when you click on something as junk, ask your reason that that particular file is junk. That way files like that in the future can be filtered as junk in the search results

whoa! beat me to it. this is the whole reason i came here!

i am COMPLETELY sick and tired of clicking "junk" on those 177KB, 3,462KB, and all the rest of those files that are obviously the same files over and over with simply /YOUR SEARCH PATTERN HERE/ echoed back as the filename.

all it needs is a really simple popup:

------------------------------------------------
Please select the criteria for this file being junk:

[this box could be the filename...] [filesize..] [some] [other] [fields]

[ OK ] [ CANCEL ]
------------------------------------------------

and then you just highlight whichever fields, or parts of the fields, you feel makes it a junk file.

like if your filename returned was
Sesame Street cute girl has shaking orgasm on webcam

you would just highlight the "cute girl has..." part.

or highlight the filesize on that one damn 3,462kb file that turns up on every damn search i've made in the last year.

PLEASE OH PLEASE YE MIGHTY LORDS OF LIMEWIRE!
YOUR HUMBLE PEONS BESEECH THEE!
SAVE US FROM THE TERRORS OF THE SPAM KINGS!

i dont think this is a good idea it would take me YEARS to do all that in the past 10 mins i have marked over 200 files as junk i am getting sick of it and i am takeing action since it we need a virus scan BEFORE upload this would solve the problem.

Make it MANDATORY. Dont let files be uploaded with out virus scan first this would solve our problem. I dont just mean with one program. I mean with Nortan, Bull Dog, AVG, Spybot search and destroy, etc . . .

Sure this would probally lag the uploads to the server but still at least it wouldnt be so junky

and where precisely would this virus scanner be put?
the upload is from the (alleged) spammer's computer. from there it goes thru the "series of tubes" to the downloaders computer.

there is no 'central location' where all the files could be scanned before being made available to downloaders.

the point of this idea is that instead of marking hundreds of identical files as junk just because they have a different name.... and then new ones continue to show up and still need to be marked as junk.... you simply point out to the system -why- said file is junk. then the system can scan any results you get back and automatically junk anything that turns up.

a virus-scanner solution will not fix this. the problem originates at the source, and the source has no interest in fixing it -- the source doesn't even consider it broken. if you put a forced virus-scan limitation into it, then you're going to get someone who writes a hack for it because they don't want to use anti-virus... or wore because they are one of the ones deliberately distributing them.


me personally, i don't consider 99% of the junk results i get to even be potential viruses, simply because most of them are not executable files.
go ahead and paste a virus into an mp3 file; you'll get a bit of static, or a corrupted file, or the player will just ignore it. the players don't execute the code, they just read it.

to make an exxagerated comparison.. it would be like taking a sheet of player-piano music, and drawing the plans for a bomb on them. the player piano will just read the sheet and play the piano music automatically. and when it gets to the bomb, it just keeps on playing and can't recognize it. maybe you've damaged some of the holes the piano is reading so it screws a few notes up.

if you took the sheet out of the piano and processed it differently, sure, you could make a dangerous bomb off it. but the automated piano isn't capable of doing it, and neither are most modern media players*

*to note; make sure you turn off "automatically open links in media" wherever possible.. once the player open some random website, all guarantees are off. but to-date, i've only ever noticed this in .wmv files, and i learend a long time ago to filter out *.wmv with the existing junk filters.

the only way i could see something like this happening is if you use a checksum on programs that contain the viruses and block the ones with that checksum. But then still you can download a RAR file that is a working game and it could still contain a virus. So the only way to really do this would be through checksums.

but again, what compares the checksums?

if it's on the sender's computer, they'll just write a hack to get around it, and then people will be less careful because they assume 'the network is protecting us'.

if it's on the receivers' computer, we need to store a checksum for.. well.. every file limewire/gnutella has ever seen. will limewire be shipping with a new HDD to store that on now? "new limewire, better compression, only a 5gig dowload". and thats still the same as it being on the senders' computer anyways. plus what keep the list itself safe from hacking thru a virus obtained elsewhere?

and if we put it on the limewire central server -- oh wait, there still isn't one.

i hope i'm not coming off condescending here, i'm certainly not trying to. it would be a perfectly acceptable method of protection in a more limited, client/server type setup. like an MMO style game sending patches to the clients. in a peer-to-peer system, where thousands of new files a day might be seen, theres just no way to keep up.

every time limewiere started up it would have to scan & compare every file against its database, to make sure it hasn't changed. that might take hours if you're sharing say 1000 files, and need to scan, hash, then compare against the 100,000,000 files in the database.

if you edit an mp3 to clip 2-3 seconds of silence off at the end, guess what, it's a "new file".

and then if it finds say 1 or 2 files you added that aren't catalogued what does it do? send a report and ban you till they get processed?
or risk lettng you onto the network to share these potentially dangerous viruses? (even tho it's just your shopping list in a txt file, that you saved in the wrong directory).

and then on top of that.. who catalogues all these tens of hundreds of thousands of files? by my guessing, you'd need a handful of people working full time jobs, just to keep the database up to date.


and then from there, you still need to get that database out to the users. every limewire user downloading an update every time they start limewire that might potentially be larger than the installer itself.

coz sooner or later someone's going to accidentally shar their c:\windows\ folder, and 50- to 100,000 files go on the pile in one dump.

if theres no center server then what is goin to keep up with the criterias of the junk filters? The program it self? couldnt their possibly be a hack made for that?

A good way to keep up with checksums would be to make a file attached with the download that only limewire reads with a checksums everytime its searched if this checksums = the checksum of a virus mark as junk or completely stop shareing

If limewire checks the checksum and it doesnt add up dont allow it to share ( to prevent people from tampering with it)

I anit saying someone would have to keep up with all the checksums of the files loaded just the checksums of viruses (this would reduce work load)

But still yes their would probally be ways to hack this but THERE IS A WAY TO "HACK" (get around) any kind of security measure we can take espically if its programed into the program itself.

O ya what keeps up with what files are junk the individual programs on every computer or a server?

If kept on a server then the checksum thing would be more feasible because limewire is also haveing problems with bungholes marking LEGIT perfectly good files as junk so if this was controlled by a central server monitoring the files for viruses or empty files Screw wrong named fie thats what bitiz ticket is for if you don't use it before u download something your bound to run into trouble sooner or later?

What would stop these bungholes from MASS JUNKING perfectly good files? Wait maybe not mass junking but AUTO JUNKING good filesif theres no center server then what is goin to keep up with the criterias of the junk filters? The program it self? couldnt their possibly be a hack made for that?

A good way to keep up with checksums would be to make a file attached with the download that only limewire reads with a checksums everytime its searched if this checksums = the checksum of a virus mark as junk or completely stop shareing

If limewire checks the checksum and it doesnt add up dont allow it to share ( to prevent people from tampering with it)

I anit saying someone would have to keep up with all the checksums of the files loaded just the checksums of viruses (this would reduce work load)

But still yes their would probally be ways to hack this but THERE IS A WAY TO "HACK" (get around) any kind of security measure we can take espically if its programed into the program itself.

O ya what keeps up with what files are junk the individual programs on every computer or a server?

If kept on a server then the checksum thing would be more feasible because limewire is also haveing problems with bungholes marking LEGIT perfectly good files as junk so if this was controlled by a central server monitoring the files for viruses or empty files Screw wrong named fie thats what bitiz ticket is for if you don't use it before u download something your bound to run into trouble sooner or later?

What would stop these bungholes from MASS JUNKING perfectly good files? Wait maybe not mass junking but AUTO JUNKING good files

*edit*

O ya what what would reasons you could type in this box?

Virus: Will not work program doesnt have virus scan to scan all the files
Size: A lot of files have the sames size but arent all junk
Program not working right: How would filter be able to determine this?
wrong name: filter cant detect this
name: some files have same names but are the real deal

If there are anymore pretty sure filter wouldnt be able to real determine if the file is the real thing or not

About the only way i can see that a filter would be able to determine if something is the real thing is through a checksum seeing as how that is what a checksum is for since a computer doesnt know what the file is suppose to do it just understands what the file is telling it to do and doesnt think like humans do

Another thing about this on top of auto junking IF this reason based filter could work think how much of a battle it would become to fight autojunking we have enough problems dealing with it already!

limewire doesn't work around a central server. it never has. that's why it's "peer to peer" filesharing. "peers" being the end user computers, that is,everyone on the network. client/server models don't work in filesharing, because if anyone uploads copyrighted or other unacceptable data, the copyright holders come n and have that server shut down.

as for what keeps up with the junk filters; the destination computer handles it. same as it does now. just with better handling. and sure you could make a hack to get around it.. but why would you hack around your own chosen filters? we aren't asking for network-wide systems to be implemented, just a more successful endpoint system.

the problem with checksums is that all they can do is verify that the file is the same as the one that is originally sent. yes if a virus is checksummed, and that checksum is recordd and blocked, it eliminates that virus. except that if you take the file and add a small block of random data to the end, the checksum changes. meaning the same virus in a slightly different wrapper keeps going.

basically the direction your checksum idea is moving in is that limewire should have a built in anti-virus, and while i don't think it would be so bad, that's a lot more complicated than i would expect the devs to have time for.

we're asking for improvements on the existing junk-filter system; a simple interface to assign more detailed rules, to better keep out what we determine is crap.

and the same as the current system, it doesn't matter what i tell my computer is junk; it's my computer handling it, and has no effect on what your computer sees as junk or 'good'. mass-junking, auto-junking.. whatever. it only limits that user's possible results.

bitiz ticket? i completely couldn't follow this part. is this a program i haven't heard of?

as for the last block;
viruses; no you can't filter that, you're right. the dev's don't have the time, i'm sure, to add a completely new and unrelated (to any existing programming) subsystem not to mention keeping it up to date and such.

size; in 20 years of working with computers, i've found that very few files actually have the exact same size without being the same file. if i were to point out "that 3,462KB file" in a room full of people dealing with this problem, i bet most people would almost instantly know the file i'm referring to.

not working right; well how can any filter determine this? it only works by actually trying to use it. but it might still be a perfectly legit program.

wrong name; again same thing. i downloaded "german techno - track 10" from somewhere once. turned out to be Eminem. but it was still a valid file.


i suggested things like..
-parts- of a name. like when spammers echoyour results back to you and append a specific phrase to it like "cute girl has shaking orgasm on webcam" -- being able to tell it "this is junk because of this part of the filename".

-exact- filesize match. 3,462KB shows up too often, and is never legit. 177kb. 399kb. etc. the files that keep showing back up with your search results echoed as the filename. currently telling the system it's junk only blocks that name, so it shows up again with a different name on your very next search.

....
actually y'know honestly these are the only 2 filters i want added. we can already filter by filetype so all those .wmv, .vbs, .bat, .pif, .scr, etc.. are already blocked.

and between filetype and filesize, i imagine we could catch 80% of the junk out there. 95% when you add in the "catchphrases". the other 5% ? well hell, nothing's perfect, and i'm prepared to handle those one-on-one.


oh and ... no. checksums are for determinng if the file you received is the same as the file was sent. if the file has a virus attached? the checksum will still be perfectly valid. it'll still say "yes this file transferred correctly". they aren't just clipping viruses on randomly before sending files. and if one file is blocked, they'll just switch it up.

as for auto-junking? uh.. what?
the junk filters exist on your system, and your system only. if you are having a problem with it, then you've mis-junked something. junking something doesn't send any reporting to the network.

The junk filter is useless. It can not recognize the same exact file. On the other hand, if you accidentally download the file and do a new search, Limewire recognizes the file as having been downloaded. Therefore, the junk filter does not work.

As you sound far more knowledgeable about computers than myself, I would like to ask what seems to me a very simple question. If Limewire can recognize the same file when it comes to placing a check mark or other icon indicating that I have downloaded it and still follow that same file every time it comes up in a search, why can't the junk file do the same? Everything I mark as junk comes back up just fine again and again.