it would be simple for limewire engineering to calculate a mathematically unique number for any file (based on contents only, excluding name and datetime), using a publically available algorithm known as Message Digest version 5 (MD5), so they could easily add a feature to block a specific file.

problem is that a spammer could simply auto-add a character to a spam file to change its MD5 signature. still, with so many users, it would only take a few flags to mark a file as spam, sex content, violent, stupid, or whatever categories they wanted to offer. the block level for each category would be directly related to the number of flags received from unique user logins, so each user would have control over block sensitivity.

blocking a host is what i do now, because most of the time, it is a spammer ip address. the problems with that approach are that spammers manufacture many addresses, so its not fully-effective. also, any legit user that downloads the spam file is also blocked, and we cannot share files if i block the ip address. perhaps the number of host blocks could be centralized in a scheme similar to the one proposed above, with sensitivity under control of the user.

spammers know that small files tend to get downloaded quickly and get their messages through, so they tend to deliver small size files, so we can at least sort on size to find suspect files before selecting what to download. this is more true of audio file or video with a single image than full ads in video format. also, there are exceptions, such as the spam that echos your text search, of size 3462 K.

for now, i'd simply be happy with a file size blocker to nip spam. even if it means blocking other legit files of the same size. at least i could block 23.7, 277, and 3462 globally.

i'm new to this forum, so wonder if our discussion has any impact on what lw engineering does. perhaps this is just an exercise in futility to apease the masses -- a place to vent frustration.

<insert smiley-bang-head-on-wall icon here>
alphakid42