> I know from reading the list that some of you have set up > your own automated systems that notify your users of held > messages and allow them to requeue them automatically.
I think you mean "user based" and not automated reviev and requeuing. That's what I've set up in the past month's 1.) weighting system up to 80 points: legit => delivery 80 to 119: spam low => subject marking & delivery 120 to 149: spam mid => subject marking & delivery 150 to 199: review => hold 200 and up: spam => hold 2.) Log file processing with MDLP (Modular Declude Logfile Processor) written by Pete McNeil This tool does extremely fast parsing of declude jm logfiles. Pete's primary intention was to write a tool that's able to analyze results of each declude test and based on the determined reliability adapt automatically the weighting system. Due to a lot of other work I haan't had time to test this part of the tool. I concentrated on the other part of MDLP. It can write CSV-files containing all processed messages, mailfrom, mailto, datetime, subject and the total weight of the weighting system. Then I've setup up some MS-SQL DTS packages that are able to import this csv-sources in a very fast way into a MS-SQL database. MDLP allows to process only e certain timerange of a daily logfile, so we import the processed message results on a hourly base. 3.) message records are keept based on their final weight. Messages below 80 points and above 200 points are deleted after max. one day. Message records in the revie range are hold for 14 days. Another script is cleaning up the hold folder for hold spam above 200 points and all hold spam older then 14 days. 4.) two times each day (07:00 am an 05:00 pm) a script is queriing all messages having a final weight in the review range. This script looks for each recipient of this messages in a separate table (notify) if the recipient has already received a notify. If not, the recipient receives a short notify with a short explanation of what notifies are. The notify contains a link to a IIS ASP SSL protected multilanguage webseite. The user can assign his own password and login to see a list of all his messages classified in the review range. (datetime, mailfrom and subject line) 5.) The user can click on a message to deliver the message immediately or he can click on a button on the end of the list to mark all remained messages as spam and so quickly clean up his review box. 6.) If a recipient has received a notify further notifies are send out only if there are new messages in the review range AND the user has already visited his review box OR the last notify was send out over 7 days ago AND The user has not choosen in his review box to not more receive review notifies 7.) is not 100% ready but under developepement Each user can choose in the reviev box his own spam filter "risk level" By selecting one of 5 levels this will add a line in a declude filter file that will add or substract some points (based on the choosen risk level) to each message send to his address and so influence the final weight (instead of moving up and down hold ranges, add or substract some points) 8.) The script creates also hourly reports on processed messages (legit, low, mid, review, spam, virus & vulnerability) The graphical visualization of this values is one of the next things on my todo list. All this functionality is part of several scripts and a long work of configuration and setup. My weighting system is able to classify only around 3% of all messages in the reviev range, so users don't have to check the review box too often. Users can click on a "forgotten your password" so the helpdesk has not more work with this user frontend. This system is working now for 14 days here and it seems reaching a stable level even if I'm working each day one some little errors, feature requests or improvements. The problem will be to find some time to bring all this "review suite" in a publicable form... Markus --- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)] --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.