The TREC spam filter submission page is now ready:

For "run identification" please enter a string constructed
using the following convention:


        ggg is a group identifier
        t   is the task (S = spam filter A = active)
        n   is the priority of the submission (1-4)
        F   means Filter        (filter submission)

The spam filter submission itself should be a single
.tgz or .zip file containing "initialize" "train" "classify"
and "finalize" as described in the guidelines.  The active
filter submission should be a .tgz or .zip file containing

NIST have created a Perl script that you may use to check
that your .zip or .tgz file includes these files; see



Later, there will be a similar "spam results" submission
page.  For that page please use a "run identification"
following this convention:


        gggtn is the same as for the filter used to generate
                   the result
        ddd identifies the public corpus to which the filter
                   was applied; one of:

            pci   (public chinese immediate feedback)
            pcd   (public chinese delayed feedback)
            pei   (public english immediate feedback)
            ped   (public english delayed feedback)