label
class label 2
classes | group
stringclasses 5
values | text
stringlengths 3
205k
|
|---|---|---|
1ham
| easy_ham | "Or, you could do this. <G>\n\nOn Mon, 9 Sep 2002, David Rees wrote:\n\n> On Mon, Sep 09, 2002 at 05:10:23PM -0400, Rob wrote:\n> > I just set up razor and spamassassin, but I keep getting this error in my \n> > mail log file\n> > \n> > razor2 check skipped: No such file or directory Can't call method \"log\" on \n> > unblessed reference at \n> > /usr/local/lib/perl5/site_perl/5.8.0/Razor2/Client/Agent.pm line 212.\n> > \n> > I have looked through the archived list and the only thing I have seen \n> > about this error is a possible permission problem on the log file.\n> \n> Yeah, the answer isn't here in this list, it's on the spamassassin-users\n> list. I just asked it late last week.\n> \n> Hint: Add -H to the spamd startup flags using the latest version of\n> spamassassin.\n> \n> -Dave\n> \n> \n> -------------------------------------------------------\n> This sf.net email is sponsored by: OSDN - Tired of that same old\n> cell phone? Get a new here for FREE!\n> https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n> _______________________________________________\n> Razor-users mailing list\n> Razor-users@lists.sourceforge.net\n> https://lists.sourceforge.net/lists/listinfo/razor-users\n> \n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by: OSDN - Tired of that same old\ncell phone? Get a new here for FREE!\nhttps://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "It's a permissions issue on the razor log file. SpamAssassin runs setuid \nto the user to whom mail is being delivered.\n\nYou'll need to make the razor log file world writable.\n\nOn Mon, 9 Sep 2002, Rob wrote:\n\n> I just set up razor and spamassassin, but I keep getting this error in my \n> mail log file\n> \n> razor2 check skipped: No such file or directory Can't call method \"log\" on \n> unblessed reference at \n> /usr/local/lib/perl5/site_perl/5.8.0/Razor2/Client/Agent.pm line 212.\n> \n> I have looked through the archived list and the only thing I have seen \n> about this error is a possible permission problem on the log file.\n> I did what it said in the archives, basically change the permission on the \n> file but its still no go.\n> \n> Any other help would be appreciated, maybe I'm missing something.\n> something I forgot to run or do.\n> \n> \n> rob\n> \n> \n> \n> \n> -------------------------------------------------------\n> This sf.net email is sponsored by: OSDN - Tired of that same old\n> cell phone? Get a new here for FREE!\n> https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n> _______________________________________________\n> Razor-users mailing list\n> Razor-users@lists.sourceforge.net\n> https://lists.sourceforge.net/lists/listinfo/razor-users\n> \n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by: OSDN - Tired of that same old\ncell phone? Get a new here for FREE!\nhttps://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Does anyone else experience this:\n\nhttp://sf.net/tracker/index.php?func=detail&aid=600311&group_id=3978&atid=10\n3978\n\nAlso, it seems that the bugs tracker on SF isn't used very much ... is there\nsomewhere else to post bugs?\n\n- Colin\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by: OSDN - Tired of that same old\ncell phone? Get a new here for FREE!\nhttps://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I have a spamd/spamc/razor/dcc setup.\n\nMy razor logs are full of:\n\nSep 10 09:48:48.411339 check[10200]: [ 1] [bootup] Logging initiated \nLogDebugLevel=3 to file:/home/omega13/.razor/razor-agent.log\nSep 10 09:48:49.391309 check[10200]: [ 3] mail 1 is not known spam.\n\nThe problem is that all the entries show \"is not known spam\" I can't believe \nI'm that lucky. So, how do I test to see that my razor installation is \nworking properly. I've been up to my ears in documentation for the past few \ndays, so, if the reference is easy, just point me to it. Otherwise a brief \nhint on how to, would be appreciated.\n-- \nRobin Lynn Frank\nParadigm-Omega, LLC\n==================================\nGandalf fell at Khazadum.\nSheridan fell at Z'ha'dum.\nAvoid high places in \nlocations ending in \"dum\"\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by: OSDN - Tired of that same old\ncell phone? Get a new here for FREE!\nhttps://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "On Tue, 10 Sep 2002, Rose, Bobby wrote:\n\n> Use the sample-spam.txt from Spamassassin and do a \"razor-check -d <\n> sample-spam.txt\"\n\nInteresting. I just upgraded to razor-agents 2.14 yesterday, ran\nrazor-admin -register. I to am seeing that I'm not getting any positives\nfrom razor. Including the sample-spam.txt's from spamassassin 2.41, 2.31,\nand 2.20.\n\nAny suggestions? thanks.\n\n-d output below\n\n\n\nmaxwell:/var/qmail/alias/gb-users%# razor-check -d <\n/usr/local/src/Mail-SpamAssassin-2.20/sample-spam.txt\n Razor-Log: Computed razorhome from env: /root/.razor\n Razor-Log: Found razorhome: /root/.razor\n Razor-Log: No /root/.razor/razor-agent.conf found, skipping.\n Razor-Log: No razor-agent.conf found, using defaults. \nSep 12 12:04:13.785852 check[52238]: [ 1] [bootup] Logging initiated\nLogDebugLevel=9 to stdout\nSep 12 12:04:13.787113 check[52238]: [ 5] computed razorhome=/root/.razor,\nconf=, ident=/root/.razor/identity-ruqVa5jbuS\nSep 12 12:04:13.787442 check[52238]: [ 2] Razor-Agents v2.14 starting\nrazor-check -d\nSep 12 12:04:13.789903 check[52238]: [ 9] uname -a: FreeBSD\nmaxwell.gta.com 4.5-RELEASE-p3 FreeBSD 4.5-RELEASE-p3 #0: Wed May 22\n14:52:29 EDT 2002 root@maxwell.gta.com:/usr/src/sys/compile/maxwell\ni386\nSep 12 12:04:13.790480 check[52238]: [ 8] reading straight RFC822 mail\nfrom <stdin>\nSep 12 12:04:13.791397 check[52238]: [ 6] read 1 mail\nSep 12 12:04:13.791917 check[52238]: [ 8] Client supported_engines: 1 2 3\n4\nSep 12 12:04:13.792948 check[52238]: [ 8] prep_mail done: mail 1\nheaders=1432, mime0=3139\nSep 12 12:04:13.793639 check[52238]: [ 6] skipping whitelist file\n(empty?): /root/.razor/razor-whitelist\nSep 12 12:04:13.794295 check[52238]: [ 5] read_file: 1 items read from\n/root/.razor/servers.discovery.lst\nSep 12 12:04:13.794871 check[52238]: [ 5] read_file: 1 items read from\n/root/.razor/servers.nomination.lst\nSep 12 12:04:13.795382 check[52238]: [ 5] read_file: 3 items read from\n/root/.razor/servers.catalogue.lst\nSep 12 12:04:13.796047 check[52238]: [ 9] Assigning defaults to\nhonor.cloudmark.com\nSep 12 12:04:13.796456 check[52238]: [ 9] Assigning defaults to\napt.cloudmark.com\nSep 12 12:04:13.796829 check[52238]: [ 9] Assigning defaults to\nfire.cloudmark.com\nSep 12 12:04:13.797180 check[52238]: [ 9] Assigning defaults to\ntruth.cloudmark.com\nSep 12 12:04:13.798322 check[52238]: [ 5] read_file: 11 items read from\n/root/.razor/server.apt.cloudmark.com.conf\nSep 12 12:04:13.799188 check[52238]: [ 5] read_file: 11 items read from\n/root/.razor/server.honor.cloudmark.com.conf\nSep 12 12:04:13.799866 check[52238]: [ 5] 96778 seconds before closest\nserver discovery\nSep 12 12:04:13.800342 check[52238]: [ 6] apt.cloudmark.com is a Catalogue\nServer srl 51; computed min_cf=1, Server se: 0A\nSep 12 12:04:13.800818 check[52238]: [ 8] Computed supported_engines: 2 4\nSep 12 12:04:13.801167 check[52238]: [ 8] Using next closest server\napt.cloudmark.com:2703, cached info srl 51\nSep 12 12:04:13.801513 check[52238]: [ 8] mail 1 Subject: Home Based\nBusiness for Grownups\nSep 12 12:04:13.805467 check[52238]: [ 6] preproc: mail 1.0 went from 3139\nbytes to 3100 \nSep 12 12:04:13.805878 check[52238]: [ 6] computing sigs for mail 1.0, len\n3100\nSep 12 12:04:13.809475 check[52238]: [ 5] Connecting to apt.cloudmark.com\n...\nSep 12 12:04:18.587441 check[52238]: [ 8] Connection established\nSep 12 12:04:18.587929 check[52238]: [ 4] apt.cloudmark.com >> 29 server\ngreeting: sn=C&srl=51&ep4=7542-10&a=l\nSep 12 12:04:18.588562 check[52238]: [ 6] apt.cloudmark.com is a Catalogue\nServer srl 51; computed min_cf=1, Server se: 0A\nSep 12 12:04:18.589041 check[52238]: [ 8] Computed supported_engines: 2 4\nSep 12 12:04:18.589432 check[52238]: [ 8] mail 1.0 e2\nsig: PGFfFte87P3Ve-CPTdu3NWgiBikA\nSep 12 12:04:18.589751 check[52238]: [ 8] mail 1.0 e4\nsig: k6oGZsa1AvVolyvalWx2AACdWb8A\nSep 12 12:04:18.590103 check[52238]: [ 8] preparing 2 queries\nSep 12 12:04:18.590638 check[52238]: [ 8] sending 1 batches\nSep 12 12:04:18.591087 check[52238]: [ 4] apt.cloudmark.com << 96\nSep 12 12:04:18.591324 check[52238]: [ 6]\n-a=c&e=2&s=PGFfFte87P3Ve-CPTdu3NWgiBikA\na=c&e=4&ep4=7542-10&s=k6oGZsa1AvVolyvalWx2AACdWb8A\n.\nSep 12 12:04:19.270870 check[52238]: [ 4] apt.cloudmark.com >> 14\nSep 12 12:04:19.271227 check[52238]: [ 6] response to sent.1\n-p=0\np=0\n.\nSep 12 12:04:19.272130 check[52238]: [ 6] mail 1.0 e=2\nsig=PGFfFte87P3Ve-CPTdu3NWgiBikA: sig not found.\nSep 12 12:04:19.272449 check[52238]: [ 6] mail 1.0 e=4\nsig=k6oGZsa1AvVolyvalWx2AACdWb8A: sig not found.\nSep 12 12:04:19.272760 check[52238]: [ 7] method 5: mail\n1.0: no-contention part, spam=0\nSep 12 12:04:19.273012 check[52238]: [ 7] method 5: mail 1: a\nnon-contention part not spam, mail not spam\nSep 12 12:04:19.273257 check[52238]: [ 3] mail 1 is not known spam.\nSep 12 12:04:19.273578 check[52238]: [ 5] disconnecting from server\napt.cloudmark.com\nSep 12 12:04:19.273982 check[52238]: [ 4] apt.cloudmark.com << 5\nSep 12 12:04:19.274203 check[52238]: [ 6] a=q\nSep 12 12:04:19.274585 check[52238]: [ 8] razor-check finished\nsuccessfully.\n\n\n\n\n\n\n\n---\ndavid raistrick\ndrais@atlasta.net\t\thttp://www.expita.com/nomime.html\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I'm taking all my razored mail today and calling any 1-800 numbers I can\nfind. I say \"Hi, I'm calling everyone that spammed me today to raise their\ncost of doing business. You have a great day!\" I'm told I should do it\nfrom a pay phone, so it costs them 50 cents per call. If more people did\nthis, we could tie up their phone lines and cost them so much money they\ncould never make spam work.\n\nAlso fun is the simple \"Die Spammer Die!\"\n\nFox\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Hi,\n\nIs it possible to use razor without filtering empty mails as spamm?\nAn mail with an attachment is considered spamm. Is this normal?\nOr I mysqlf like to send emails to myself where all important is said in\nsubject and body is empty (for remaining smth).\n\nThanks,\nRaido Kurel\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "> >Tom R:\n>>> \n>>><http://www.cliktrik.com/people/family/me/0419.jpg>http://www.cliktrik.com/people/family/me/0419.jpg\n>>>\n>>Which one's you?\n>\n>I'm actually taking the photo -- both figures are in fact waxworks.\n>\n>This was in Mme Tussaud's in, of all places, Sydney Australia.\n>\n> /t\n>--\n\ndamn it Tom!! I had my kids believing you knew Albert Einstein!!\n\nWell, until the smart one asked just how old you were now.\n-- \n\n\nFel\nNEW!! Cafe Forteana is back: http://www.frogstone.net/Cafe/CafeForteana.html\nhttp://www.frogstone.net\nWeird Page: http://my.athenet.net/~felinda/WeirdPage.html\n\n[Non-text portions of this message have been removed]\n\n\n------------------------ Yahoo! Groups Sponsor ---------------------~-->\nPlan to Sell a Home?\nhttp://us.click.yahoo.com/J2SnNA/y.lEAA/MVfIAA/7gSolB/TM\n---------------------------------------------------------------------~->\n\nTo unsubscribe from this group, send an email to:\nforteana-unsubscribe@egroups.com\n\n \n\nYour use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ \n\n\n\n" |
1ham
| easy_ham | "Razor won't filter empty emails. What version of the agents are you using?\n\ncheers,\nvipul.\n\nOn Mon, Sep 16, 2002 at 04:16:31PM +0300, Raido Kurel wrote:\n> Hi,\n> \n> Is it possible to use razor without filtering empty mails as spamm?\n> An mail with an attachment is considered spamm. Is this normal?\n> Or I mysqlf like to send emails to myself where all important is said in\n> subject and body is empty (for remaining smth).\n> \n> Thanks,\n> Raido Kurel\n> \n> \n> \n> \n> -------------------------------------------------------\n> This sf.net email is sponsored by:ThinkGeek\n> Welcome to geek heaven.\n> http://thinkgeek.com/sf\n> _______________________________________________\n> Razor-users mailing list\n> Razor-users@lists.sourceforge.net\n> https://lists.sourceforge.net/lists/listinfo/razor-users\n\n-- \n\nVipul Ved Prakash | \"The future is here, it's just not \nSoftware Design Artist | widely distributed.\"\nhttp://vipul.net/ | -- William Gibson\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | ["Gary Funck <gary@intrepid.com> [2002-09-18 13:57:00 -0700]:\n> In my experience, there are spam messages that sneak past Spam Assassin,\n> that Razor will pick up. Those are the ones that I'm calling \"marginal\".\n> Basically, I'm hoping that \"the collective\" of Razor users make a better\n> judge of spam than any single program like SA can, and therefore I can\n> benefit from their judgement and get more extensive spam filtering. I've\n> seen examples of this already, where SA doesn't score the spam high enough\n> to bounce it, but Razor does.\n\nI think perhaps you missed the fact that SA scores are adjustable. If\nyou want SA to tag all messages listed in Razor then you can put this\nin your ~/.spamassassin/user_prefs file.\n\n score RAZOR_CHECK 10\n\nThe default score is 3 and the default threshold needed is 5.\nTherefore if you wish to have any razor listed messages tagged by SA\nthen setting a score for any razor listed messages to anything above 5\nwould be sufficient.\n\nIf you are already using SA then the above would be more efficient.\nOtherwise you are running all of the mail through razor twice, once\nfor SA and once again afterward. If you really want to run Razor\nindividually then you should set the 'score RAZOR_CHECK 0' so that SA\nwon't do it and avoid the double network hit.\n\nHowever, one of the benefits of using SA in combination with Razor has\nbeen the history of false positive reports in the razor database. The\ncurrent score of 3 is hefty, but not enough by itself to tag as spam.\nBut for any real spam is usually enough to push it over the threshold.\nRazor2 addresses the false positive problem but is not yet in as wide\nof use as Razor1.\n\nBob\n", "…"] |
1ham
| easy_ham | "> I suggest that you use a normal user to start. Once you are familiar\n> enough with razor and understand how it works, try the complex things\n> like integration with the MTA and the like.\n\nOK, but do I have to launch that command as \"root\" or what?\n\nThank you\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Using -lm 4 is yielding an extra 20% a day, but it gets false positives\nwhere it shouldn't.\n\nSuch as an email with a Word doc and the signature below. After looking at\nthe Word doc, directions to the sender's cabin, I am convinced it marked the\nbody, which contains no next except the \"IncrediMail advertisment\"\nsignature, as spam.\n\nSo I have to turn off -lm 4. Razor has been getting other strange emails it\nshouldn't with -lm 4 on.\n\n See the incredimail ad signature I am talking about below:\n\n Fox\n\n\n\n____________________________________________________\n IncrediMail - Email has finally evolved - Click Here\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "\nHere is my recipe for Maildrop:\n\n# Vipul's Razor check\n#\nlog `/usr/bin/razor-check`\nlog \"razor-check = $RETURNCODE\"\nif ( $RETURNCODE == 0 )\n{\n xfilter \"reformail -a'X-Razor: SPAM'\"\n log \"SPAM caught by Vipul's Razor\"\n to \"$FOLDERS/.SPAM/\"\n}\n\nI used this with Razor version 1. I'm not sure if it was ever used\nwith Razor2. I am now using SpamAssassin w/ Razor2. You may wish to\nremove the log statements once running.\n\nI use both. SpamAssassin (SA) catches things that Razor does not. There\nwere enough false positives with Razor that I do not trust it alone.\nAfter whitelisting all the newsletters I receive, SA w/ Razor2 does a\nvery good job.\n\nHTH,\n Jeffrey\n\n\nQuoting Sunil William Savkar <sunil@inthespace.com>:\n> Hi.\n>\n> I just finished installing and getting running maildrop with my virtual\n> users. I was thinking to set up razor with maildrop as my first test of\n> its filtering capabilities..\n>\n> I have seen documentation out there for procmail, but is there similar\n> documentation for integrating maildrop with razor?\n> \n> \n> \n> Separately, it looks like many people use spamassassin with razor. If I\n> am using razor2, is there still an advantage to this?\n> \n> \n> \n> Thanks in advance.\n> \n> \n> \n> Sunil\n> \n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Jeff--\n\nWhat is the maildrop recipe you use with spamassassin? I was trying to\nset mine up but I am running into a few difficulties.\n\nI think I will try spamassassin first and see how it goes, and then\nperhaps fold in razor2... is it easy to fold in?\n\nSunil\n\n>\n> Here is my recipe for Maildrop:\n>\n> # Vipul's Razor check\n> #\n> log `/usr/bin/razor-check`\n> log \"razor-check = $RETURNCODE\"\n> if ( $RETURNCODE == 0 )\n> {\n> xfilter \"reformail -a'X-Razor: SPAM'\"\n> log \"SPAM caught by Vipul's Razor\"\n> to \"$FOLDERS/.SPAM/\"\n> }\n>\n> I used this with Razor version 1. I'm not sure if it was ever used with\n> Razor2. I am now using SpamAssassin w/ Razor2. You may wish to remove\n> the log statements once running.\n>\n> I use both. SpamAssassin (SA) catches things that Razor does not.\n> There were enough false positives with Razor that I do not trust it\n> alone. After whitelisting all the newsletters I receive, SA w/ Razor2\n> does a very good job.\n>\n> HTH,\n> Jeffrey\n>\n>\n> Quoting Sunil William Savkar <sunil@inthespace.com>:\n>> Hi.\n>>\n>> I just finished installing and getting running maildrop with my\n>> virtual users. I was thinking to set up razor with maildrop as my\n>> first test of its filtering capabilities..\n>>\n>> I have seen documentation out there for procmail, but is there similar\n>> documentation for integrating maildrop with razor?\n>>\n>>\n>>\n>> Separately, it looks like many people use spamassassin with razor. If\n>> I am using razor2, is there still an advantage to this?\n>>\n>>\n>>\n>> Thanks in advance.\n>>\n>>\n>>\n>> Sunil\n>>\n>\n>\n> -------------------------------------------------------\n> This sf.net email is sponsored by:ThinkGeek\n> Welcome to geek heaven.\n> http://thinkgeek.com/sf\n> _______________________________________________\n> Razor-users mailing list\n> Razor-users@lists.sourceforge.net\n> https://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I received a spam email that had apparently forged the From header with\nmy own email address. After reviewing the message I forwarded it with\nthe rest of a batch of spam ti the database with razor-report. Now of,\ncourse, my own email address is listed int he Razor database. How do I\ngo about getting it removed.\n\nI have seen this ploy (forged From headers) several time since then.\nPerhaps the razor-report need to detect this and emit a warning.\n\ndave\n\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | ">That Goddess Chick wrote:\n>>\n>> >Thanks Fel. Got no scanner. My photo is in that group of 100 obsessive\n>> >compulsive clipsters in FT, 1996 or 1997.\n>> >\n>> >Terry\n>>\n>> Great, and right now all my pre '98s are in Washington state, in a\n>> cardboard box in a shed in the back of Sydde's garage. Probably mice\n>> nests by now. :-( Put a scanner on your Christmas list right above\n>> world peace!\n>> --\n>>\n>> Fel\n>> NEW!! Cafe Forteana is back: \n>><http://www.frogstone.net/Cafe/CafeForteana.html>http://www.frogstone.net/Cafe/CafeForteana.html\n>\n>Maybe a kind soul with access to that issue and a scanner could scan it and\n>forward to you.\n>\n>Terry\n>\n\nI would appreciate that very much as I won't be getting back to \nWashington until December.\n-- \n\n\nFel\nNEW!! Cafe Forteana is back: http://www.frogstone.net/Cafe/CafeForteana.html\nhttp://www.frogstone.net\nWeird Page: http://my.athenet.net/~felinda/WeirdPage.html\n\n[Non-text portions of this message have been removed]\n\n\n------------------------ Yahoo! Groups Sponsor ---------------------~-->\nSell a Home with Ease!\nhttp://us.click.yahoo.com/SrPZMC/kTmEAA/MVfIAA/7gSolB/TM\n---------------------------------------------------------------------~->\n\nTo unsubscribe from this group, send an email to:\nforteana-unsubscribe@egroups.com\n\n \n\nYour use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ \n\n\n\n" |
1ham
| easy_ham | "I'm getting \"no servers available\" about half the time in the last few days.\nThis is with Razor 2. Is there something I need adjust in the installation\nhere, or are the servers just down/overloaded?\n\nThanks,\nWhit\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I just installed razor 2.152 on a FreeBSD 4.4-RELEASE box and having\nproblems with razor-check. Any time razor-check is run, (with or without\narguments), i get this error:\n\nCan't use an undefined value as a symbol reference at\n/usr/local/lib/perl5/site_perl/5.005/i386-freebsd/Razor2/Client/Agent.pm\nline 756.\n\nrazor-admin runs just fine and the make test before the install was all\nok.\n\nHas anyone seen this before?\n\nModule versions:\nDigest-HMAC-1.01\nDigest-MD5-2.20\nDigest-Nilsimsa-0.06\nDigest-SHA1-2.01\nMIME-Base64-2.12\nNet-DNS-0.23\nTest-Simple-0.44\nTime-HiRes-01.20\nURI-1.19\n\n\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "> Date: Mon, 30 Sep 2002 14:00:12 -0400 (EDT)\n> From: Dayv Gastonguay <noghri@nauticom.net>\n>\n> I just installed razor 2.152 on a FreeBSD 4.4-RELEASE box and having\n> problems with razor-check. Any time razor-check is run, (with or without\n> arguments), i get this error:\n>\n> Can't use an undefined value as a symbol reference at\n> /usr/local/lib/perl5/site_perl/5.005/i386-freebsd/Razor2/Client/Agent.pm\n> line 756.\n>\n Try installing the latest Perl (at least 5.6.1) port on Freebsd and make\nsure you set the system to use perl from ports (i.e. in the\nports/lang/perl5/files directory run ./use.perl port. Reinstall the relevant\nperl modules needed by razor and try again.\n\nSven\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I noticed a drop in checks, and did some tests. If I move truth up in\nmy catalog list, a check comes back postive, but if I move fire up in\nthe list, then I don't get the positive check back. It's almost like\nit's not syncing up with Hubris which is the one my reports are getting\nsent to.\n\n-=Bobby\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "What does this mean? I set up procmailrc for a spamtrap but I'm getting\nan error. I also am reporting to pyzor and dcc and they aren't\nregistering an error. What's weird is that it works sometimes.\n\n.\nOct 02 23:46:11.470523 report[14051]: [ 4] honor.cloudmark.com >> 20\nOct 02 23:46:11.470805 report[14051]: [ 6] response to sent.3\n-res=1\nerr=230\n.\nOct 02 23:46:11.471825 report[14051]: [ 5] mail 1, orig_email, special\ncase eng 1: Server accept\ned report.\nOct 02 23:46:11.472228 report[14051]: [ 8] mail 1.0, eng 4: err 230 -\nserver wants mail\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "\nI recently brought up a new system - SuSe 7.3 and I am running\nspamassassin and procmail, which work fine.\n\nI tried bringing up razor (2.14) but I am not able to get the SDK to\ninstall and razor reports missing modules. They appear to be there. I\nfollowed the installation directions. There appears to be an endless list\nof errors. Any ideas on what to do or where to start?\n\nDoug\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "If you examine the log further, you'll see debug messages generated by the\n(content) reporting process that follows error 230.\n\nOn Wed, Oct 02, 2002 at 10:20:49PM -0700, Jordan Ritter wrote:\n> Error 230 occurs when you report a signature, but the server doesn't\n> know about the signature, so it wants the full content. It's\n> basically an optimization. \n> \n> Beyond that, I'm not sure how to interpret that output.. what version?\n> Vipul?\n> \n> On Wed, Oct 02, 2002 at 11:59:01PM -0400, Rose, Bobby wrote:\n> # What does this mean? I set up procmailrc for a spamtrap but I'm getting\n> # an error. I also am reporting to pyzor and dcc and they aren't\n> # registering an error. What's weird is that it works sometimes.\n> # \n> # .\n> # Oct 02 23:46:11.470523 report[14051]: [ 4] honor.cloudmark.com >> 20\n> # Oct 02 23:46:11.470805 report[14051]: [ 6] response to sent.3\n> # -res=1\n> # err=230\n> # .\n> # Oct 02 23:46:11.471825 report[14051]: [ 5] mail 1, orig_email, special\n> # case eng 1: Server accept\n> # ed report.\n> # Oct 02 23:46:11.472228 report[14051]: [ 8] mail 1.0, eng 4: err 230 -\n> # server wants mail\n\n\n\n\n-- \n\nVipul Ved Prakash | \"The future is here, it's just not \nSoftware Design Artist | widely distributed yet.\"\nhttp://vipul.net/ | -- William Gibson\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I'm still seeing razor reporting bomb out with a small variety of messages.\nWould reinitializing something here help, or does razor do that as needed\nanyway, and this just reflects trouble at the servers?\n\nAlso, since these are reports I'm making from within Mutt, it's annoying\nthat the error messages garble the X window a bit. If there's nothing\nconstructive to do about razor failing, can I at least turn off the failure\nmessages? One way or the other I'd rather not see them.\n\nWhit \n\nOn Wed, Sep 25, 2002 at 12:57:13PM -0400, Whit Blauvelt wrote:\n> I'm getting \"no servers available\" about half the time in the last few days.\n> This is with Razor 2. Is there something I need adjust in the installation\n> here, or are the servers just down/overloaded?\n> \n> Thanks,\n> Whit\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "trying to report spam [razor chooses hubris] I timeout on the connection\n(which seems to have gotten slower all morning) and receive the following\nerror message:\n\nrazor-report error: connect4: nextserver: discover1: Error reading socket\nconnect4: nextserver: discover1: Error reading socket\n\nI then try to run razor-admin -discover and receive the same error .....\nproblems with the servers today? only one discovery server?\n\nSven\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | ["For everyone's benefit/edification, regardless of Razor usage.\n\n--jordan\n\n-----Original Message-----\nFrom: InterNAP Backbone Status Reports [mailto:noc@internap.com]\nSent: Thursday, October 03, 2002 11:30 AM\nTo: InterNAP Backbone Status Reports\nSubject: [EVENT NOTIFICATION] UUnet North American Backbone Problems\n20021003@06:00 PDT [TIX106448] (fwd)\n\n\nHello,\n\nI have just spoken with an UUnet technician at 11:23 PDT and he was unable\nto provide me with any new information regarding the outage. He did\nconfirm that problems had spread beyond the east coast and are now seen\nnationwide through out their network. They are currently working with\ntheir hardware vendor to try and isolate the issue.\n\nA large number of Internap's connections to UUnet remain shutdown and\ntraffic is being routed over alternate providers. The scale of this\noutage is causing a large number of peering points with UUnet and other\nproviders to be flooded due to the large traffic shifts.\n\nIf you have problems reaching specific sites please send an email to\nnoc@internap.com containing a traceroute and source/destination IP\naddresses and we will investigate. If it is possible will will attempt to\nmove specific prefixes onto an alternate provider leaving an Internap PNAP\nif that will provide better performance.\n\nRegards,\nAndrew\n\n--------------------------------------------------------------------\nAndrew Dul Network Operations Center\nInterNAP Network Services E-Mail: noc@internap.com\nandrew@internap.com 206.256.9500 - 1.877.THE.INOC\n\nThe contents of this email message are confidential and proprietary.\n\n---------- Forwarded message ----------\nDate: Thu, 3 Oct 2002 08:51:26 -0700 (PDT)\nFrom: Andrew Dul <andrew@internap.com>\nSubject: [EVENT NOTIFICATION] UUnet North American Backbone Problems\n [TIX106448]\n\n\nUUnet is currently experiencing a large number of problems in their North\nAmerican backbone. Internap has been aware of these problems since approx\n06:00 PST. Most of the problems appear to be concentrated on the East\nCoast but we have reports of problems from other geographic regions.\n\nThis is being tracked under Internap ticket 106448 and UUnet master ticket\n651751.\n\nUUnet currently does not have any ETA for a fix for this event.\n\nIf you have issues reaching a specific site please send an email to\nnoc@internap.com with a traceroute showing the path which has a problem.\n\nInternap has shutdown peerings to UUnet is various cities to help reduce\nthe number of problems that customers will experience.\n\nRegards,\nAndrew\n\n--------------------------------------------------------------------\nAndrew Dul Network Operations Center\nInterNAP Network Services E-Mail: noc@internap.com\nandrew@internap.com 206.256.9500 - 1.877.THE.INOC\n\nThe contents of this email message are confidential and proprietary.\n\n", "…"] |
1ham
| easy_ham | "This is happening due to insufficient write access to the\n\"razor-agent.log\" file. I was getting the same error, but\nonly as a non-root user. As a quick workaround, you can do\n\"chmod go+w razor-agent.log\".\n\nIn Agent.pm, when then the Logger object is created, it\ndoesn't check whether the logfile is writable by the current\nuser. Then, when a write attempt is made, it bails out with\nthe \"unblessed reference\" error.\n\nHope that helps,\nMichael\n\n> I just noticed the following log entries in my syslog with the latest\n> Spamassassin CVS (set up using spamc/spamd) and razor-agents 2.14:\n>\n> Jul 26 17:30:09 timmy spamd[54928]: razor2 check skipped: No such file or\n> directory Can't call method \"log\" on unblessed reference at\n> /usr/local/lib/perl5/site_perl/5.6.1/Razor2/Client/Agent.pm line 211,\n> <STDIN> line 25.\n>\n> I saw this after checking if my upgrade from razor-agents 2.12 to 2.14 went\n> okay, but the problem is still there after downgrading back to 2.12. I\n> don't really know when this started happening, :-/\n>\n> Any ideas on the problem?\n>\n> - Robert\n\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n" |
1ham
| easy_ham | "-----BEGIN PGP SIGNED MESSAGE-----\nHash: SHA1\n\nde'i Thursday 29 August 2002 13:36 la Fox cusku di'e\n> The following was personal correspondence between two people. I can't\n>\n> fathom how Razor thinks it is spam:\n\nWas it sent in HTML? If so, and it had a background, the background may have \nbeen sent in a spam.\n\ncmeclax\n-----BEGIN PGP SIGNATURE-----\nVersion: GnuPG v1.0.7 (GNU/Linux)\n\niD8DBQE9bq9f3/k1hdmG9jMRAk4XAJ9CheEA+/hLIU9zTzfJbPyoPUm+XwCfXgZ1\ntg7Fn8JcG9Q13UlKVfaOJzk=\n=Mw8+\n-----END PGP SIGNATURE-----\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n" |
1ham
| easy_ham | "\nUsing Razor2 via SpamAssasin.\n\nSystem is Solaris 2.7, with qmail. Spamassassin run via user's procmail.\nAll users who use SA have run razor-register.\n\nRazor2 is failing, and I can't find anything in the limited docs or on \ngoogle on it,\nand I'm hoping someone can help.\n\nThe error (which doesn't prevent SA from working) is:\n\nOct 2 06:38:22 sancho2 qmail: 1033565902.186041 delivery 4588: success: \nrazor2_check_skipped:_Bad_file_number_Can't_locate_object_m\nethod_\"new\"_via_package_\"Razor2::Client::Agent\"_(perhaps_you_forgot_to_load_\"Razor2::Client::Agent\"?)_at_/usr/local/lib/perl5/site_p\nerl/5.6.1/Mail/SpamAssassin/Dns.pm_line_374./did_0+0+1/\n\nLooking at Dns.pm doesn't really help me, and Razor2::Client::Agent appears \nto be in the right place,\nin /usr/local/lib/perl5/site_perl/5.6.1/Razor2/Client.\n\nIdeas?\n\n...Chris\n\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I've been testing Razor, invoked from sendmail/procmail and so far it\nseems pretty copacetic. Last night's spam to the list provided a good test\n- the spam itself as well as several of the responses were flagged, as\nother list members reported.\n\nThis morning I piped the messages out from pine, being careful to use the\nRAW mode, to razor-check -d. None of the messages come back as spam, even\nthe spam. Since folks revoked the false positives, I understand why they\nwould not come up, but not the spam itself, unless that also was revoked.\n\nIs this spam just a bad one to test against, or is there some setting in\npine or razor that I am missing?\n\n-- \nsc\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Once upon a time, Roi wrote :\n\n> RPM build errors:\n> user dude does not exist - using root\n> user dude does not exist - using root\n> user dude does not exist - using root\n> user dude does not exist - using root\n> user dude does not exist - using root\n\nThis I would guess is normal, but don't you get it at the very beginning of\nthe build? Isn't this at the end just a reminder?\n\n> File not found: /var/tmp/xine-root/usr/bin/aaxine\n\nArgh, I forgot to exclude aaxine from the %files when using \"--without\naalib\" :-(\nThe current \"fr6.1\" spec file fixes this...\n\nMatthias\n\n-- \nClean custom Red Hat Linux rpm packages : http://freshrpms.net/\nRed Hat Linux release 7.3 (Valhalla) running Linux kernel 2.4.18-10acpi\nLoad : 0.02 0.03 0.00\n\n_______________________________________________\nRPM-List mailing list <RPM-List@freshrpms.net>\nhttp://lists.freshrpms.net/mailman/listinfo/rpm-list\n\n\n" |
1ham
| easy_ham | "On Wed, 9 Oct 2002 09:15:03 -0400 (EDT), Samuel Checker <sc@pffcu.org> wrote:\n\n> I've been testing Razor, invoked from sendmail/procmail and so far it\n> seems pretty copacetic. Last night's spam to the list provided a good test\n> - the spam itself as well as several of the responses were flagged, as\n> other list members reported.\n> \n> This morning I piped the messages out from pine, being careful to use the\n> RAW mode, to razor-check -d. None of the messages come back as spam, even\n> the spam. Since folks revoked the false positives, I understand why they\n> would not come up, but not the spam itself, unless that also was revoked.\n> \n> Is this spam just a bad one to test against, or is there some setting in\n> pine or razor that I am missing?\n\n Are you using Spamassassin on the input side? I've just changed my sendmail installation and am looking for the 'proper' way to pass it through there, systemwide, before accepting it and sending it to the users. It's kinda problematic to set up procmail scripts for every user, when the user's home directories are NFS mounted....and the source is on my own machine, on which I try new things. (And it's the only machine with the drivespace...)\n \n------------------------------------------------------------------------\nBrian Fahrländer Linux Zealot, Conservative, and Technomad\nEvansville, IN My Voyage: http://www.CounterMoon.com\nICQ 5119262\n------------------------------------------------------------------------\nangegangen, Schlange-Hüften, sein es ganz rüber jetzt. Bügel innen fest,\nweil es eine lange, süsse Fahrt ist. \n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "I am somewhat puzzled by a phone call I (or rather the CIO at the ISP for\nwhom I work) received from an individual claiming to represent Cloudmark.\nThe gist of the call was that since we were using razor and checking our\nmail against the razor servers and that since those servers contain\ninformation proprietary to Cloudmark, we would [in the near future?] be\nrequired to begin paying Cloudmark/spamnet $1.00/user per year. I was\nwondering if anyone else has received such a call?\n\nI am curious as to whether a spammer has begun to try contacting razor users\nwith the above tactic in an effort to get them to stop using razor or\nwhether the open source/community aspect of the razor project is going by\nthe wayside in lieu of a strictly commercial approach (ala brightmail and\nthe likes).\n\nSven\n\n\n\n-------------------------------------------------------\nThis sf.net email is sponsored by:ThinkGeek\nWelcome to geek heaven.\nhttp://thinkgeek.com/sf\n_______________________________________________\nRazor-users mailing list\nRazor-users@lists.sourceforge.net\nhttps://lists.sourceforge.net/lists/listinfo/razor-users\n\n\n" |
1ham
| easy_ham | "Hi,\n\nI was inspired by a mode of operation supported by VMWare. You can have a\nbase disk image shared by multiple virtual machine (vm) instances. That base\nimage is never altered by a vm instance. Instead, each vm instance writes\nchanges to its own \"redo\" log. Future hard disk reads from that vm instance\nincorporate both the base image and the appropriate redo log to present the\ncurrent disk image for that specific virtual machine.\n\nThis is described here (thanks to Duane for providing this link on the\nhoneypots mailing list)\nhttp://www.vmware.com/support/reference/common/disk_sharing.html\n\nCould this basic concept be used to easily make self-fixing client/server\napplications that efficiently and automatically recover from most attacks,\neven before those attacks have been discovered? Here is what I imagine.\n\nThe physical architectures of most production client/server systems are\nlayered. For example, your basic web application might have a web server\nrunning Apache, connected to an application server running some J2EE or .Net\nbusiness logic, connected to a database server for persistence. The only one\nof these whose disk image really should evolve over time is the database\nserver, and even here you often put the static RDBMS software on one\npartition and the changeable datafiles on another partition. It is only the\npartition with the volatile datafiles that must be allowed to change from\none boot to the next. Other paritions may need to be writable for, say, swap\nspace, but these changes could be eliminated on each reboot.\n\nWhen someone cracks this system, they will probably change an image that\nshouldn't be changed. E.g., they might leverage a buffer overflow in IIS or\nApache to install a trojan or a backdoor on the more exposed web server. But\nwhat if the web server ran off a base image, writing changes to a \"delta\" or\n\"redo\" partition? And then what if every night it automatically erased the\nredo partition and rebooted? The downtime involved for each machine would be\nminimal, because it is only deleting data - rather than restoring from\nbackup. In a system with redundant web servers for load balancing or high\navailability, this could be scheduled in a way such that the system is\nalways accessible. This base/redo partition concept could be implemented at\nthe same level as a feature of hardware RAID, allowing for greater\nperformance, reliability, and hack resistance. This concept could also be\napplied to the application servers, and even the database server partitions\n(except for those partitions which contain the table data files, of course.)\n\nDoes anyone do this already? Or is this a new concept? Or has this concept\nbeen discussed before and abandoned for some reasons that I don't yet know?\nI use the physical architecture of a basic web application as an example in\nthis post, but this concept could of course be applied to most server\nsystems. It would allow for the hardware-separation of volatile and\nnon-volatile disk images. It would be analogous to performing nightly\nghosting operations, only it would be more efficient and involve less (or\nno) downtime.\n\nThanks for any opinions,\nBen\n\n\n" |
1ham
| easy_ham | "On Tue, Sep 03, 2002 at 09:03:40PM -0400, Yannick Gingras wrote:\n> This make me wonder about the relative protection of smart cards. They have \n> an internal procession unit around 4MHz. Can we consider them as trusted \n> hardware ? The ability to ship smart cards periodicaly uppon cashing of a \n> monthly subscription fee would not raise too much the cost of \"renting\" the \n> system. Smart card do their own self encryption. Can they be used to \n> decrypt data needed by the system ? The input of the system could me mangled \n> and the would keep a reference of how long it was in service.\n> \n> This sounds really feasible but I may be totaly wrong. I may also be wrong \n> about the safety of a smart card. \n> \n> What do you think ?\n\nThat's similar to using hard-locks (either the old parallel, or the new\nusb).\n\nThe problem is that that piece of hardware is trustworthy, but the rest of\nthe PC isn't, so a cracker just needs to simulate the lock/smart card, or\npeek at the executable after the lock has been deactivated.\n\nRegards,\nLuciano Rocha\n\n-- \nConsciousness: that annoying time between naps.\n\n" |
1ham
| easy_ham | "Scott MacKenzie wrote:\n\n>There is a software package that is used (or was up through w2k) \n>on MicroSloth for this purpose. Ghost, or some such. One essentially \n>\"takes a picture\" of the machine's proper config, and then upon \n>schedule or demand replaces the machine's current config with the \n>proper picture. It essentially over-writes the entire disk drive. \n>Especially good for student access machines at libraries, etc.\n>\nAnd it is pretty common practice in some environments with public \nworkstations to just wipe and re-install Windows machines on a weekly \n(or even daily) basis. It's easier than trying to maintain Windows.\n\nCrispin\n\n-- \nCrispin Cowan, Ph.D.\nChief Scientist, WireX http://wirex.com/~crispin/\nSecurity Hardened Linux Distribution: http://immunix.org\nAvailable for purchase: http://wirex.com/Products/Immunix/purchase.html\n\n\n\n\n" |
1ham
| easy_ham | "the only way to insure a safe key is to use all the storage space in the\nuniverse. too big to decrypt.\n\nmy point is there will never be a \"safe\" key. what I would consider is how\nlong does the data need to be protected. if you need to protect the data for\nlonger than 6 months, do not release it to the public. if you are trying to\nstop the general public (your customer) from coping the data then use what\nis available on the market. If you want to stop the bootleggers do not\nrelease the data to the public.\n\nI have never seen a lock that could not be unlocked. the act of unlocking\nthe key gives away it's secret.\n\nthe tougher the lock the more pissed-off your customers will be. take MS-XP\nfor example. only the home user is forced to register. think of the\nnightmare if business had to register every copy. how many times have we\nneeded to reinstall our laptop OS? notice the amount of Mac's sold after the\nXP release. these where mostly home users that converted to Mac OS.\n\nthe new Audio CD's that have digital copy protection so not play on my\ncomputer. does this stop me from copying the CD? no. however it does make me\nreturn them and get my money back.\n\nthe more popular the software the more likely it is to be cracked.\n\njef\n\n\n\n" |
1ham
| easy_ham | ["\n\n> Simple approxmation to this: make /usr a separate partion, and mount it \n> read-only:\n> \n> * The good news: attackers that want to trojan your software have to\n> reboot, at least.\n> * The bad news: administrators that want to update your software\n> have to reboot, at least.\n\nNo reboot is required, you just need to remount it:\n\n\t# mount -o remount,rw /usr\n\nThis requires root access, but presumably /usr is safe from non-root\nusers anyway.\n\nOnly way to disable this is to have the kernel compiled with something\nthat compartmentalizes capabilities (LIDS/etc on Linux for example) or to\nremove CAP_SYS_ADMIN with lcap, which would definately require a reboot,\nand possibly break some other functionatily to boot. (Pun intended. My\napologies.)\n\n--\nBrian Hatch \"Are you expected?\"\n Systems and \"No. Dreaded.\"\n Security Engineer\nwww.hackinglinuxexposed.com\n\nEvery message PGP signed\n", "…"] |
1ham
| easy_ham | "> However, cracking and reverse engineering tools are not so ubiquitous on\n> UNIX as they are on Windows platform for two main reasons:\n>\n> 1. The main customers of commercial Unices (Solaris, HP-UX, Aix, SCO...)\n> are respectable companies. They are ready to pay big bucks for software\n> they need: the reputation matters.\n>\n> 2. Most software for free and open source Unices like Linux and xBSD (this\n> software often may be used on commercial unices as well) is, well, free and\n> open source.\n\nThanks to your answers, I start to see where I should head for. What are your \nsugestions for protecting a 100% online requirement system ?\n\n-- \nYannick Gingras\nCoder for OBB : Observing Bantu-speaking Butanone\nhttp://OpenBeatBox.org\n\n\n" |
1ham
| easy_ham | "Once upon a time, Roi wrote :\n\n> oh xmms didn't work for me also\n> i used mpg123 i tought its something from me\n\nNope, this is \"normal\" as Red Hat removed all mp3 support from 8.0 because\nof patent and royalty issues :-(\nOn freshrpms.net, you can find the xmms mp3 plugin as I said, but also\nlibmad, lame (mp3 encoder), and soon mpg321 as I often used it myself.\nMany other players can also play mp3 files, like alsaplayer, xine, mplayer.\n\n> like mplayer not working also and gives black screen\n\nThis is not normal though...\nTry \"mplayer -vo help\" then try usinf various output methods to see if some\ndo work or not.\n\nMatthias\n\n-- \nClean custom Red Hat Linux rpm packages : http://freshrpms.net/\nRed Hat Linux release 7.3 (Valhalla) running Linux kernel 2.4.18-10acpi\nLoad : 0.00 0.05 0.01\n\n_______________________________________________\nRPM-List mailing list <RPM-List@freshrpms.net>\nhttp://lists.freshrpms.net/mailman/listinfo/rpm-list\n\n\n" |
1ham
| easy_ham | "Ben Mord wrote:\n\n> -----Original Message-----\n> *From:* Crispin Cowan [mailto:crispin@wirex.com]\n> *Sent:* Wednesday, September 04, 2002 5:46 PM\n> *To:* Ben Mord\n> *Cc:* Webappsec Securityfocus.Com; SECPROG Securityfocus\n> *Subject:* Re: use of base image / delta image for automated\n> recovery from attacks\n>\n> Ben Mord wrote:\n>\n>>I was inspired by a mode of operation supported by VMWare. [use VMWare's ability to rolll back state to recover from intrusions]\n>>\n> I did my dissertation work in this area (Optimistic Computing\n> <http://www.cse.ogi.edu/%7Ecrispin/hope.pubs.html>) and so was\n> interested in applying it to the security problem. Unfortunately,\n> you hit a bunch of problems:\n>\n> * When can you \"commit\" a state as being \"good\"? You can't\n> run from a redo log forever; the performance and storage\n> penalties accumulate. Even log structured file systems\n> garbage collect eventually. So you have to commit sometime.\n> The problem is that if you commit too eagerly, you might\n> commit corrupted state. If you commit too conservatively,\n> you eat performance and storage penalties.\n> * What do you do if you discover that there is corrupted state\n> in the *middle* of your redo log, and you want some of the\n> critical state that comes after it? You need some way to dig\n> the corruption out of the middle and save the rest. My\n> dissertation solves this problem, but you have to re-write\n> everything in my programming language :)\n> * Just doing this at all imposes substantial performance\n> penalties. I love VMWare, and use it every day (the best\n> $200 I ever spent on software) but it is not very fast. \n>\n> My proposed solution to the first two problems you mention is to be \n> less ambitious. The idea is that you *never* commit - instead, you \n> simply revert to base state on reboot.\n\nAh. In that case, you can use something considerably less powerful than \nVMWare. All you need is a machine configured to boot from CD-ROM and use \na RAM disk for scratch space. Numerous Linux distros are available that \nlet you boot a stateless but functional system from CD-ROM.\n\n> Obviously, you can't do this with partitions that accrue important \n> state, e.g. a partition that stores database table data.\n\n... but if you *do* want some state to persist, then you need a \nmountable writable partition. To protect it, you need some kind of \naccess control management to decide who can do what to the writable \npartition, blah blah blah ... and before you know it, the security \nproblem starts to look just like it does for conventional servers.\n\nSimple approxmation to this: make /usr a separate partion, and mount it \nread-only:\n\n * The good news: attackers that want to trojan your software have to\n reboot, at least.\n * The bad news: administrators that want to update your software\n have to reboot, at least.\n\nCrispin\n\n-- \nCrispin Cowan, Ph.D.\nChief Scientist, WireX http://wirex.com/~crispin/\nSecurity Hardened Linux Distribution: http://immunix.org\nAvailable for purchase: http://wirex.com/Products/Immunix/purchase.html\n\n\n\n" |
1ham
| easy_ham | "\n-----Original Message-----\nFrom: Crispin Cowan [mailto:crispin@wirex.com]\nSent: Wednesday, September 04, 2002 7:30 PM\nTo: Ben Mord\nCc: Webappsec Securityfocus.Com; SECPROG Securityfocus\nSubject: Re: use of base image / delta image for automated recovery from\nattacks\n\n\nBen Mord wrote:\n\n>> -----Original Message-----\n>> *From:* Crispin Cowan [mailto:crispin@wirex.com]\n>> *Sent:* Wednesday, September 04, 2002 5:46 PM\n>> *To:* Ben Mord\n>> *Cc:* Webappsec Securityfocus.Com; SECPROG Securityfocus\n>> *Subject:* Re: use of base image / delta image for automated\n>> recovery from attacks\n>>\n>> Ben Mord wrote:\n>>\n>> My proposed solution to the first two problems you mention is to be\n>> less ambitious. The idea is that you *never* commit - instead, you\n>> simply revert to base state on reboot.\n\n>Ah. In that case, you can use something considerably less powerful than\n>VMWare. All you need is a machine configured to boot from CD-ROM and use\n>a RAM disk for scratch space. Numerous Linux distros are available that\n>let you boot a stateless but functional system from CD-ROM.\n\nBut RAM is expensive, and the directory structures of many systems (e.g.\nWindows) are not sufficiently organized and standardized to make this\ncombination of bootable CDs and RAM drives practical. Even if you are\nfortunate enough to be using Linux (or another FHS-compliant *nix), you\nstill can't fit a lot on a CD. Its not unusual today to have gigabytes of\nstatic multimedia content on the web server. This particular problem can be\nalleviated somewhat by using DVDs, but this is a temporary solution at best\nwhich will become outdated quickly as our data requirements grow and hard\ndrives become cheaper.\n\n>> Obviously, you can't do this with partitions that accrue important\n>> state, e.g. a partition that stores database table data.\n\n>... but if you *do* want some state to persist, then you need a\n>mountable writable partition. To protect it, you need some kind of\n>access control management to decide who can do what to the writable\n>partition, blah blah blah ... and before you know it, the security\n>problem starts to look just like it does for conventional servers.\n\nRight. This is why you would consolidate all state of any long-term\nsignificance on just a couple partitions, and why you would not put static\napplication code on these changeable partitions. Fortunately, most large\nclient/server application physical architectures do this anyhow, because\nthese are two fundamentally different kinds of state with two very different\nsets of administrative, security, RAID, and backup requirements. People also\ntend to do this anyhow because layered logical architectures are popular\nwith the GUI at one end, business logic in the middle, and persistence\nservices at the other. This logical architecture maps naturally to a\nphysical architecture that has a static web server, a static application\nserver, and a database server that has static and changeable partitions. (I\nuse the word static versus changeable instead of writeable versus unwritable\nbecause the \"unchangeable\" partitions might be written to for temporary swap\nspace. Who knows what Windows does internally?)\n\nMy point is that there should be a market out there for a hardware RAID\ndevice that can split designated partitions into a permanent base image\npartition and a temporary delta image partition, that has some simple but\nsolid security measures to prevent the unauthorized remote modification of\nbase images, and that can be configured to clear the delta image when the\nserver is rebooted. If some vendor wished to implement this, they could then\nmarket this as a mechanism to help frustrate broad classes of attack that\nrely on the permanent modification of system or application files via buffer\noverflows, platform and middleware bugs, etc. The prevention of unauthorized\nmodification of application data, of course, would not be addressed by this\nparticular product. But there are many other techniques out there to defend\napplication data. But those techniques all assume that your system itself\nhas not been compromised at a lower level, which is where this product could\nhelp.\n\nI would have to think that these features would be relatively easy for a\nhardware RAID vendor to implement. (I'm just guessing, of course, with no\nknowledge of how hardware RAID works internally.) If anyone knows of such a\nproduct, I'd love to hear about it.\n\nBen\n\n\n" |
1ham
| easy_ham | "> Ben Mord said:\n> \n> >Ah. In that case, you can use something considerably less powerful than\n> >VMWare. All you need is a machine configured to boot from CD-ROM and use\n> >a RAM disk for scratch space. Numerous Linux distros are available that\n> >let you boot a stateless but functional system from CD-ROM.\n> \n> But RAM is expensive, and the directory structures of many systems (e.g.\n> Windows) are not sufficiently organized and standardized to make this\n> combination of bootable CDs and RAM drives practical. Even if you are\n> fortunate enough to be using Linux (or another FHS-compliant *nix), you\n> still can't fit a lot on a CD. Its not unusual today to have gigabytes of\n> static multimedia content on the web server. This particular problem can be\n> alleviated somewhat by using DVDs, but this is a temporary solution at best\n> which will become outdated quickly as our data requirements grow and hard\n> drives become cheaper.\n\nSo, just write-protect the hard disk for partitions that are static.\nI seem to recall an article on this (early 80's, Byte magazine, perhaps?)\nfor BBS systems or for testing unknown (perhaps trojan horse) software.\n\n - George\n\n-- \n ----------------------------------------------------------------------\n George Dinwiddie gdinwiddie@alberg30.org\n The gods do not deduct from man's allotted span those hours spent in\n sailing. http://www.Alberg30.org/\n ----------------------------------------------------------------------\n\n\n" |
1ham
| easy_ham | "On 5 Sep 2002, Richard Bartlett wrote:\n\nRichard,\n\n> I have a customer who is developing some printer driver code to allow \n> custom driver settings (n-up, booklet, duplex etc.) to be saved up to the \n> server to be retrieved by other users. The data is being written, by a \n> printer driver (using the logged on users authentication, to a registry \n> key) HKLM\\SYSTEM\\CurrentControlSet\\Control\\Print\\Environments\\Windows NT \n> x86\\Drivers\\Version-3\\{Driver Name}\\{Custom Key}\\Subkey).\n\nLet me get this straight; a registry key is loaded from the server onto \nthe client workstations who can modify it, then write it back onto the \nserver's own registry - which is not going to use it?\n\n> The question is, what are the security risks of allowing users to write \n> to this key? The data is string data, in the form of delimited numeric \n> values. This data is then retrieved by capable printer drivers and \n> interpreted.\n> \n> The risks as I see it are twofold;\n> (1) The risks of a compromise to the server using this registry key. I \n> think this is unlikeley as the server itself does not use this data, only \n> client PC's do. Unless someone knows a way to travel out of a hive up \n> the registry bypassing the permissions set using regedt32.\n\nWhat is the reason to write a registry key to a server if the server \nitself is not using it?\nI don't think you should worry too much about someone travelling out of \nthe hive, but again, I'm curious as to how the driver actually modifies \nthe keys on the server.\n\n> (2) The risks of a compromise to the client (far more likely). This \n> would probably be by a malformed or extremely long string in the key \n> value, which would presumably lead to either DOS or system compromise by \n> buffer overflow on the client system.\n\nAnd if the client writes the key back onto the server, yes, there's wide \nopen for something nasty here.\nTwo other things spring to mind; \n1) If anyone can modify the key, how do you make sure that two users are \nnot overwriting the same key, thereby causing undesirable effects.\n2) If anyone have permissions to write to the key (and below), anyone can \ncreate thousands of extra keys under this key, thereby filling up the \nregistry. The result of such a thing is obvious.\n\nIf I got this all wrong, I'd be happy that you clarify a bit more and tell \nme where I might have misunderstood.\n\n\nMed venlig hilsen / Best regards,\n-Allan Jensen\n\nSi hoc signum legere potes, operis boni in rebus Latinus alacribus et \nfructuosis potiri potes!\n\n\n\n\n" |
1ham
| easy_ham | "Bryan Feir [mailto:bryan@sgl.crestech.ca] wrote: \n\n>Of course, once one player key was broken, dealing with the rest became\n> a known plaintext attack, and the rest of the player keys went down\nlike\n> a row of dominos.\n\nThe actual follow-up to the Xing player break was more interesting than\nthat. The mere knowledge of known plaintext (a corresponding input and\noutput) does not necessarily make it trivial to break a properly\ndesigned systems and/or algorithm. The primary reason it was easy for\nCSS is because the CSS key was only 40-bits, and thereby easy to break\nwith exhaustive search attacks. It was only 40-bits (speculated)\nbecause of a misunderstanding of the government cryptography export\nrules at the time.\n\nEven more interesting, to me at least, was that soon after the Xing\nplayer break, people started studying the CSS algorithm itself. They\nrapidly found serious design flaws which left the 40-bit CSS algorithm\nwith an actual strength of around 23-bits (from memory, and new attacks\nmight have further reduced the strength). This is another great example\nshowing why proprietary cryptography algorithms should be viewed with\nthe greatest of suspicion.\n\n\nOn Tue, Sep 03, 2002 at 09:03:40PM -0400, Yannick Gingras wrote:\n> This make me wonder about the relative protection of smart cards.\nThey have \n> an internal procession unit around 4MHz. Can we consider them as\ntrusted \n> hardware ? \n\nYes and no. You can put a limited amount of trust in a smart card.\nThere have been any number of very clever attacks against smartcards\n(Ross Anderson in the UK has documented quite a few of these), and\nsmartcard manufactures are usually one step behind these attacks. A\nwell designed system assumes that a system smartcard will be completely\ncompromised at some point, giving an adversary all of the secrets\ncontained in the smartcard. The cryptography industry has developed a\nvariety of techniques that can reduce the impact of a compromise,\nincluding unique keys per smartcard and forward security techniques.\n\n\nLuciano Rocha [strange@nsk.yi.org] wrote: \n\n> The problem is that that piece of hardware is trustworthy, but the\nrest of\n> the PC isn't, so a cracker just needs to simulate the lock/smart card,\nor\n> peek at the executable after the lock has been deactivated.\n\nGoing back to the original question, once the encrypted material goes\noutside the trusted hardware, it is impossible to \"unbreakably\" protect\nit. There may be some mitigation steps you can take, such as the SDMI\nwatermarking, but most schemes to date have been easily broken. \n\nAnother consideration is the value of what you are trying to protect.\nWhile there is no such thing as unbreakable, adding more cost (both in\nterms of price and hassle-factor) can greatly improve the protection.\nSince you are talking about the use of standard PC workstations, I\npresume what you are trying to protect is not THAT valuable. I'm afraid\nmost security measures don't come for free.\n\n\nMichael McKay \nDirector of Software Development \nmmckay@iscubed.com \n\n Information Security Systems & Services Inc. \n 19925 Stevens Creek Blvd. Cupertino, CA 95014 \n Phone: 408.725.7136 x 4138 Fax: 408.973.7239 www.iscubed.com\n\n\n" |
1ham
| easy_ham | "reply to the mail from Ben Mord (bmord@icon-nicholson.com):\n\n> Hi,\n\nHello,\n\n< ... snipped for brevity ... >\n\n> ... This concept could also be\n> applied to the application servers, and even the database server partitions\n> (except for those partitions which contain the table data files, of course.)\n\n\tAlthough the data might just be the information that needs protecting.\n\n> Does anyone do this already? Or is this a new concept?\n\n\tI've seen this implemented for a shell server, although they chose\nto have their root on a CD-WR in a CD-R drive. Which meant that even\nwhen compromised it was only possible to examine other users data.\n\tAFAIR(emember) they just swapped CD's when a root exploit was found.\n\n> Thanks for any opinions,\n\nNP\n\nblaze your trail\n--\nredhat\n\n'I am become Shiva, destroyer of worlds'\n\n\n" |
1ham
| easy_ham | "Hi everybody!\n\nI'm writing a web application in java (tomcat + jsp/servlets + database \naccess with postgreSQL).\n\nThis will be released under the GPL and will eventually be useful as a \nframework for other web applications.\n\nThe application main focus is e-commerce, but not limited to that.\n\nI would like to use some form of cryptography to protect data on the \ndatabase, but I have some problem figuring out the right approach.\n\nAbove all, how to store passwords and keys in a shared web server.\n\nA problem that I was unable to solve is how to store keys for \nencryption/decryption. The api that I'm using is the jca (jdk1.4.x), \nand the methods of saving generated keys in keystores fails always. \n\nI can serialize the object, and store in the database, but this is not \nthe most secure approach: this key is needed to decrypt data in the \ndatabase, but the database is accessible from the web application. \nAssuming that I can find a good and secure place where to store the \ndatabase password, I can use a different database with different \nuser... Argh... to complex and doesn't really solve the problem.\n\nWhere I can found good documentation about this topic?\n\nThere is another approach that I would share with the list, something I \nthought that can be of bit interest, but probabily wrong and insecure. \nAfter all, I'm a real beginner in secure programming, and I'm here to \nlearn methods and technics.\n\nFirst of all, I need a secure way to keep database passwords secure, so \nI have to keep them separate from the main server. The right approach \ncould be using a small java bean application that run as normal user \n(not tomcat, so it is not shared with other web services or, worst, the \nnobody user), that has no shell login, but has a default home directory \nor a place where it can hold passwords and keys.\n\nThe web application could then open an ssl connection (could be done in \nthe init method at server startup) to get database passwords. The small \nbean could check via code signature/rmi/whatever else that the source \nis the right one, and handle all the database connections, or give the \ndb connection/password to the main web application.\n\nIn this way, we solve the problem of keeping the keys and passwords in \nshared directories, and also, an attacker should get root/bean user \naccount to read data. This is not perfect, and works only if your \nprovider gives the opportunity to configure a separated java \napplication (that means, really, another server running in the \nbackground).\n\nAny suggestions?\n\nThank you,\nMario Torre\n-- \nPlease avoid sending me Word or PowerPoint attachments.\nSee http://www.fsf.org/philosophy/no-word-attachments.html \n\n" |
1ham
| easy_ham | "Hi,\n\nThank you for the useful replies, I have found some interesting \ntutorials in the ibm developer connection.\n\nhttps://www6.software.ibm.com/developerworks/education/j-sec1\n\nand\n\nhttps://www6.software.ibm.com/developerworks/education/j-sec2\n\nRegistration is needed.\n\nI will post the same message on the Web Application Security list, as \nsuggested by someone.\n\nFor now, I thing I will use md5 for password checking (I will use the \napproach described in secure programmin fo linux and unix how-to).\n\nI will separate the authentication module, so I can change its \nimplementation at anytime.\n\nThank you again!\n\nMario Torre\n-- \nPlease avoid sending me Word or PowerPoint attachments.\nSee http://www.fsf.org/philosophy/no-word-attachments.html \n\n" |
1ham
| easy_ham | "URL: http://diveintomark.org/archives/2002/10/08.html#teach_a_man_to_fish\nDate: 2002-10-08T00:22:08-05:00\n\n_Kevin Hemenway_: Finding More Channels[1]. “In simple terms, there are \nthousands of web sites that are actively providing their news and headlines in \na format AmphetaDesk can understand [RSS]. And while AmphetaDesk knows about a \ngood number of these sites, it'd be impossible to hunt down each and every \nsingle possibility. So, this page is here to teach you how to fish.”\n\n\n\n[1] http://www.disobey.com/amphetadesk/finding_more.html\n\n\n" |
1ham
| easy_ham | "On 05 September 2002, Tim Peters said:\n> Greg Ward is\n> currently capturing a stream coming into python.org, and I hope we can get a\n> more modern, and cleaner, test set out of that.\n\nNot yet -- still working on the required config changes. But I have a\ncunning plan...\n\n> But if that stream contains\n> any private email, it may not be ethically possible to make that available.\n\nIt will! Part of my cunning plan involves something like this:\n\n if folder == \"accepted\": # ie. not suspected junk mail\n if (len(recipients) == 1 and\n recipients[0] in (\"guido@python.org\", \"barry@python.org\", ...)):\n folder = \"personal\"\n\nIf you (and Guido, Barry, et. al.) prefer, I could change that last\nstatement to \"folder = None\", so the mail won't be saved at all. I\n*might* also add a \"and sender doesn't look like -bounce-*, -request,\n-admin, ...\" clause to that if statement.\n\n> Can you think of anyplace to get a large, shareable ham sample apart from a\n> public mailing list? Everyone's eager to share their spam, but spam is so\n> much alike in so many ways that's the easy half of the data collection\n> problem.\n\nI believe the SpamAssassin maintainers have a scheme whereby the corpus\nof non-spam is distributed, ie. several people have bodies of non-spam\nthat they use for collectively evolving the SA score set. If that\nsounds vague, it matches my level of understanding.\n\n Greg\n-- \nGreg Ward <gward@python.net> http://www.gerg.ca/\nReality is for people who can't handle science fiction.\n" |
1ham
| easy_ham | "> I believe the SpamAssassin maintainers have a scheme whereby the corpus\n> of non-spam is distributed, ie. several people have bodies of non-spam\n> that they use for collectively evolving the SA score set. If that\n> sounds vague, it matches my level of understanding.\n\nSee if you can get a hold of that so we can do a level-playing-field\ncompetition. :-)\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "On 06 September 2002, Anthony Baxter said:\n> A snippet, hopefully not enough to trigger the spam-filters.\n\nAs an aside: one of the best ways to dodge SpamAssassin is by having an\nIn-Reply-To header. Most list traffic should meet this criterion.\n\nAlternately, I can whitelist mail to spambayes@python.org -- that'll\nwork until spammers get ahold of the list address, which usually seems\nto take a few months.\n\n Greg\n-- \nGreg Ward <gward@python.net> http://www.gerg.ca/\nGee, I feel kind of LIGHT in the head now, knowing I can't make my\nsatellite dish PAYMENTS!\n" |
1ham
| easy_ham | "\n>>>>> \"TP\" == Tim Peters <tim.one@comcast.net> writes:\n\n >> Any thought to wrapping up your spam and ham test sets for\n >> inclusion w/ the spambayes project?\n\n TP> I gave it all the thought it deserved <wink>. It would be\n TP> wonderful to get several people cranking on the same test\n TP> data, and I'm all in favor of that. OTOH, my Data/ subtree\n TP> currently has more than 35,000 files slobbering over 134\n TP> million bytes -- even if I had a place to put that much stuff,\n TP> I'm not sure my ISP would let me email it in one msg <wink>.\n\nCheck it into the spambayes project. SF's disks are cheap <wink>.\n\n-Barry\n" |
1ham
| easy_ham | "> Check it into the spambayes project. SF's disks are cheap <wink>.\n\nPerhaps more useful would be if Tim could check in the pickle(s?)\ngenerated by one of his training runs, so that others can see how\nTim's training data performs against their own corpora. This could\nalso be the starting point for a self-contained distribution (you've\ngot to start with *something*, and training with python-list data\nseems just as good as anything else).\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "\n>>>>> \"GW\" == Greg Ward <gward@python.net> writes:\n\n GW> If you (and Guido, Barry, et. al.) prefer, I could change that\n GW> last statement to \"folder = None\", so the mail won't be saved\n GW> at all.\n\nI don't care if the mail is foldered on python.org, but personal\nmessages regardless of who they're for, shouldn't be part of the\npublic spambayes repository unless specifically approved by both the\nrecipient and sender.\n\nNote also that we are much more liberal about python.org/zope.org\nmailing list traffic than most folks. Read list-managers for any\nlength of time and you'll find that there are a lot of people who\nassert strict copyright over their collections, are very protective of\ntheir traffic, and got really pissed when gmane just started\ngatewaying their messages without asking.\n\nWhich might be an appropriate for their lists, but not for ours (don't\nthink I'm suggesting we do the same -- I /like/ our laissez-faire\napproach).\n\nBut for personal email, we should be more careful.\n-Barry\n" |
1ham
| easy_ham | "Quite independently from testing and tuning the algorithm, I'd like to\nthink about deployment.\n\nEventually, individuals and postmasters should be able to download a\nspambayes software distribution, answer a few configuration questions\nabout their mail setup, training and false positives, and install it\nas a filter.\n\nA more modest initial goal might be the production of a tool that can\neasily be used by individuals (since we're more likely to find\nindividuals willing to risk this than postmasters).\n\nThere are many ways to do this. Some ideas:\n\n- A program that acts both as a pop client and a pop server. You\n configure it by telling it about your real pop servers. You then\n point your mail reader to the pop server at localhost. When it\n receives a connection, it connects to the remote pop servers, reads\n your mail, and gives you only the non-spam. To train it, you'd only\n need to send it the false negatives somehow; it can assume that\n anything is ham that you don't say is spam within 48 hours.\n\n- A server with a custom protocol that you send a copy of a message\n and that answers \"spam\" or \"ham\". Then you have a little program\n that is invoked e.g. by procmail that talks to the server. (The\n server exists so that it doesn't have to load the pickle with the\n scoring database for each message. I don't know how big that pickle\n would be, maybe loading it each time is fine. Or maybe\n marshalling.)\n\n- Your idea here.\n\nTakers? How is ESR's bogofilter packaged? SpamAssassin? The Perl\nBayes filter advertised on slashdot?\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "On 6 Sep 2002 at 10:31, Guido van Rossum wrote:\n\n> your mail, and gives you only the non-spam. To train it, you'd only need\n> to send it the false negatives somehow; it can assume that anything is\n> ham that you don't say is spam within 48 hours.\n\nI have folks who leave their email programs running 24 hours a day, constantly polling \nfor mail. If they go away for a long weekend, lots of \"friday night spam\" will become \nham on sunday night. (Friday night seems to be the most popular time)\n\n\n> - Your idea here.\n\nUltimately I'd like to see tight integration into the \"most popular email clients\".. As a \nstop-gap to the auto-ham ..\n\nHow about adding an IMAP server with a spam and deleted-ham folder. Most email \nclients can handle IMAP. Users should be able to quickly move \"spam\" into the spam \nfolder. \n\nInstead of deleting messages (or, by reprogramming the delete function) they can \nquickly move ham into the ham folder.\n\nIn either case, the message would be processed and then destroyed.\n\n\nBrad Clements, bkc@murkworks.com (315)268-1000\nhttp://www.murkworks.com (315)268-9812 Fax\nAOL-IM: BKClements\n\n" |
1ham
| easy_ham | "[Tim]\n> OTOH, my Data/ subtree currently has more than 35,000 files slobbering\n> over 134 million bytes -- even if I had a place to put that much stuff,\n> I'm not sure my ISP would let me email it in one msg <wink>.\n\n[Skip]\n> Do you have a dialup or something more modern <wink>?\n\nMuch more modern: a cable modem with a small upload rate cap. There's a\nreason the less modern uncapped @Home went out of business <wink>.\n\n> 134MB of messages zipped would probably compress pretty well - under 50MB\n> I'd guess with all the similarity in the headers and such. You could zip\n> each of the 10 sets individually and upload them somewhere.\n\nI suppose this could finish over the course of an afternoon. Now where's\n\"somewhere\"? I expect we'll eventually collect several datasets;\nSourceForge isn't a good place for it (they expect projects to distribute\nrelatively small code files, and complain if even those get big).\n\n> ...\n> How about random sampling lots of public mailing lists via gmane or\n> something similar, manually cleaning it (distributing that load over a\n> number of people) and then relying on your clever code and your\n> rebalancing script to help further cleanse it?\n\nWhat then are we training the classifier to do? Graham's scoring scheme is\nbased on an assumption that the ham-vs-spam task is *easy*, and half of that\nis due to that the ham has a lot in common. It was an experiment to apply\nhis scheme to all the comp.lang.python traffic, which is a lot broader than\nhe had in mind (c.l.py has long had a generous definition of \"on topic\"\n<wink>). I don't expect good things to come of making it ever broader,\n*unless* your goal is to investigate just how broad it can be made before it\nbreaks down.\n\n> The \"problem\" with the ham is it tends to be much more tied to one person\n> (not just intimate, but unique) than the spam.\n\nWhich is \"a feature\" from Graham's POV: the more clues, the better this\n\"smoking guns only\" approach should work.\n\n> I save all incoming email for ten days (gzipped mbox format) before it\nrolls\n> over and disappears. At any one time I think I have about 8,000-10,000\n> messages. Most of it isn't terribly personal (which I would cull before\n> passing along anyway) and much of it is machine-generated, so would be of\n> marginal use. Finally, it's all ham-n-spam mixed together. Do we call\n> that an omelette or a Denny's Grand Slam?\n\nUnless you're volunteering to clean it, tag it, package it, and distribute\nit, I'd call it irrelevant <wink>.\n\n" |
1ham
| easy_ham | "> > your mail, and gives you only the non-spam. To train it, you'd only need\n> > to send it the false negatives somehow; it can assume that anything is\n> > ham that you don't say is spam within 48 hours.\n> \n> I have folks who leave their email programs running 24 hours a day,\n> constantly polling for mail. If they go away for a long weekend,\n> lots of \"friday night spam\" will become ham on sunday night.\n> (Friday night seems to be the most popular time)\n\nSo we'll make this a config parameter.\n\n> > - Your idea here.\n> \n> Ultimately I'd like to see tight integration into the \"most popular\n> email clients\".. As a stop-gap to the auto-ham ..\n\nWhat's an auto-ham?\n\n> How about adding an IMAP server with a spam and deleted-ham\n> folder. Most email clients can handle IMAP. Users should be able to\n> quickly move \"spam\" into the spam folder.\n\nI personally don't think IMAP has a bright future, but for people who\ndo use it, that's certainly a good approach.\n\n> Instead of deleting messages (or, by reprogramming the delete\n> function) they can quickly move ham into the ham folder.\n\nYes.\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "\n Guido> Takers? How is ESR's bogofilter packaged? SpamAssassin? The\n Guido> Perl Bayes filter advertised on slashdot?\n\nDunno about the other tools, but SpamAssassin is a breeze to incorporate\ninto a procmail environment. Lots of people use it in many other ways. For\nperformance reasons, many people run a spamd process and then invoke a small\nC program called spamc which shoots the message over to spamd and passes the\nresult back out. I think spambayes in incremental mode is probably fast\nenough to not require such tricks (though I would consider changing the\npickle to an anydbm file).\n\nBasic procmail usage goes something like this:\n\n :0fw\n | spamassassin -P\n\n :0\n * ^X-Spam-Status: Yes\n $SPAM\n\nWhich just says, \"Run spamassassin -P reinjecting its output into the\nprocessing stream. If the resulting mail has a header which begins\n\"X-Spam-Status: Yes\", toss it into the folder indicated by the variable\n$SPAM.\n\nSpamAssassin also adds other headers as well, which give you more detail\nabout how its tests fared. I'd like to see spambayes operate in at least\nthis way: do its thing then return a message to stdout with a modified set\nof headers which further processing downstream can key on.\n\nSkip\n\n" |
1ham
| easy_ham | "Did you want this on the list? I'm replying to the list..\n\nOn 6 Sep 2002 at 10:43, Guido van Rossum wrote:\n\n> What's an auto-ham?\n\nAutomatically marking something as ham after a given timeout.. regardless of how long \nthat timeout is, someone is going to forget to submit the message back as spam. \n\nHow many spams-as-hams can be accepted before the f-n rate gets unacceptable?\n\n\n> > How about adding an IMAP server with a spam and deleted-ham\n> > folder. Most email clients can handle IMAP. Users should be able to\n> > quickly move \"spam\" into the spam folder.\n> \n> I personally don't think IMAP has a bright future, but for people who\n> do use it, that's certainly a good approach.\n> \n> > Instead of deleting messages (or, by reprogramming the delete\n> > function) they can quickly move ham into the ham folder.\n> \n> Yes.\n\nI view IMAP as a stop-gap measure until tighter integration with various email clients \ncan be achieved.\n\nI still feel it's better to require classification feedback from the recipient, rather than \nmake any assumptions after some period of time passes. But this is an end-user issue \nand we're still at the algorithm stage.. ;-)\n\n\n\n\nBrad Clements, bkc@murkworks.com (315)268-1000\nhttp://www.murkworks.com (315)268-9812 Fax\nAOL-IM: BKClements\n\n" |
1ham
| easy_ham | "> > What's an auto-ham?\n> \n> Automatically marking something as ham after a given\n> timeout.. regardless of how long that timeout is, someone is going\n> to forget to submit the message back as spam.\n\nOK, here's a refinement. Assuming very little spam comes through, we\nonly need to pick a small percentage of ham received as new training\nham to match the new training spam. The program could randomly select\na sufficient number of saved non-spam msgs and ask the user to\nvalidate this selection. You could do this once a day or week (config\nparameter).\n\n> How many spams-as-hams can be accepted before the f-n rate gets\n> unacceptable?\n\nConfig parameter.\n\n> I view IMAP as a stop-gap measure until tighter integration with\n> various email clients can be achieved.\n> \n> I still feel it's better to require classification feedback from the\n> recipient, rather than make any assumptions after some period of\n> time passes. But this is an end-user issue and we're still at the\n> algorithm stage.. ;-)\n\nI'm trying to think about the end-user issues because I have nothing\nto contribute to the algorithm at this point. For deployment we need\nboth!\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "\n >> Ultimately I'd like to see tight integration into the \"most popular\n >> email clients\"..\n\nThe advantage of using a kitchen sink (umm, make that highly programmable)\neditor+email package like Emacs+VM is that you can twiddle your key bindings\nand write a little ELisp (or Pymacs) glue to toss messages in the right\ndirection (spam or ham). For this, spambayes would have to operate in an\nincremental fashion when fed a single ham or spam message.\n\n(No, I have no idea what an \"auto-ham\" is. A pig run over by a car,\nperhaps?)\n\ngive-a-dog-a-bone-ly, y'rs,\n\nSkip\n" |
1ham
| easy_ham | "\n >> Dunno about the other tools, but SpamAssassin is a breeze ...\n >> SpamAssassin also adds other headers as well, which give you more\n >> detail ...\n\n Guido> Do you feel capable of writing such a tool? It doesn't look too\n Guido> hard.\n\nSure, but at the moment I have to stop reading email for a few hours and do\nsome real work. ;-) I'll see if I can modify GBayes.py suitably over the\nweekend.\n\nSkip\n" |
1ham
| easy_ham | "URL: http://www.oblomovka.com/entries/2002/10/07#1034055900\nDate: 2002-10-07T22:45:00-0700\n\n... for not updating: I'm doing the guestblog at Boing Boing[1]. Now to find an \nexcuse for missing last week.\n\n[1] http://www.boingboing.net/\n\n\n" |
1ham
| easy_ham | "Guido van Rossum wrote:\n> I personally don't think IMAP has a bright future, but for people who\n> do use it, that's certainly a good approach.\n\nWriting an IMAP server is a non-trivial task. The specification is huge\nand clients do all kinds of weird stuff. POP is very easy in\ncomparison. Perhaps you could forward messages to a special address or\nsave them in a special folder to mark them as false negatives.\n\nAlternatively, perhaps there could be a separate protocol and client\nthat could be used to review additions to the training set. Each day a\nfew random spam and ham messages could be grabbed as candidates.\nSomeone would periodically startup the client, review the candidates,\nreclassify or remove any messages they don't like and add them to the\ntraining set.\n\n Neil\n" |
1ham
| easy_ham | "On Fri, 6 Sep 2002, Guido van Rossum wrote:\n\n>Quite independently from testing and tuning the algorithm, I'd like to\n>think about deployment.\n>\n>Eventually, individuals and postmasters should be able to download a\n>spambayes software distribution, answer a few configuration questions\n>about their mail setup, training and false positives, and install it\n>as a filter.\n>\n>A more modest initial goal might be the production of a tool that can\n>easily be used by individuals (since we're more likely to find\n>individuals willing to risk this than postmasters).\n\nMy impression is that a pre-collected corpus would not fit most individuals\nvery well, but each individual (or group?) should collect their own corpus.\n\nOne problem that comes upp immediately: individuals are lazy.\n\nIf I currently get 50 spam and 50 ham a day, and I'll have to\npress the 'delete' button once for each spam, I'll be happy\nto press the 'spam' button instead. However, if in addition\nhave to press a 'ham' button for each ham, it starts to look\nmuch less like a win to me. Add the time to install and setup\nthe whole machinery, and I'll just keep hitting delete.\n\nThe suggestions so far have been to hook something on the delete\naction, that adds a message to the ham corpus. I see two problems\nwith this: the ham will be a bit skewed; mail that I keep around\nwithout deleting will not be counted. Secondly, if I by force of\nhabit happen to press the 'delete' key instead of the 'spam' key,\nI'll end up with spam in the ham, anyways.\n\nI would like to look for a way to deal with spam in the ham.\n\nThe obvious thing to do is to trigger on the 'spam' button,\nand at that time look for messages similar to the deleted one\nin the ham corpus, and simply remove them. To do this we\nneed a way to compare two word count histograms, to see\nhow similar they are. Any ideas ?\n\nAlso, I personally would prefer to not see the spam at all.\nIf they get bounced (preferably already in the SMTP),\nfalse positives become the senders problem, to rewrite\nto remove the spam smell.\n\nIn a well tuned system then, there spam corpus will be much\nsmaller than the ham corpus, so it would be possible to be\nslightly over-agressive when clearing potential spam from\nthe ham corpus. This should make it easier to keep it clean.\n\nHaving a good way to remove spam from the ham corpus,\nthere's less need to worry about it getting there by mistake,\nand we might as well simply add all messages to the ham corpus,\nthat didn't get deleted by the spam filtering.\n\nIt might also be useful to have a way to remove messages from\nthe spam corpus, in case of user ooops.\n\n\t/Paul\n\n" |
1ham
| easy_ham | "> I guess MUA-level filtering is just a fallback for people who don't have\n> 1) a burning, all-consuming hatred of junk mail, 2) root access to all\n> mail servers they rely on, and 3) the ability and inclination to install\n> an MTA with every bell and whistle tweaked to keep out junk mail.\n\nSure. But for most people, changing their company's or ISP's server\nrequires years of lobbying, while they have total and immediate\ncontrol over their own MUA.\n\nThat said, I agree that we should offer a good solution to\npostmasters, and I trust that your ideas are right on the mark!\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "I think one step towards deployment is creating a re-usable tokenizer\nfor mail messages. The current codebase doesn't expose an easy-to-use\nor easy-to-customize tokenizer.\n\nThe timtest module seems to contain an enormous body of practical\nknowledge about how to parse mail messages, but the module wasn't\ndesigned for re-use. I'd like to see a module that can take a single\nmessage or a collection of messages and tokenize each one.\n\nI'd like to see the tokenize by customizable, too. Tim had to exclude\nsome headers from his test data, because there were particular biases\nin the test data. If other people have test data without those\nbiases, they ought to be able to customize the tokenizer to include\nthem or exclude others.\n\nJeremy\n\n" |
1ham
| easy_ham | "[Anthony Baxter]\n> The other thing on my todo list (probably tonight's tram ride home) is\n> to add all headers from non-text parts of multipart messages. If nothing\n> else, it'll pick up most virus email real quick.\n\nSee the checkin comments for timtest.py last night. Adding this code gave a\nmajor reduction in the false negative rate:\n\ndef crack_content_xyz(msg):\n x = msg.get_type()\n if x is not None:\n yield 'content-type:' + x.lower()\n\n x = msg.get_param('type')\n if x is not None:\n yield 'content-type/type:' + x.lower()\n\n for x in msg.get_charsets(None):\n if x is not None:\n yield 'charset:' + x.lower()\n\n x = msg.get('content-disposition')\n if x is not None:\n yield 'content-disposition:' + x.lower()\n\n fname = msg.get_filename()\n if fname is not None:\n for x in fname.lower().split('/'):\n for y in x.split('.'):\n yield 'filename:' + y\n\n x = msg.get('content-transfer-encoding:')\n if x is not None:\n yield 'content-transfer-encoding:' + x.lower()\n\n\n...\n\n t = ''\n for x in msg.walk():\n for w in crack_content_xyz(x):\n yield t + w\n t = '>'\n\nI *suspect* most of that stuff didn't make any difference, but I put it all\nin as one blob so don't know which parts did and didn't help.\n\n" |
1ham
| easy_ham | "\n TP> A false positive *really* has to work hard then, eh? The long\n TP> quote of a Nigerian scam letter is one of the two that made\n TP> it, and spamprob() looked at all this stuff before deciding it\n TP> was spam:\n\nHere's an interesting thing to test: discriminate words differently if\nthey are on a line that starts with `>' or, to catch styles like\nabove, that the first occurance on a line of < or > is > (to eliminate\nhtml).\n\nThen again, it may not be worth trying to un-false-positive that\nNigerian scam quote.\n\n-Barry\n" |
1ham
| easy_ham | "So then, Guido van Rossum <guido@python.org> is all like:\n\n> > Basic procmail usage goes something like this:\n> > \n> > :0fw\n> > | spamassassin -P\n> > \n> > :0\n> > * ^X-Spam-Status: Yes\n> > $SPAM\n> > \n> \n> Do you feel capable of writing such a tool? It doesn't look too hard.\n\nNot to beat a dead horse, but that's exactly what my spamcan package\ndid. For those just tuning in, spamcan is a thingy I wrote before I\nknew about Tim & co's work on this crazy stuff; you can download it from\n<http://woozle.org/~neale/src/spamcan/spamcan.html>, but I'm not going\nto work on it anymore.\n\nI'm currently writing a new one based on classifier (and timtest's\nbooty-kicking tokenizer). I'll probably have something soon, like maybe\nhalf an hour, and no, it's not too hard. The hard part is storing the\ndata somewhere. I don't want to use ZODB, as I'd like something a\nperson can just drop in with a default Python install. So anydbm is\nlooking like my best option.\n\nI already have a setup like this using Xavier Leroy's SpamOracle, which\ndoes the same sort of thing. You call it from procmail, it adds a new\nheader, and then you can filter on that header. Really easy.\n\nHere's how I envision this working. Everybody gets four new mailboxes:\n\n train-eggs\n train-spam\n trained-eggs\n trained-spam\n\nYou copy all your spam and eggs* into the \"train-\" boxes as you get it.\nHow frequently you do this would be up to you, but you'd get better\nresults if you did it more often, and you'd be wise to always copy over\nanything which was misclassified. Then, every night, the spam fairy\nswoops down and reads through your folders, learning about what sorts of\nthings you think are eggs and what sorts of things are spam. After she's\ndone, she moves your mail into the \"trained-\" folders.\n\nThis would work for anybody using IMAP on a Unix box, or folks who read\ntheir mail right off the server. I've spoken with some fellows at work\nabout Exchange and they seem to beleive that Exchange exports\nappropriate functionality to implement a spam fairy as well.\n\nAdvanced users could stay ahead of the game by reprogramming their mail\nclient to bind the key \"S\" to \"move to train-spam\" and \"H\" to \"move to\ntrain-eggs\". Eventually, if enough people used this sort of thing, it'd\nstart showing up in mail clients. That's the \"delete as spam\" button\nPaul Graham was talking about.\n\n* The Hormel company might not think well of using the word \"ham\" as the\n opposite of \"spam\", and they've been amazingly cool about the use of\n their product name for things thus far. So I propose we start calling\n non-spam something more innocuous (and more Monty Pythonic) such as\n \"eggs\".\n\nNeale\n" |
1ham
| easy_ham | "I think that when considering deployment, a solution that supports all\nPython platforms and not just the L|Unix crowd is desirable.\n\nMac and PC users are more apt to be using a commercial MUA that's unlikely\nto offer hooking ability (at least not easily). As mentioned elsewhere, even\nL|Unix users may find an MUA solution easier to use then getting it added to\ntheir MTA. (SysOps make programmers look like flaming liberals ;).)\n\nMy notion of a solution for Windows/Outlook has been, as Guido described, a\nclient-server. Client side does pop3/imap/mapi fetching (of which, I'm only\ngoing to implement pop3 initially) potentially on several hosts, spamhams\nthe incoming mail and puts it into one file per message (qmail style?). The\nMUA accesses this \"eThunk\" as a server to obtain all the ham. Spam is\nretained in the eThunk and a simple viewer would be used for manual\noversight on the spam for ultimate rejection (and training of spam filter)\nand the ham will go forward (after being used for training) on the next MUA\nfetch. eThunk would sit on a timer for 'always online' users, but I am not\nclear on how to support dialup users with this scheme.\n\nOutbound mail would use a direct path from the MUA to the MTA. Hopefully all\nMUAs can split the host fetch/send URL's\n\nIMO, end users are likely to be more intested in n-way classification. If\nthis is available, the \"simple viewer\" could be enhanced to support viewing\nvia folders and (at least for me) the Outlook nightmare is over - I would\nuse this as my only MUA. (N.B. according to my recent readings, the best\nn-way classifier uses something called a \"Support Vector Machine\" (SVM)\nwhich is 5-8% more accurate then Naive Bayes (NB) ).\n\nI wonder if the focus of spambayes ought not to be a classifier that leaves\nthe fetching and feeding of messages to auxillary code? That way, it could\nbe dropped into whatever harness that suited the user's situation.\n\nDavid LeBlanc\nSeattle, WA USA\n\n\n" |
1ham
| easy_ham | "> I wonder if the focus of spambayes ought not to be a classifier that\n> leaves the fetching and feeding of messages to auxillary code? That\n> way, it could be dropped into whatever harness that suited the\n> user's situation.\n\nI see no reason to restrict the project to developing the classifier\nand leave the deployment to others. Attempts at deployment in the\nreal world will surely provide additional feedback for the classifier.\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "On 06 September 2002, Tim Peters said:\n> > Note that header names are case insensitive, so this one's no\n> > different than \"MIME-Version:\". Similarly other headers in your list.\n> \n> Ignoring case here may or may not help; that's for experiment to decide.\n> It's plausible that case is significant, if, e.g., a particular spam mailing\n> package generates unusual case, or a particular clueless spammer\n> misconfigures his package.\n\nCase of headers is definitely helpful. SpamAssassin has a rule for it\n-- if you have headers like \"DATE\" or \"SUBJECT\", you get a few more\npoints.\n\n Greg\n-- \nGreg Ward <gward@python.net> http://www.gerg.ca/\nGod is omnipotent, omniscient, and omnibenevolent\n---it says so right here on the label.\n" |
1ham
| easy_ham | "On Friday 06 September 2002 20:48, Skip Montanaro wrote:\n> Greg> In case it wasn't obvious, I'm a strong proponent of\n> filtering Greg> junk mail as early as possible, ie. right after the\n> SMTP DATA Greg> command has been completed. Filtering spam at the\n> MUA just seems Greg> stupid to me -- by the time it gets to me MUA,\n> the spammer has Greg> already stolen my bandwidth.\n>\n> The two problems I see with filtering that early are:\n>\n> 1. Everyone receiving email via that server will contribute ham\n> to the stew, making the Bayesian classification less effective.\n>\n> 2. Given that there will be some false positives, you absolutely\n> have to put the mail somewhere. You can't simply delete it. (I also\n> don't like the TMDA-ish business of replying with a msg that says,\n> \"here's what you do to really get your message to me.\" That puts an\n> extra burden on my correspondents.) As an individual, I would prefer\n> you put spammish messages somewhere where I can review them, not an\n> anonymous sysadmin who I might not trust with my personal email\n> (nothing against you Greg ;-).\n>\n> I personally prefer to manage this stuff at the user agent level. \n> Bandwidth is a heck of a lot cheaper than my time.\n>\n\nI see no reason why both approaches could and should not be used.\nMTA level filtering would just need to use a different corpus, one that \nwould contain illegal or otherwise commonly unapproved material for the \ngroup of people using that MTA. I'm sure that such an approach would \nsignificantly reduce the mail traffic as a first step, without giving \nfalse positives.\n\nMUA corpus would then be personally trained -- although I'd like the \noption of 'down-loadable' corpuses and merge functionality.\n\nHarri\n\nPS. Just joined the list, so pardon if my thoughts have been hashed \nthrough before.\n\n\n" |
1ham
| easy_ham | "[Guido]\n> ...\n> - A program that acts both as a pop client and a pop server. You\n> configure it by telling it about your real pop servers. You then\n> point your mail reader to the pop server at localhost. When it\n> receives a connection, it connects to the remote pop servers, reads\n> your mail, and gives you only the non-spam.\n\nFYI, I'll never trust such a scheme: I have no tolerance for false\npositives, and indeed do nothing to try to block spam on any of my email\naccounts now for that reason. Deliver all suspected spam to a Spam folder\ninstead and I'd love it.\n\n" |
1ham
| easy_ham | "> > - A program that acts both as a pop client and a pop server. You\n> > configure it by telling it about your real pop servers. You then\n> > point your mail reader to the pop server at localhost. When it\n> > receives a connection, it connects to the remote pop servers, reads\n> > your mail, and gives you only the non-spam.\n> \n> FYI, I'll never trust such a scheme: I have no tolerance for false\n> positives, and indeed do nothing to try to block spam on any of my email\n> accounts now for that reason. Deliver all suspected spam to a Spam folder\n> instead and I'd love it.\n\nAnother config parameter.\n\nThe filter could add a header file. Or a ~ to the subject if you like\nthat style. :-)\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "[Guido]\n> Takers? How is ESR's bogofilter packaged? SpamAssassin? The Perl\n> Bayes filter advertised on slashdot?\n\nWRT the last, it's a small pile of Windows .exe files along with\ncygwin1.dll. The .exes are cmdline programs. One is a POP3 proxy. If I\ncurrently have an email server named, say, mail.comcast.net, with user name\ntimmy, then I change my email reader to say that my server is 127.0.0.1, and\nthat my user name on that server is mail.comcast.net:timmy. In that way the\nproxy picks up both the real server and user names from what the mail reader\ntells it the user name is.\n\nThis is an N-way classifier (like ifile that way), and \"all it does\" is\ninsert a\n\n X-Text-Classification: one_of_the_class_names_you_picked\n\nheader into your email before passing it on to your mail reader. The user\nthen presumably fiddles their mail reader to look for such headers and \"do\nsomething about it\" (and even Outlook can handle *that* much <wink>).\n\nThe user is responsible for generating text files with appropriate examples\nof each class of message, and for running the cmdline tools to train the\nclassifier.\n\n" |
1ham
| easy_ham | "You missed the part that said that spam is kept in the \"eThunk\" and was\nviewable by a simple viewer for final disposition?\n\nOf course, with Outbloat, you could fire up PythonWin and stuff the spam\ninto the Junk Email folder... but then you loose the ability to retrain on\nthe user classified ham/spam.\n\nDavid LeBlanc\nSeattle, WA USA\n\n> -----Original Message-----\n> From: spambayes-bounces+whisper=oz.net@python.org\n> [mailto:spambayes-bounces+whisper=oz.net@python.org]On Behalf Of Tim\n> Peters\n> Sent: Friday, September 06, 2002 12:24\n> To: spambayes@python.org\n> Subject: RE: [Spambayes] Deployment\n>\n>\n> [Guido]\n> > ...\n> > - A program that acts both as a pop client and a pop server. You\n> > configure it by telling it about your real pop servers. You then\n> > point your mail reader to the pop server at localhost. When it\n> > receives a connection, it connects to the remote pop servers, reads\n> > your mail, and gives you only the non-spam.\n>\n> FYI, I'll never trust such a scheme: I have no tolerance for false\n> positives, and indeed do nothing to try to block spam on any of my email\n> accounts now for that reason. Deliver all suspected spam to a Spam folder\n> instead and I'd love it.\n>\n>\n> _______________________________________________\n> Spambayes mailing list\n> Spambayes@python.org\n> http://mail.python.org/mailman-21/listinfo/spambayes\n\n" |
1ham
| easy_ham | "So then, Tim Peters <tim.one@comcast.net> is all like:\n\n> [Guido]\n> > ...\n> > I don't know how big that pickle would be, maybe loading it each time\n> > is fine. Or maybe marshalling.)\n> \n> My tests train on about 7,000 msgs, and a binary pickle of the database is\n> approaching 10 million bytes.\n\nMy paltry 3000-message training set makes a 6.3MB (where 1MB=1e6 bytes)\npickle. hammie.py, which I just checked in, will optionally let you\nwrite stuff out to a dbm file. With that same message base, the dbm\nfile weighs in at a hefty 21.4MB. It also takes longer to write:\n\n Using a database:\n real 8m24.741s\n user 6m19.410s\n sys 1m33.650s\n\n Using a pickle:\n real 1m39.824s\n user 1m36.400s\n sys 0m2.160s\n\nThis is on a PIII at 551.257MHz (I don't know what it's *supposed* to\nbe, 551.257 is what /proc/cpuinfo says).\n\nFor comparison, SpamOracle (currently the gold standard in my mind, at\nleast for speed) on the same data blazes along:\n\n real 0m29.592s\n user 0m28.050s\n sys 0m1.180s\n\nIts data file, which appears to be a marshalled hash, is 448KB.\nHowever, it's compiled O'Caml and it uses a much simpler tokenizing\nalgorithm written with a lexical analyzer (ocamllex), so we'll never be\nable to outperform it. It's something to keep in mind, though.\n\nI don't have statistics yet for scanning unknown messages. (Actually, I\ndo, and the database blows the pickle out of the water, but it scores\nevery word with 0.00, so I'm not sure that's a fair test. ;) In any\ncase, 21MB per user is probably too large, and 10MB is questionable. \n\nOn the other hand, my pickle compressed very well with gzip, shrinking\ndown to 1.8MB.\n\nNeale\n" |
1ham
| easy_ham | "URL: http://boingboing.net/#85537496\nDate: Not supplied\n\nCurcumin, the chemical that makes curry yellow, turns out to be a good compound \nfor treating radiation burns resulting from cancer therapy. Link[1] Discuss[2] \n(_Thanks, Cheryl!_)\n\n[1] http://www.alertnet.org/thenews/newsdesk/N07347915\n[2] http://www.quicktopic.com/boing/H/THKNJnrnHdDd\n\n\n" |
1ham
| easy_ham | "So then, Tim Peters <tim.one@comcast.net> is all like:\n\n> [Tim]\n> > My tests train on about 7,000 msgs, and a binary pickle of the database is\n> > approaching 10 million bytes.\n> \n> That shrinks to under 2 million bytes, though, if I delete all the WordInfo\n> records with spamprob exactly equal to UNKNOWN_SPAMPROB. Such records\n> aren't needed when scoring (an unknown word gets a made-up probability of\n> UNKNOWN_SPAMPROB). Such records are only needed for training; I've noted\n> before that a scoring-only database can be leaner.\n\nThat's pretty good. I wonder how much better you could do by using some\ncustom pickler. I just checked my little dbm file and found a lot of\nwhat I would call bloat:\n\n >>> import anydbm, hammie\n >>> d = hammie.PersistentGrahamBayes(\"ham.db\")\n >>> db = anydbm.open(\"ham.db\")\n >>> db[\"neale\"], len(db[\"neale\"])\n ('ccopy_reg\\n_reconstructor\\nq\\x01(cclassifier\\nWordInfo\\nq\\x02c__builtin__\\nobject\\nq\\x03NtRq\\x04(GA\\xce\\xbc{\\xfd\\x94\\xbboK\\x00K\\x00K\\x00G?\\xe0\\x00\\x00\\x00\\x00\\x00\\x00tb.', 106)\n >>> d.wordinfo[\"neale\"], len(`d.wordinfo[\"neale\"]`)\n (WordInfo'(1031337979.16197, 0, 0, 0, 0.5)', 42)\n\nIgnoring the fact that there are too many zeros in there, the pickled\nversion of that WordInfo object is over twice as large as the string\nrepresentation. So we could get a 50% decrease in size just by using\nthe string representation instead of the pickle, right?\n\nSomething about that logic seems wrong to me, but I can't see what it\nis. Maybe pickling is good for heterogeneous data types, but every\nvalue of our big dictionary is going to have the same type, so there's a\nton of redundancy. I guess that explains why it compressed so well.\n\nNeale\n" |
1ham
| easy_ham | "\n >> > ##Remove: jeremy@alum.mit.edu##\n\n Tim> Yuck: it got two 0.01's from embedding your email address at the\n Tim> bottom here.\n\nWhich suggests that tagging email addresses in To/CC headers should be\nhandled differently than in message bodies?\n\nSkip\n\n\n\n" |
1ham
| easy_ham | "> >> > ##Remove: jeremy@alum.mit.edu##\n>\n> Tim> Yuck: it got two 0.01's from embedding your email address at the\n> Tim> bottom here.\n>\n> Which suggests that tagging email addresses in To/CC headers should be\n> handled differently than in message bodies?\n\nI don't know whether it suggests that, but they would be tagged differently\nin to/cc if I were tagging them at all right now. If I were tagging To:\naddresses, for example, the tokens would look like\n\n 'to:email addr:mit'\n\ninstead of\n\n 'email addr:mit'\n\nas they appear when an email-like thingie is found in the body. Whether\nemail addresses should be stuck in as one blob or split up as they are now\nis something I haven't tested.\n\n" |
1ham
| easy_ham | " http://sf.net/project/showfiles.php?group_id=61702\n\n This is the binary pickle of my classifier after training on\n my first spam/ham corpora pair. All records with\n spamprob == UNKNOWN_SPAMPROB have been purged.\n\nIt's in a zip file, and is only half a meg.\n\nJeremy, it would be interesting if you tried that on your data. The false\nnegative rates across my other 4 test sets when run against this are:\n\n 0.364%\n 0.400%\n 0.400%\n 0.909%\n\n" |
1ham
| easy_ham | ">>>>> \"TP\" == Tim Peters <tim.one@comcast.net> writes:\n\n >> The false positive rate is 0-3%. (Finally! I had to scrub a\n >> bunch of previously unnoticed spam from my inbox.) Both\n >> collections have about 1100 messages.\n\n TP> Does this mean you trained on about 1100 of each?\n\nThe total collections are 1100 messages. I trained with 1100/5\nmessages. \n\n TP> Can't guess. You're in a good position to start adding more\n TP> headers into the analysis, though. For example, an easy start\n TP> would be to uncomment the header-counting lines in tokenize()\n TP> (look for \"Anthony\"). Likely the most valuable thing it's\n TP> missing then is some special parsing and tagging of Received\n TP> headers.\n\nI tried the \"Anthony\" stuff, but it didn't make any appreciable\ndifference that I could see from staring at the false negative rate.\nThe numbers are big enough that a quick eyeball suffices.\n\nThen I tried a dirt simple tokenizer for the headers that tokenize the\nwords in the header and emitted like this \"%s: %s\" % (hdr, word).\nThat worked too well :-). The received and date headers helped the\nclassifier discover that most of my spam is old and most of my ham is\nnew.\n\nSo I tried a slightly more complex one that skipped received, data,\nand x-from_, which all contained timestamps. I also skipped the X-VM-\nheaders that my mail reader added:\n\nclass MyTokenizer(Tokenizer):\n\n skip = {'received': 1,\n 'date': 1,\n 'x-from_': 1,\n }\n\n def tokenize_headers(self, msg):\n for k, v in msg.items():\n k = k.lower()\n if k in self.skip or k.startswith('x-vm'):\n continue\n for w in subject_word_re.findall(v):\n for t in tokenize_word(w):\n yield \"%s:%s\" % (k, t)\n\nThis did moderately better. The false negative rate is 7-21% over the\ntests performed so far. This is versus 11-28% for the previous test\nrun that used the timtest header tokenizer.\n\nIt's interesting to see that the best descriminators are all ham\ndiscriminators. There's not a single spam-indicator in the list.\nMost of the discriminators are header fields. One thing to note is\nthat the presence of Mailman-generated headers is a strong non-spam\nindicator. That matches my intuition: I got an awful lot of\nMailman-generated mail, and those lists are pretty good at surpressing\nspam. The other thing is that I get a lot of ham from people who use\nXEmacs. That's probably Barry, Guido, Fred, and me :-).\n\nOne final note. It looks like many of the false positives are from\npeople I've never met with questions about Shakespeare. They often\nstart with stuff like:\n\n> Dear Sir/Madam,\n> \n> May I please take some of your precious time to ask you to help me to find a\n> solution to a problem that is worrying me greatly. I am old science student\n\nI guess that reads a lot like spam :-(.\n\nJeremy\n\n\n238 hams & 221 spams\n false positive: 2.10084033613\n false negative: 9.50226244344\n new false positives: []\n new false negatives: []\n\n best discriminators:\n 'x-mailscanner:clean' 671 0.0483425\n 'x-spam-status:IN_REP_TO' 679 0.01\n 'delivered-to:skip:s 10' 691 0.0829876\n 'x-mailer:Lucid' 699 0.01\n 'x-mailer:XEmacs' 699 0.01\n 'x-mailer:patch' 699 0.01\n 'x-mailer:under' 709 0.01\n 'x-mailscanner:Found' 716 0.0479124\n 'cc:zope.com' 718 0.01\n \"i'll\" 750 0.01\n 'references:skip:1 20' 767 0.01\n 'rossum' 795 0.01\n 'x-spam-status:skip:S 10' 825 0.01\n 'van' 850 0.01\n 'http0:zope' 869 0.01\n 'email addr:zope' 883 0.01\n 'from:python.org' 895 0.01\n 'to:jeremy' 902 0.185401\n 'zope' 984 0.01\n 'list-archive:skip:m 10' 1058 0.01\n 'list-subscribe:skip:m 10' 1058 0.01\n 'list-unsubscribe:skip:m 10' 1058 0.01\n 'from:zope.com' 1098 0.01\n 'return-path:zope.com' 1115 0.01\n 'wrote:' 1129 0.01\n 'jeremy' 1150 0.01\n 'email addr:python' 1257 0.01\n 'x-mailman-version:2.0.13' 1311 0.01\n 'x-mailman-version:101270' 1395 0.01\n 'python' 1401 0.01\n\n" |
1ham
| easy_ham | "[Neale Pickett]\n> I hacked up something to turn WordInfo into a tuple before pickling,\n\nThat's what WordInfo.__getstate__ does.\n\n> and then turn the tuple back into WordInfo right after unpickling.\n\nLikewise for WordInfo.__setstate__.\n\n> Without this hack, my database was 21549056 bytes. After, it's 9945088\nbytes.\n> That's a 50% savings, not a bad optimization.\n\nI'm not sure what you're doing, but suspect you're storing individual\nWordInfo pickles. If so, most of the administrative pickle bloat is due to\nthat, and doesn't happen if you pickle an entire classifier instance\ndirectly.\n\n> So my question is, would it be too painful to ditch WordInfo in favor of\n> a straight out tuple? (Or list if you'd rather, although making it a\n> tuple has the nice side-effect of forcing you to play nice with my\n> DBDict class).\n>\n> I hope doing this sort of optimization isn't too far distant from the\n> goal of this project, even though README.txt says it is :)\n>\n> Diff attached. I'm not comfortable checking this in,\n\nI think it's healthy that you're uncomfortable checking things in with\n\n> + # XXX: kludge kludge kludge.\n\ncomments <wink>.\n\n> since I don't really like how it works (I'd rather just get rid of\nWordInfo).\n> But I guess it proves the point :)\n\nI'm not interested in optimizing anything yet, and get many benefits from\nthe *ease* of working with utterly vanilla Python instance objects. Lots of\ncode all over picks these apart for display and analysis purposes. Very few\npeople have tried this code yet, and there are still many questions about it\n(see, e.g., Jeremy's writeup of his disappointing first-time experiences\ntoday). Let's keep it as easy as possible to modify for now. If you're\ndesparate to save memory, write a subclass?\n\nOther people are free to vote in other directions, of course <wink>.\n\n" |
1ham
| easy_ham | "[Barry]\n> Here's an interesting thing to test: discriminate words differently if\n> they are on a line that starts with `>' or, to catch styles like\n> above, that the first occurance on a line of < or > is > (to eliminate\n> html).\n\nGive me a mod to timtoken.py that does this, and I'll be happy to test it.\n\n> Then again, it may not be worth trying to un-false-positive that\n> Nigerian scam quote.\n\nIf there's any sanity in the world, even the original poster would be glad\nto have his kneejerk response blocked <wink>. OTOH, you know there are a\ngreat many msgs on c.l.py (all over Usenet) that do nothing except quote a\nprevious post and add a one-line comment. Remove the quoted sections from\nthose, and there may be no content left to judge except for the headers. So\nI can see this nudging the stats in either direction. The only way to find\nout for sure is for you to write some code <wink>.\n\n" |
1ham
| easy_ham | "[Jeremy]\n> The total collections are 1100 messages. I trained with 1100/5\n> messages.\n\nWhile that's not a lot of training data, I picked random subsets of my\ncorpora and got much better behavior (this is rates.py output; f-p rate per\nrun in left column, f-n rate in right):\n\nTraining on Data/Ham/Set1 & Data/Spam/Set1 ... 220 hams & 220 spams\n 0.000 1.364\n 0.000 0.455\n 0.000 1.818\n 0.000 1.364\nTraining on Data/Ham/Set2 & Data/Spam/Set2 ... 220 hams & 220 spams\n 0.455 2.727\n 0.455 0.455\n 0.000 0.909\n 0.455 2.273\nTraining on Data/Ham/Set3 & Data/Spam/Set3 ... 220 hams & 220 spams\n 0.000 2.727\n 0.455 0.909\n 0.000 0.909\n 0.000 1.818\nTraining on Data/Ham/Set4 & Data/Spam/Set4 ... 220 hams & 220 spams\n 0.000 2.727\n 0.000 0.909\n 0.000 0.909\n 0.000 1.818\nTraining on Data/Ham/Set5 & Data/Spam/Set5 ... 220 hams & 220 spams\n 0.000 1.818\n 0.000 1.364\n 0.000 0.909\n 0.000 2.273\ntotal false pos 4 0.363636363636\ntotal false neg 29 2.63636363636\n\nAnother full run with another randomly chosen (but disjoint) 220 of each in\neach set was much the same. The score distribution is also quite sharp:\n\nHam distribution for all runs:\n* = 74 items\n 0.00 4381 ************************************************************\n 2.50 3 *\n 5.00 3 *\n 7.50 1 *\n 10.00 0\n 12.50 0\n 15.00 1 *\n 17.50 1 *\n 20.00 1 *\n 22.50 0\n 25.00 0\n 27.50 0\n 30.00 1 *\n 32.50 0\n 35.00 0\n 37.50 0\n 40.00 1 *\n 42.50 0\n 45.00 0\n 47.50 0\n 50.00 0\n 52.50 0\n 55.00 0\n 57.50 1 *\n 60.00 0\n 62.50 0\n 65.00 0\n 67.50 1 *\n 70.00 0\n 72.50 0\n 75.00 0\n 77.50 0\n 80.00 0\n 82.50 0\n 85.00 0\n 87.50 1 *\n 90.00 0\n 92.50 2 *\n 95.00 0\n 97.50 2 *\n\nSpam distribution for all runs:\n* = 73 items\n 0.00 13 *\n 2.50 0\n 5.00 4 *\n 7.50 5 *\n 10.00 0\n 12.50 2 *\n 15.00 1 *\n 17.50 1 *\n 20.00 2 *\n 22.50 1 *\n 25.00 0\n 27.50 1 *\n 30.00 0\n 32.50 3 *\n 35.00 0\n 37.50 0\n 40.00 0\n 42.50 0\n 45.00 1 *\n 47.50 3 *\n 50.00 16 *\n 52.50 0\n 55.00 0\n 57.50 0\n 60.00 1 *\n 62.50 0\n 65.00 2 *\n 67.50 1 *\n 70.00 1 *\n 72.50 0\n 75.00 1 *\n 77.50 0\n 80.00 3 *\n 82.50 2 *\n 85.00 1 *\n 87.50 2 *\n 90.00 2 *\n 92.50 4 *\n 95.00 4 *\n 97.50 4323 ************************************************************\n\nIt's hard to say whether you need better ham or better spam, but I suspect\nbetter spam <wink>. 18 of the 30 most powerful discriminators here were\nHTML-related spam indicators; the top 10 overall were:\n\n '<font' 266 0.99\n 'content-type:text/plain' 275 0.172932\n '</body>' 312 0.99\n '</html>' 329 0.99\n 'click' 334 0.99\n '<html>' 335 0.99\n 'wrote:' 381 0.01\n 'skip:< 10' 398 0.99\n 'python' 428 0.01\n 'content-type:text/html' 454 0.99\n\nThe HTML tags come from non-multipart/alternative HTML messages, from which\nHTML tags aren't stripped, and there are lots of these in my spam sets.\n\nThat doesn't account for it, though. If I strip HTML tags out of those too,\nthe rates are only a little worse:\n\nraining on Data/Ham/Set1 & Data/Spam/Set1 ... 220 hams & 220 spams\n 0.000 1.364\n 0.000 1.818\n 0.455 1.818\n 0.000 1.818\nraining on Data/Ham/Set2 & Data/Spam/Set2 ... 220 hams & 220 spams\n 0.000 1.364\n 0.455 1.818\n 0.455 0.909\n 0.000 1.818\nraining on Data/Ham/Set3 & Data/Spam/Set3 ... 220 hams & 220 spams\n 0.000 2.727\n 0.000 0.909\n 0.909 0.909\n 0.455 1.818\nraining on Data/Ham/Set4 & Data/Spam/Set4 ... 220 hams & 220 spams\n 0.000 1.818\n 0.000 0.909\n 0.455 0.909\n 0.000 1.364\nraining on Data/Ham/Set5 & Data/Spam/Set5 ... 220 hams & 220 spams\n 0.000 2.727\n 0.000 1.364\n 0.455 2.273\n 0.455 2.273\notal false pos 4 0.363636363636\notal false neg 34 3.09090909091\n\nThe 4th-strongest discriminator *still* finds another HTML clue, though!\n\n 'subject:Python' 164 0.01\n 'money' 169 0.99\n 'content-type:text/plain' 185 0.2\n 'charset:us-ascii' 191 0.127273\n \"i'm\" 232 0.01\n 'content-type:text/html' 248 0.983607\n ' ' 255 0.99\n 'wrote:' 372 0.01\n 'python' 431 0.01\n 'click' 519 0.99\n\nHeh. I forgot all about .\n\n" |
1ham
| easy_ham | "\n> > Note that header names are case insensitive, so this one's no\n> > different than \"MIME-Version:\". Similarly other headers in your list.\n> \n> Ignoring case here may or may not help; that's for experiment to decide.\n> It's plausible that case is significant, if, e.g., a particular spam mailing\n> package generates unusual case, or a particular clueless spammer\n> misconfigures his package.\n\nI found it made no difference for my testing.\n\n> The brilliance of Anthony's \"just count them\" scheme is that it requires no\n> thought, so can't be fooled <wink>. Header lines that are evenly\n> distributed across spam and ham will turn out to be worthless indicators\n> (prob near 0.5), so do no harm.\n\nzactly. I started off doing clever clever things, and, as always with\nthis stuff, found that stupid with a rock beats smart with scissors,\nevery time.\n\n\n-- \nAnthony Baxter <anthony@interlink.com.au> \nIt's never too late to have a happy childhood.\n\n" |
1ham
| easy_ham | "\n> Looks like your ham corpus by and large has To: jeremy@alum.mit.edu in\n> a header while your spam corpus by and large doesn't. But this one\n> does.\n\nInterestingly, for me, one of the highest value spam indicators was\nthe name of the mail host that the spam was delivered to, in the To:\nline. So mail to info@gin.elax2.ekorp.com was pretty much a dead cert\nfor the filters.\n\n\n-- \nAnthony Baxter <anthony@interlink.com.au> \nIt's never too late to have a happy childhood.\n\n" |
1ham
| easy_ham | "\n\n> course, it instantly picked up on 'received:2001' as a non-ham. \n -spam.\n\n*sigh*\n\n\n-- \nAnthony Baxter <anthony@interlink.com.au> \nIt's never too late to have a happy childhood.\n\n" |
1ham
| easy_ham | "There seem to be two \"drivers\" for the classifier now: Neale Pickett's\nhammie.py, and the original GBayes.py. According to the README.txt,\nGBayes.py hasn't been kept up to date. Is there anything in there\nthat isn't covered by hammie.py? About the only useful feature of\nGBayes.py that hammie.py doesn't (yet) copy is -u, which calculates\nspamness for an entire mailbox. This feature can easily be copied\ninto hammie.py. (GBayes.py also has a large collection of tokenizers;\nbut timtoken.py rules, so I'm not sure how interesting that is now.)\n\nTherefore I propose to nuke GBayes.py, after adding a -u feature.\n\nAnyone against? (I imagine that Skip or Barry might have a stake in\nGBayes.py; Tim seems to have moved all code he's working to other\nmodules.)\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |
1ham
| easy_ham | "\nI actually like Neale's X-Spam-Disposition header, I just wonder if maybe we\nshould choose something with a different prefix than \"X-Spam-\" so that\npeople don't confuse it with SpamAssassin, all of whose headers begin with\nthat prefix.\n\nAlso, some sort of version information would be useful. Perhaps the CVS\nversion number for the classifier module could be tacked on.\n\nSkip\n" |
1ham
| easy_ham | "\n Guido> Therefore I propose to nuke GBayes.py, after adding a -u feature.\n\n Guido> Anyone against? (I imagine that Skip or Barry might have a stake\n Guido> in GBayes.py; Tim seems to have moved all code he's working to\n Guido> other modules.)\n\nNo argument here.\n\nSkip\n" |
1ham
| easy_ham | "\nIf the frequency of my laptop's disk chirps are any indication, I'd say\nhammie is about 3-5x faster than SpamAssassin.\n\nSkip\n\n" |
1ham
| easy_ham | "URL: http://boingboing.net/#85534626\nDate: Not supplied\n\nAfter discovering an open wireless net available from his sofa, Steven \n\"Hackers\" Levy interviewed lawmen, academics and WiFi activists about the \nlegality and ethics of using open wireless access points. \n\n I downloaded my mail and checked media news on the Web. When I confessed \n this to FBI agent Bill Shore, he spared the handcuffs. \"The FBI wouldn't \n waste resources on that,\" he sniffed. Now I know that if it did, it would \n be hard to argue that I broke a law. What's more, I certainly didn't feel \n illegal. Because—and this is the point of all that war-driving and \n -chalking and node-stumbling—when you get used to wireless, the experience \n feels more and more like a God-given right. One day we may breathe \n bandwidth like oxygen—and arguing its illegality will be unthinkable. \n\nLink[1] Discuss[2] (_Thanks, Steven[3]!_)\n\n[1] http://www.msnbc.com/news/816606.asp\n[2] http://www.quicktopic.com/boing/H/f5ZpuJU5975K\n[3] http://www.echonyc.com/~steven/\n\n\n" |
1ham
| easy_ham | "I'm listed as a developer on SF and have the spambayes CVS module checked\nout using my SF username, but I'm unable to write to the repository. CVS\ncomplains:\n\n % cvs add unheader.py\n cvs [server aborted]: \"add\" requires write access to the repository\n\nAny thoughts?\n\nSkip\n" |
1ham
| easy_ham | "> Yeah, that's exactly what I was doing--I didn't realize I was\n> incurring administrative pickle bloat this way. I'm specifically\n> trying to make things faster and smaller, so I'm storing individual\n> WordInfo pickles into an anydbm dict (keyed by token). The result\n> is that it's almost 50 times faster to score messages one per run\n> our of procmail (.408s vs 18.851s).\n\nThis is very worthwhile.\n\n> However, it *does* say all over the place that the goal of this\n> project isn't to make the fastest or the smallest implementation, so\n> I guess I'll hold off doing any further performance tuning until the\n> goal starts to point more in that direction. .4 seconds is probably\n> fast enough for people to use it in their procmailrc, which is what\n> I was after.\n\nMaybe. I batch messages using fetchmail (don't ask why), and adding\n.4 seconds per message for a batch of 50 (not untypical) feels like a\nreal wait to me...\n\n--Guido van Rossum (home page: http://www.python.org/~guido/)\n" |