MozillaZine

If I offered money for a Mozilla build, would anyone accept

Discussion about Seamonkey builds
drosoph
 
Posts: 147
Joined: November 4th, 2002, 10:28 pm

Post Posted November 5th, 2002, 3:31 pm

Ive been wanting to get my hands on a Spam-filtering copy of Mozilla (Built with the code from Bug # 169638) ... If I offered up 20$ for a win32 build ... would I get any takers ???

Or maybe, just maybe ... one of you generous souls would even be willing to give me a copy for free :) ...

I hate SPAM THAT much ! ... Im one of the lucky ones ... getting over 1500 spam per day :) ... (lots of accounts, domains, etc) ...

Kommet

User avatar
 
Posts: 112
Joined: November 5th, 2002, 1:15 pm
Location: San Jose, CA

Post Posted November 6th, 2002, 12:42 am

If I had the resources available I'd compile with the patch for free and post it somewhere.

Unfortunately, by the time I pulled the source in over my modem and compiled on my slow-ass machine (a few times as I'm sure I'd miss a config or two the first time) the pre-1.3a builds will be available.

The good news is that this bug is supposed to have already landed a patch, and the patch is only waiting for 1.3 to open. Since the branch just happened and the trunk is open again for 1.3 development work, you should have SPAM filtering to test in one of the very next builds.

Kommet

User avatar
 
Posts: 112
Joined: November 5th, 2002, 1:15 pm
Location: San Jose, CA

Post Posted November 6th, 2002, 1:13 am

Whoops. Looks like there are several more steps to take. <a href="http://bugzilla.mozilla.org/show_bug.cgi?id=169638" title="front end work for spam / junk mail">The bug you referenced</a> seems to deal only with the spam-filtering UI, not the backend. Reading a <a href="http://bugzilla.mozilla.org/show_bug.cgi?id=163188" title="Add Bayesian antispam filters per Paul Graham's design">bug dealing with Bayesian filtering</a> indicates that the actual engine work is still in progress.

<a href="http://bugzilla.mozilla.org/show_bug.cgi?id=163188#c63">Comment 63</a> in the backend bug points to a <a href="http://spambayes.sourceforge.net/" title="SpamBayes: Bayesian anti-spam classifier written in Python.">project at SourceForge</a> where Bayesian filtering has been investigated and even alludes to some heavy lifting done already in the R&D arena to refine the <a href="http://www.paulgraham.com/spam.html" title="A Plan for Spam">original proposed algorithm</a> and <a href="http://radio.weblogs.com/0101454/stories/2002/09/16/spamDetection.html" title="Spam Detection">subsequent rebuttal piece</a>. Looks like the folks at SF are looking at a much more detailed statistical sampling method which better determines what is spam, what is not spam, and what it simply cannot make a judgement on.

I assume that we won't get REAL spam filters until a backend gets checked in for the frontend, and from the way the comments sounded, that will be a little while longer while they look at algorithms.

Return to SeaMonkey Builds


Who is online

Users browsing this forum: No registered users and 1 guest