You might have already learnt from other sources that Virus Bulletin is planning to implement anti-spam testing in the near future and I am writing to you to see if there might be any interest to participate in this new and exciting venture with your open source anti-spam product. I have attached by way of information the article published in the January edition of the Virus Bulletin magazine which outlines the proposed test methodology, as well as a recent press release announcing the new tests. Please feel free to pass this information onto any other potentially interested parties within your organisation and if you, or they, should have any questions or feedback please contact Martijn Grooten, Anti-spam Test Director (martijn.grooten@virusbtn.com<mailto:martijn.grooten@virusbtn.com>) for any technical based queries or myself for any financial/administrative queries. Look forward to hearing from you. Best regards Allison Allison Sketchley Sales Executive www.virusbtn.com<http://www.virusbtn.com/> Tel + 44 1235 555139 Fax + 44 1865 543153 ________________________________ Virus Bulletin Ltd, The Pentagon, Abingdon, OX14 3YP, England. Company Reg No: 2388295. VAT Reg No: GB 532 5598 33.
Allison Sketchley schreef:
You might have already learnt from other sources that Virus Bulletin is planning to implement anti-spam testing in the near future and I am writing to you to see if there might be any interest to participate in this new and exciting venture with your open source anti-spam product.
I have attached by way of information the article published in the January edition of the Virus Bulletin magazine which outlines the proposed test methodology, as well as a recent press release announcing the new tests.
Please feel free to pass this information onto any other potentially interested parties within your organisation and if you, or they, should have any questions or feedback please contact Martijn Grooten, Anti-spam Test Director (martijn.grooten@virusbtn.com <mailto:martijn.grooten@virusbtn.com>) for any technical based queries or myself for any financial/administrative queries.
Look forward to hearing from you.
In the proposed test methodology you write the following: "To take part in Virus Bulletin’s anti-spam tests, products must be able to accept SMTP transfers and classify email into two categories: ham and spam." According to the Spambayes FAQ on http://spambayes.sourceforge.net/faq.html, Spambayes is not an MTA (postfix, exchange,...) spam filter. It works by sitting between the MTA or the MDA (procmail,...) and the MUA (outlook, thunderbird, mutt,...) as a POP or IMAP proxy. In case of the Outlook plugin and the upcoming Thunderbird plugin, it even works *after* the MUA. I know that there have been some people who have tried to implement Spambayes as a MTA spamfilter, but these attempts were all very experimental, and they were done by individual Spambayes users. I don't think that there will ever be an "official" build for Spambayes at the SMTP. There is always SpamAssassin for those who want that. However, I am very interested to see how other Software Libre products perform that do work on SMTP level, like SpamAssassin. How well do they compare to commercial products? Disclaimer: I am not a Spambayes developer, I'm just a regular user. By sending your email to a public mailing list, you have agreed that any user of Spambayes may read your email, and that any user may reply to it. Kind regards, Amedee Van Gasse
Dear Amedee, others, Thank you for your email. To answer your question: if I understand it correctly, Spambayes takes an email message (header + body) as input and outputs either 'spam' or 'ham'. (I know many products actually output a likeliness percentage, but by setting a threshold this has the same effect.) This does not necessarily have to be a problem: we could possibly install Spambayes on the server which redistributes the email and, once an email is redistributed, make it check whether it thinks the email is ham or spam and store that in the database. Note, however, that for various reasons, we do not 'teach' the filters by providing them with end user feedback. On the (separate) test where the filters are exposed to the stream of a large spamtrap, we could use any free SMTP daemon. I assume that the actual daemon used should not matter for its performance. I do have one concern though: we are happy to include free, open source products in our test (and, unlike commercial products, will test them for free). However, we want to be sure that the developers agree on submitting the product. In particular, we don't want to be the source of an argument among developers about submitting a product. Best Regards, Martijn Grooten Anti-spam Test Director www.virusbtn.com Email: martijn.grooten@virusbtn.com Tel: +44 1235 540235 / +44 7872 674989 -----Original Message----- From: Amedee Van Gasse [mailto:amedee@amedee.be] Sent: 13 January 2009 22:55 To: Allison Sketchley Cc: spambayes@python.org; Martijn Grooten Subject: Re: [Spambayes] Virus Bulletin Anti-spam Testing Allison Sketchley schreef:
You might have already learnt from other sources that Virus Bulletin is planning to implement anti-spam testing in the near future and I am writing to you to see if there might be any interest to participate in this new and exciting venture with your open source anti-spam product.
I have attached by way of information the article published in the January edition of the Virus Bulletin magazine which outlines the proposed test methodology, as well as a recent press release announcing the new tests.
Please feel free to pass this information onto any other potentially interested parties within your organisation and if you, or they, should have any questions or feedback please contact Martijn Grooten, Anti-spam Test Director (martijn.grooten@virusbtn.com <mailto:martijn.grooten@virusbtn.com>) for any technical based queries or myself for any financial/administrative queries.
Look forward to hearing from you.
In the proposed test methodology you write the following: "To take part in Virus Bulletin’s anti-spam tests, products must be able to accept SMTP transfers and classify email into two categories: ham and spam." According to the Spambayes FAQ on http://spambayes.sourceforge.net/faq.html, Spambayes is not an MTA (postfix, exchange,...) spam filter. It works by sitting between the MTA or the MDA (procmail,...) and the MUA (outlook, thunderbird, mutt,...) as a POP or IMAP proxy. In case of the Outlook plugin and the upcoming Thunderbird plugin, it even works *after* the MUA. I know that there have been some people who have tried to implement Spambayes as a MTA spamfilter, but these attempts were all very experimental, and they were done by individual Spambayes users. I don't think that there will ever be an "official" build for Spambayes at the SMTP. There is always SpamAssassin for those who want that. However, I am very interested to see how other Software Libre products perform that do work on SMTP level, like SpamAssassin. How well do they compare to commercial products? Disclaimer: I am not a Spambayes developer, I'm just a regular user. By sending your email to a public mailing list, you have agreed that any user of Spambayes may read your email, and that any user may reply to it. Kind regards, Amedee Van Gasse Virus Bulletin Ltd, The Pentagon, Abingdon, OX14 3YP, England. Company Reg No: 2388295. VAT Reg No: GB 532 5598 33.
Martijn Grooten schreef:
Dear Amedee, others,
Thank you for your email.
To answer your question: if I understand it correctly, Spambayes takes an email message (header + body) as input and outputs either 'spam' or 'ham'. (I know many products actually output a likeliness percentage, but by setting a threshold this has the same effect.)
Spambayes outputs a number between 0.00 and 0.99.
This does not necessarily have to be a problem: we could possibly install Spambayes on the server which redistributes the email and, once an email is redistributed, make it check whether it thinks the email is ham or spam and store that in the database. Note, however, that for various reasons, we do not 'teach' the filters by providing them with end user feedback.
Teaching is essential for Spambayes. It simply cannot work if it doesn't get a statistically relevant corpus of ham and spam to train on. Without training, all emails will get a score of 0.50. I don't want to speak on behalf of the developers, but I'm afraid that Spambayes falls outside the rules of your test. Kind regards, Amedee Van Gasse
Amedee> I don't want to speak on behalf of the developers, but I'm Amedee> afraid that Spambayes falls outside the rules of your test. Keep going. You seem to be doing a very good job. ;-) -- Skip Montanaro - skip@pobox.com - http://smontanaro.dyndns.org/
participants (4)
-
Allison Sketchley -
Amedee Van Gasse -
Martijn Grooten -
skip@pobox.com