lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: LYNX-DEV lynx-dev list status/problems


From: Al Gilman
Subject: Re: LYNX-DEV lynx-dev list status/problems
Date: Mon, 14 Jul 1997 11:37:45 -0400 (EDT)

[Tom Z. is concerned about spammers mining his edress out of the
lynx-dev archives.]

Tom,

I approach this in terms of immediate and continuing time-frames.

IMMEDIATE:

I would like to see the list revert to open posting, i.e. the
From: and Sender: headers don't figure in message acceptance for
distribution, outside of some possibility of domain blacklisting,
but I don't think we will need to filter by domain.

So long as this is the case, you can at least use the spambait@
address to post from and possibly reduce your exposure to spammers.

CONTINUING:

Filter rules: 

Any spam that I get these days has invoked my edress in a Bcc:
header.  It doesn't appear in either To: or Cc: headers.  This is
my idea of the first filter we should apply to messages which
arrive at SigNet for delivery to lynx-dev.  This is a rule that
messages are suspect unless they contain either "lynx-dev" or
"address@hidden" in their To: or Cc: header value.  I am hopeful that
that kind of rule can be done either in standard MajorDomo
configuration media or via Perl patched into MajorDomo that
SigNet will tolerate (and MajorDomo will want to incorporate).

I would consider it up to Bob and SigNet to decide whether they
would agree to patch MajorDomo at all or whether they would allow
the lynx-dev login to run Procmail to perform this filtering if
MajorDomo can't do it.  

There are other flows that are possible if MajorDomo does not
comprehend open posting as an option.  We could create an
auto-approve agent by a login which runs Procmail and is known to
Majordomo as an approval authority.  There are many ways to skin
a cat.

I would still have the messages that get kicked out by automatic
screening be reviewed by a human (team).  But I want to reduce
the set of bonafide walk-up questions that get sidetracked for
clearance reduced to a small fraction of the total walk-up
traffic.  It seems that with current spamming practices we can
sidetrack most spam and not sidetrack most posts from strangers
at the same time.

Edress exposure at the archive:

Any sweep that comes across my FAQ gets into the message pages of
the archive, and could crawl from there to the indices and all
messages.  Putting a form at the front door doesn't mean that
there aren't back doors.

Putting a form at the front door makes a significant difference
to the ease of use by spambots (good) and blind Internauts (bad).
Really buggering up access to the archive such that messages
there could not be bookmarked would hurt its use as a Help annex.

I am still interested in having the archive offer the opportunity
of follow-ups that get threaded right.  This is a topic in work
with the WAI and not just lynx-dev.  So if there are spam-related
downsides I want to understand them and try to find a balanced
response.  Since the critical form of follow-up is to the list
address, maybe we should only put that mailto: in the pages
and leave it to people who post but don't subscribe to either

        - read follow-ups via the threading in the archives, or
        - manually insert their email and request a manual Cc:

--
Al Gilman
;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;

reply via email to

[Prev in Thread] Current Thread [Next in Thread]