Based on the forwarded message below, I think it is necessary that you
inform list subscribers that the list is to be publicly archived at a
place where their addresses will be in full view for crawlers, in
particular to email address collectors that sell the addresses to
spammers and whatnot.
In contrast, at mail-archive.com, they do everything they can to
1) hide email addresses
2) disallow crawlers.
In general, the decent thing to do is first discuss it with existing
subscribers before you channel the lyx lists to newsgroups and
mailarchives. In particular, let them know the risks as far as
spamming and viruses are concerned.
I do not see any messages at the devel or users' list about this
archiving at marc.theaimsgroup.com. (My grep for theaims turned up
nothing on the list archives.)
Also, the lyx webpage may want to discuss the risks of subscribing to
the lyx lists.
Let me know the result of the discussion; then I will act according to
the result.
Mate
----- Forwarded message from Hank Leininger <[EMAIL PROTECTED]> -----
Date: Tue, 18 Jun 2002 02:06:01 -0400 (EDT)
From: Hank Leininger <[EMAIL PROTECTED]>
Subject: Re: [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
To: <[EMAIL PROTECTED]>
cc: John Levon <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]>,
<[EMAIL PROTECTED]>, <[EMAIL PROTECTED]>
On Mon, 17 Jun 2002 [EMAIL PROTECTED] wrote:
> On Sun, Jun 16, 2002 at 03:08:00PM -0400, Hank Leininger wrote:
[snip]
> > Could you guys disable the X-No-Archive setting for the lyx lists?
>
> I'll address this tomorrow. What I do for the existing achiving at
> mail-archive.com is I set up a sublist which sends mail only to the
> webarchive---and I remove the header X-No-Archive only for the sublist.
Ah, cool. OK. Currently our archives are subscribed as
'lyx-*@progressive-comp.com' (should be the only @progressive-comp.com
address on each list). I guess you could move those subscriptions to
the non-header'ed sublist?
> What do I need to do to have the sublists send mail to your archive?
> Do you guys do something against email and other address collector
> software? Do you hide addresses?
Good question.
We don't do a whole lot. Since most of our list archives are
unofficial, we've never done any address-munging--for one, people could
then complain that they were being quoted w/o proper attribution--and
sue for copyright violations. (This has actually happened, though not
to us.) For list-owners I think they can safely set whatever policy
they want-- by joining and posting to a list a participant agrees to
explicit and implicit list terms. But for third parties it's stickier
:-P Also, MARC currently knows about >890,000 authors (from about 9.2
million emails). Since we do make authors searchable (no wildcards
though), and trackable (click on an author and see whatever else they've
posted that we've got), chopping addresses before the @ would cause lots
and lots of collisions and therefore misattributions (the John Wilson
problem[1]).
But anyway, as for what we do do: the interface tries to be harvester-
unfriendly. Any view which lists multiple people (lists of messages,
lists of threads, lists of authors, lists of search results, etc) do not
display full addresses, only the comment, lhs, whatever. Individual
messages have the full sender (and any email addresses within the body
are unmolested), though. However message IDs and author IDs are not
sequential[2], so one can't simply iterate through ++'ing each time
harvesting addresses. We also monitor request volume/rate per source,
and throttle (or simply block) people running unfriendly robots (this is
not yet as automated as it should be). Sure, you could write something
a bit more intelligent which walks the site correctly and slowly enough,
but at that point I think there are far, far easier ways to harvest
addresses elsewhere. If that did start to happen it should still be a
noticable blip in the monthly usage stats, so it'd be apparent this was
no longer enough.
Of course, what brought this all up is: we honor X-No-Archive's, so if
an individual chooses to use them, we'll silently skip over their
messages. Until someone replies and quotes them and includes their
entire mail including .sig without any trimming :(
Let me know if that makes you feel reasonably comfortable; I'd
understand if it didn't.
Thanks,
Hank Leininger <[EMAIL PROTECTED]>
E407 AEF4 761E D39C D401 D4F4 22F8 EF11 861A A6F1
[1] http://marc.theaimsgroup.com/?l=cryptography&m=102269473203419&w=2
[2] With the exception of bulk-added mails; message IDs are sequential
for batches of mail inserted together :( Author numbers are still
not, however. I plan to move to md5sum-derived external keys for
most things, which will address this along with many other things.
But no promises, or even wild guesses about when that will happen.
----- End forwarded message -----
--
---
Mate Wierdl | Dept. of Math. Sciences | University of Memphis
