ietf
[Top] [All Lists]

Re: Eliminating Virus Spam

2001-01-04 10:30:02
From: "Matt Crawford" <crawdad(_at_)fnal(_dot_)gov>

Please point to an example of a useful multipart message seen in
this list or that might someday be useful in this mailing list.

I have sent to wg lists a multipart containing a preamble and an
internet-draft or similar file.  This makes it easy for recipients to
save the draft as-is.  Sometimes I have made the file a
content/external-body to avoid stuffing O(100kB) into many mailboxes
where it wouldn't be wanted.

You said "this list", but the above scenario would apply if the
document were pertinent to a BOF or brand-new working group.

I'm having trouble parsing that.  
I agree that depending on the I-D and the WG, sending I-D's to working
group lists is sometimes a good thing.  I don't think MIME is required
for that and wouldn't choose it myself, but I agree a working group could
agree to tolerate or prefer it.  However, sending I-D's to this list is
not only not kosher, but an anathema.

Sending content/external-body instead of the 100KB thing itself anywhere
could be swell, except that the state of word puts it somewhere between
irrelevant and a bad idea:
  1. many MUA's automatically dereference all pointers they find anywhere,
     whether in separate MIME attachments or in-line
  2. many slimey outfits use that fact with "web bugs" for violating privacy,
     and have convinced other outfits to help them do that (e.g. my
     soon-to-be-ex on-line stock broker)
  3. so anyone with any sense turns off all automatic dereferencing and
     uses the manual hooks to #1, and that makes content/external-body
     irrelevant.

In other words, who has the least trouble dereferencing a URL in a 7-bit
ASCII message without a single MIME header?

   ...........


] From: James M Galvin <galvin(_at_)acm(_dot_)org>

] ...
] To choose to filter content (restrict it to text/plain or some other
] limited set) would be changing the policy of this elist.  That is more
] than an operational decision and probably deserves more debate, although
] not here.

Someone should tell Mr. Galvin that this is not a "general elist," although
it has been identified by some as a haunt of "elitist legacy programmers."

] ...
] viruses, but it is the application data that is the virus not the OS.
] Microsoft Word macros can be run on UNIX and text/html with javascript
] runs virtually anywhere.

That seems to be based on a non-standard definition of "virus."  Neither
MS Word macros nor text/html javascript are likely to be infectious on
reasonable or even most UNIX boxes.  Just running MS Word macros on most
UNIX boxes is between difficult and impossible.  (Yes, I've heard about
StarOffice.)

] ...
] A virus scanner scans content, a sequence of bytes.  It can do this on
] any platform, regardless of the intended destination platform of the
] content.

If wishes were horses then beggars would ride.
Or what MIME attachment virus scanners are there for NetBSD?

]           The virus may be platform specific but that is irrelevant.

That sounds reasonable only to members of the Church of Redmond.


Vernon Schryver    vjs(_at_)rhyolite(_dot_)com



<Prev in Thread] Current Thread [Next in Thread>