At 11.07 -0800 0-05-07, ned(_dot_)freed(_at_)innosoft(_dot_)com wrote:
Well, I was there, and I question the validity of your assessment of what was
going on. While it is true that there was a clear concensus opposed to adding
wiretapping facilities in the RAVEN sense, it was by no means 95-98 percent.
Perhaps I misunderstood the question being asked at the
meeting. I understood the question to be if we wanted to
develop protocols to help police trace net villains, you
understood it to be more restricted in only helping police
perform viretapping.
At 21.39 -0700 0-05-07, James P. Salsman wrote:
I fully agree and have decided to sponsor a contest to correct
the situation. I will give one share of Microsoft stock to the
first person who posts, to this IETF Discussion list, a draft
shareholder resolution that would, in the opinion of Keith Moore
or his designated alternate, correct the situation if it were
adopted by Microsoft Corporation as we currently know it.
Certainly, Microsoft software could be designed to make it more
difficult for virus spreading. However, the villains will learn
to get around such features. Compare with spammers. A few years
ago, you could easily detect spamming by checking if a message
did not come from a mailing list you subscribed to, and did not
have your name in any of the recipient field. Today, more and
more spam messages even contain your name in the text in some
kind of greeting "Hella Jacob", which obviously was put there
to confound spam checkers who detect spam by checking if many
identical messages are sent.
At 06.38 -0400 0-05-08, Garreth Jeremiah wrote:
The "Java" sandbox idea in my mind is a great one.
My experience is that almost ever where I wanted to do something
useful with applets, what I wanted to do was forbidden by
security restrictions.
At 15.05 +0200 0-05-08, Magnus Danielson wrote:
What you really would like to have is a common
accept/deny type of list. This would trim down the
required OK's quite alot. Those which are on the deny list
would be silently denied and those on the accept list
would be silently accepted. Only those not existing on
either of the lists would actually require manual
intervention in approving.
This will only work if the identity of the allowed senders
was identified with crypthographic methods. Otherwise,
the virus senders will find ways to make believe being
the people you trust.
---
Methods helping the police track virus makers:
(1) Making software more restrictive in accepting foreign
code. Comment: Will help, unless the virus producers
learn to circumwent it. Hass the risk of making life
for ordinary legal users more difficult.
(2) Improve (1) with strong crypthographic methods to
identify trusted senders. Comment: A promising method,
if only strong crypthographic methods get commonly
used. Note however, those of you who want to
protect anonymity: Strong crypthographic methods
are methods to identify people securely, not methods
to allow people to be anonymous.
(3) Tracing and logging feature to find out where the
virus came from. Comment: Virus makers will certainly
try to cheat such systems by incorrect identification
such as senders IP address. But I still believe this
is one of the most promising methods.
(4) Sandbox environments for executing possibly dangerous
code. Comment: Every good programming language should be
designed as a "virtual machine" where a program, when
executed, cannot do anything outside of this protected
environment. I wrote this already in a paper published in
Datamation, December 1975, pp 77-80, with the title
"Languages for Reliable Software". However, the safest
sandboxes are also those most restrictive against doing
legal things well.
(5) Create anti-bodies which scan incoming data and detect
known viruses. This is the main methods of the anti-
virus software sold today. It is, however, becoming
more difficult since the number of viruses is getting
so large that the anti-body creators have problems
keeping up with it.
I do not think this is an either/or situation. To stop
the proliferation of viruses, we should do all of this.
And IETF can certainly help, by designing methods to
support all of these anti-virus activities.
I do not think we can ever stop people from producing
viruses. If, however, we make the risk of getting caught
large enough, most of them will find other methods of
venting their anger at society, like stealing hubcaps
or crashing windows.
There is an obvious conflict between anonymity, privacy,
and detecting criminal behaviour. Different people position
themselves at different places on this scale, but you
cannot deny that the conflict exists. Crime is much
more common in urban than rural areas - just because
people are easier anonymous in the urban areas.
--
Jacob Palme <jpalme(_at_)dsv(_dot_)su(_dot_)se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/