[2796] in WWW Security List Archive
Re: A problem with Navigator's cache -Reply
daemon@ATHENA.MIT.EDU (hallam@ai.mit.edu)
Sun Aug 25 20:44:17 1996
From: hallam@ai.mit.edu
To: jsw@netscape.com, www-security@ns2.rutgers.edu
Cc: hallam@ai.mit.edu
Date: Sun, 25 Aug 96 19:27:07 -0400
Errors-To: owner-www-security@ns2.rutgers.edu
> Since Marc and Kipp have not been involved in our security
>efforts for at least a year and a half (since I arrived here),
>I don't understand why your experiences with them so long ago
>are relevant to our current security efforts.
Last I heard Marc was your Senior VP in charge of Technology.
When I raised the point about wanting to turn off Java and
Javascript someone at Netscape made a very pointed reply about
the user being able to do this but never responded to my point
that as security officer it is my decision.
Another problem on the horizon is that we do not currently have
a security solution that interacts with firewalls. As one
manager put it to me "SSL will not go through our firewall
period.". Basically an opaque encryption protocol conflicts
with the primary role of a firewall - restricting outgoing
bandwidth.
Mind you I don't think that implementing S-HTTP is much fun
either :-(
>You can also change the User-Agent field
>to indicate that this is your locked version.
Ah thats fair enough.
It would be nice if this was incorporated into some sort of
client capabilities field so that servers generally were aware
of which features were available. Dave Ragget has been trying
to get a proposal of this sort implemented for a very long time
but with little support.
One of the features that worries me about Java generally is that
there does not seem to have been much thought about how to upgrade
the virtual machine. This is of particular concern to language
designers like myself for whom the Java VM is inadequate to
implement language models with richer inheritance features.
Although I have never believed that the mobile code concept that
Java initially took wing on was ready for prime time, I like
many others was willing to give Java a nudge in the hope that
it could kill C++ which seemed like it was about to take over
the world, even though nobody actually liked it. Now that C++
is competing against Java on its merits I think Java is the
clear leader, if I was to teach a programming course for non
specialists it would be my first choice as a teaching language
since it is the first language with clean semantics that has
succeeded commercially.
This brings a problem for the Java VM, the Redmont club having
built their way into OLE, COM, DCOM etc have heavilly invested
into a multiple inheritance model that Java VM does not support.
Given that Java represents the best chance to clean up the resulting
mess and produce a programming environment for windows that is
usable the commercial and technical pressures are likely to force
certain extensions on the VM.
Tagging which VM a browser supports would seem to me to be a
good start.
Phill