W3C home > Mailing lists > Public > w3c-ietf-xmldsig@w3.org > July to September 1999

RE: Re[2]: How to sign several resources (XML and XSL)?

From: Ed Simon <ed.simon@entrust.com>
Date: Thu, 23 Sep 1999 13:37:50 -0400
Message-ID: <01E1D01C12D7D211AFC70090273D20B105E77B@sothmxs06.entrust.com>
To: "'rhimes@nmcourt.fed.us'" <rhimes@nmcourt.fed.us>, miltonma@gte.net, david.burdett@commerceone.com, jevans@differential.com, winchel@mindspring.com, andreas.siglreithmayr@ixos.de, w3c-ietf-xmldsig@w3.org
Cc: ietf-trade@lists.eListX.com
Just as signing an entity like "&bank;" may not be sufficient
when the user actually sees "Bank of Bells Corners", signing
a link to stylesheet may also be insufficient.

Though an XML instance can specify one or more stylesheets
(using the Recommendation Rich links to), that does not
mean applications will actually
use one of those stylesheets when formatting the
data for a user.  For example, it is predicted that PCs will
make up less than half of the devices connect to the Internet
by 2002 (others would include handheld devices like PalmPilots
and cell-phones (with mini screens like those on modern
video cameras).  To accomodate the wide range of output devices,
vendors are looking at dynamically selecting the stylesheet
based on the output device.  So, if you want to trade stocks
using a cell-phone's mini screen, the stock-trade-to-cell-phone
stylsheet would be used.  An application may then want to sign
that stylesheet used for that transaction rather than one of the
ones specified in the XML instance.

Anyway, I like the way Rich makes the point about the
"level of trust" being up to the application.  The XML
Signature spec may make it possible to sign things beyond
the core data but doesn't require one do so.  Applications
should sign whatever they believe is necessary.  For
legally-binding signatures, guiding principles might be
"beyond a reasonable doubt" or
"on the balance of probabilities".

Ed
-----Original Message-----
From: rhimes@nmcourt.fed.us [mailto:rhimes@nmcourt.fed.us]
Sent: September 23, 1999 12:42 PM
To: miltonma@gte.net; david.burdett@commerceone.com;
jevans@differential.com; winchel@mindspring.com;
andreas.siglreithmayr@ixos.de; w3c-ietf-xmldsig@w3.org
Cc: ietf-trade@lists.eListX.com
Subject: Re[2]: How to sign several resources (XML and XSL)? 



XSL is an XML format, so we haven't excluded signature of XSL, have we?
Also,
I'm aware of the W3C recommendation at

http://www.w3.org/TR/xml-stylesheet/

that ties a stylesheet to an XML document.  However, a processing
instruction is
used for this purpose, and we have proposed elimination of PI from our c14n.
I
believe this needs to be revisited.  Applications will definitely want the
option to locate and sign an associated stylesheet.  IMO, It is not for us
to
judge whether or not this makes sense to the application just because our
committee(s) can't guarantee a trusted browser.  The level of trust required
is
up to the application (they may restrict their application to specific
signed
and trusted browsers, they may certify their application for specific common
browsers, etc.)

Rich

____________________Reply Separator____________________
Subject:    Re: How to sign several resources (XML and XSL)? 
Author: "Milton M. Anderson" <miltonma@gte.net>
Date:       9/23/99 7:44 AM


-----Original Message-----
From: David Burdett <david.burdett@commerceone.com>
Date: Wednesday, September 22, 1999 11:39 PM
Subject: RE: How to sign several resources (XML and XSL)?


>... otherwise how do you know the context for the XML data?
>
>You know the context because interpretation of the XML data is being done
by
>software from a presumably reliable source to do the interpretation that is
>built according to a specification that describes the semantics of data
>
>... I now feel that we're getting very close to the topic of "trusted"
>applications  and I'm not sure we want to go there ...

Even if XSL is signed, you still have to assume a "trusted" browser.
It's impossible not to go there...

Milt




 
Received on Thursday, 23 September 1999 13:38:01 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.29 : Thursday, 13 January 2005 12:10:07 GMT