- From: Jon Ferraiolo <jferrai@us.ibm.com>
- Date: Fri, 6 Jun 2008 08:28:04 -0700
- To: Web Application Formats Working Group WG <public-appformats@w3.org>
- Cc: public-appformats@w3.org, public-appformats-request@w3.org
- Message-ID: <OF1944C18A.0C359958-ON88257460.00542D46-88257460.0054F7D9@us.ibm.com>
(Re-expressing what I have said before multiple times)
The extra complexity with opting in to cookies is worth it. I realize that
(as has been pointed out) servers really need to implement CSRF protection
regardless of Access Control, but AC has the potential of greatly
increasing community awareness and adoption of enabling cross-site
requests, which not only means more novice developers (who don't know about
CSRF) adding AC support to their Web site, but also more hackers who will
scour the Web looking for sites that are vulnerable to CSRF attacks. It is
much better to not transmit cookies by default and require that the server
developer do a little extra work (presumably one additional header or
slightly different markup within an existing header) to trigger the
transmission of cookies. Also, don't forget to put STRONG WARNINGS in the
spec in the section that explains how to activate transmission of cookies
(or opt-in to other headers).
Jon
Web Application
Formats Working
Group Issue To
Tracker <sysbot public-appformats@w3.org
+tracker@w3.org> cc
Sent by:
public-appformats Subject
-request@w3.org ISSUE-26: Opting into cookies
[Access Control]
06/06/08 05:23 AM
Please respond to
Web Application
Formats Working
Group WG
<public-appformat
s@w3.org>
ISSUE-26: Opting into cookies [Access Control]
http://www.w3.org/2005/06/tracker/waf/issues/
Raised by: Anne van Kesteren
On product: Access Control
It has been suggested that because Access-Control also allows read access
and not just making the request explicit optin into cookies specifically is
desired.
The benefit would be that servers can more safely enable Access Control
functionality.
The drawback would be that the model becomes more complicated and therefore
more prone to errors and implementation bugs.
Attachments
- image/gif attachment: graycol.gif
- image/gif attachment: pic03849.gif
- image/gif attachment: ecblank.gif
Received on Friday, 6 June 2008 15:31:18 UTC