- From: Peter Nunn <peter.nunn@vistic.net>
- Date: Sat, 21 Apr 2007 07:55:11 +1200
- To: "Klotz, Leigh" <Leigh.Klotz@xerox.com>
- CC: www-forms@w3.org
- Message-ID: <46291A9F.5060808@vistic.net>
Thanks leigh, I am aware of the man-in-the middle hazard. The encrypt function is only one part of the solution. The problem is that just saying 'use https' has in my mind the following issues: 1. It assumes the xforms is used as part of a web page. 2. It ignores the use of transports over smtp or some form of message queueing. 3. It ignores offline forms where the form may be stored on a device that could be lost/stolen and have fragments of sensitive information within them. 4. Most security breaches occur inside an organization, not from people inside a telecom company. HTTPS is generally used on the internet and then the form/submission etc is silently processed inside a company from where most information such as credit card numbers/passwords etc are stolen. So my question is about providing the ability to encrypt the contents of an xml fragment, or instance using xml:encryption, not using it it replace https or any other internet based security protocol. How the encrypt function works with the certificate store on the browser was not part of the question. I would expect (and this is how I am doing it in MozzIE) is that if a certificate is provided for say an RSA encrypted element and the certificate is not from a trusted authority and it is not provided in a means that can be authenticated, the user is warned. This uses the same authentication mechanisms as used by https to authenticate the digital certificates. Hence my question, about whether the wg has considered an encrypt function along with all of the issues above, as it is an important function of xforms unless of course the intent is to only have xforms transmitted over the Internet, and only securable using https. If it has not been considered, then I would like to start a dialog on the issue for consideration of inclusion into a future xforms spec, otherwise it will have to fall into the catch-all extension function bucket. regards
Received on Friday, 20 April 2007 19:55:11 UTC