Re: ACTION-640: clarification to interaction model?

Good point. I agree that the current text does not define what is considered correct behavior when content is not retrieved over HTTP. In the example you mention where some resources are retrieved over FTP, a browser could be considered compliant if it reported that website as TLS-secured.

If I am reading that correctly and we are ok with that interpretation, I have no further objection.

Joe

On Apr 9, 2010, at 3:54 PM, Thomas Roessler wrote:

Joe Steele wrote:
The 4.1 proposed text is fine.

I can see your point on the change to the first two paragraphs, but I still think something more is needed to narrow the scope. The text in 5.3 refers to "all content" and "all other resources". It does not acknowledge that some content might not be coming through an HTTP transaction. That is the root of the problem. We need to remove that "all" qualifier and replace it with something more narrow. Any suggestions as to what? Providing some explanatory text elsewhere doesn't seem to solve the problem.


The change that you suggest here has an interesting side effect: If, for
example, a resource that's part of a web page is retrieved through FTP
(or some future insecure network protocol), then that resource's
security wouldn't matter for the determination of mixed content.  In the
text as currently written, one could argue that the behavior is
undefined -- which is entirely fine for the purposes of this specification.

That's why I'm extremely reluctant to do the tighter scoping in the way
in which you suggest it.


Joe

On Apr 9, 2010, at 5:34 AM, Thomas Roessler wrote:


Joe Steele wrote:

I would propose this change to paragraph 1 in section 5.3:
If a given Web page consists of a single resource only, then all
content *<change>retrieved through an HTTP transaction</change>* that
the user interacts with has security properties derived from the HTTP
transaction used to retrieve the content.
And similar changes to the two following definitions:

[Definition: A Web page is called *TLS-secured* if the top-level
resource and all other resources that can affect or control the page's
content and presentation *<change>and are retrieved through an HTTP
transaction</change>  *have been retrieved through strongly TLS
protected HTTP transactions.


Looking at these two definitions, the changes seem tautological,
therefore -1.

However, I'd suggest we say this in 4.1, after the fourth paragraph: "In
interactive Web applications, the presentation to the user might also
depend on state that is local to the client - be it local storage of
structured data, or be it recent interactions with the Web page. The
security properties of those data will depend on the security properties
of the client computer itself, and are out of scope for this specification."

Thoughts?


[Definition: A Web page is called *mixed content* if the top-level
resource was retrieved through a strongly TLS protected HTTP
transaction, but some dependent resources were *<change>retrieved
through a weakly protected or unprotected HTTP transaction</change>*]


+1