<blockquote style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;" class="gmail_quote"> "An agent MUST NOT ignore or override authoritative metadata<br></blockquote><div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
without the consent of the party employing the agent."<br><br>is a good thing, but please don't claim by having this default to true<br>constitutes consent.<br><br>Best regards, Julian</blockquote><div><br>Prompting a user for any sort of consent would be useless and confusing, because users don't know what MIME types are. Even a dialog that says "This document claims to be plain text, but looks like a hypertext document. Do you want to render it as a hypertext document?" would be useless and confusing because, frankly, users don't know the difference between plain text, web pages, and Microsoft Word.
<br><br>So, if you interpret that "consent" to mean some sort of explicit interaction by the user after the User Agent is installed, then that clause is impossible to satisfy in any useful way. For the purpose of this discussion and this section of HTML5, "consent" should be defined as Anne said - by using a browser, you consent to it being useful. But there's an option to make it less useful if you want.
<br><br>Logging Content-type discrepencies in an error console should be enough to satisfy the other requirement in section 4.<br><br>I understand your concern. You want authors to correct mistakes in their code (and server configurations) to comply with standards. Authors should be encouraged to do so, but only in ways that are not detrimental to end users. End users don't make a good middle man for telling an author when his code doesn't agree with a spec. End users tend to blame the browser first.
<br></div></div><br>-- <br>Jon Barnett