Re: [ietf-types] Update to text/html registration

["Michael[tm] Smith" <[e-mail]>]

Mark Baker <[e-mail]>, 2012-08-07 10:00 -0400:

> Hi Mike,
> 
> I left the WG when HTML5 was deliberately backwards-incompatible,

Not sure what in particular you thought was deliberately backwards-
incompatible, but OK.

> making the update of text/html an error. AFAICT, this appears to have
> since been addressed, at least judging by the text here;
> 
> http://www.w3.org/TR/html5-diff/#backwards-compatible

I think that's just stating something that's always been true. The HTML5
spec has always intended to be backwards-compatible for existing browser
behavior. That has been the number one design principle of this entire
effort from day one:

  http://www.w3.org/TR/html-design-principles/#support-existing-content

> I recommend that the interoperability considerations section mention
> this. Something as simple as;
> 
> "HTML5 is backwards compatible with all previous versions of HTML as
> implemented. See
> http://www.w3.org/TR/html5-diff/#backwards-compatible"

It's not intended to be "backwards compatible with all previous versions of
HTML". Judging whether anything is compatible with HTML4 has always been a
guessing-game exercise in creative interpretation. The HTML4 spec does not
actually state user-agent requirements in such as way that it's possible to
evaluate or test conformance to the HTML4 spec.

> I included "as implemented" because I know the editors were always
> keen to point out where multiple independent implementations strayed
> from the specifications (and rightly so).

Outside of those differences, the HTML5 spec has always intended to be
compatible with existing behavior in browsers, which is generally not open
to being changed. So I'm not sure I'd actually call that "backwards"
compatibility. It's more like just compatibility with the Web.

Anyway, as a I pointed out to Paul, the details for this registration are
defined in the HTML5 spec itself. So if you want your proposed wording
added, I suggest you raise a bug against the HTML5 spec:

  http://w3.org/brief/MjA2

-- 
Michael[tm] Smith http://people.w3.org/mike