Re: Backward-compatibility of text/html media type (ACTION-334, ACTION-364)

On Tue, 02 Feb 2010 18:13:07 +0100, <noah_mendelsohn@us.ibm.com> wrote:
> Anne van Kesteren writes
>> On Tue, 02 Feb 2010 16:58:03 +0100, <noah_mendelsohn@us.ibm.com> wrote:
>> > That's the choice, I think.  I prefer #1.
>>
>> The reason I prefer #2 is that we have had reason to obsolete features
>> over time. Given that it makes sense that conformance also evolves over
>> time as we learn more about the medium.
>
> OK, maybe fair enough in certain edge cases.  I think there are at least
> two questions on the table:
>
> 1.  Is there an intention to have at least the vast majority of the older
> content work and be considered conforming?

I don't think that is necessarily the case. E.g. only the strictest  
HTML4/XHTML1 doctypes will give standards mode rendering of all the legacy  
doctypes. Almost standards mode and quirks mode are underspecified and  
undesirable features so pages should migrate away from them.


> 2.  If so, what is the appropriate editorial means to be used in the  
> media type registration and associated specifications to document such
> conformance rules.  I.e. should the media type registration continue to
> refer explicitly to the specifications for the older forms or not.
>
> If there is a conscious decision to obsolete particular features, then I
> think that can be handled in any case.  That is, the media type
> registration could do something like explicitly reference HTML 4, but
> indicate "however, features X, Y, Z have been made obsolete and are thus
> no longer conforming."

I'm not sure how that is different from HTML5 allowing the HTML4/XHTML1  
doctypes and obsoleting some of their features.


> I'm not offering an opinion as to whether the set of features to be
> obsoleted should in fact be non-empty; I am saying that I think we should
> take what I described as option #1 as the baseline, and if necessary,
> document specific deviations explicitly.  IMO, option #2 comes too close
> to:  HTML 5 is the new standard;  it's an exercise for the reader to
> figure out how much old stuff is still supported compatibly.  I think
> users want some more explicit guarantee that, unless warned to the
> contrary on clearly identified specifics, backwards compatibility is
> maintained.

The specifics are clearly identified though. E.g. the doctype needs to be  
changed, <font> needs to be replaced with CSS. Validators are pretty good  
at this stuff.


-- 
Anne van Kesteren
http://annevankesteren.nl/

Received on Wednesday, 3 February 2010 13:45:52 UTC