W3C home > Mailing lists > Public > public-html@w3.org > September 2007

Re: HDP: Revised "Support Existing Content" Principle

From: Marghanita da Cruz <marghanita@ramin.com.au>
Date: Tue, 18 Sep 2007 07:03:16 +0000
Message-ID: <46EF38C5.1070801@ramin.com.au>
To: Maciej Stachowiak <mjs@apple.com>
Cc: "public-html@w3.org WG" <public-html@w3.org>

Maciej Stachowiak wrote:
> On Sep 17, 2007, at 6:21 PM, Marghanita da Cruz wrote:
>> To me, handling of poor markup by browsers is an example of
>> the "Degrade Gracefully" Principle rather than the
>> support existing content principle.
> That's not the intent of the principle. "Degrade Gracefully" says 
> nothing about the interaction of existing markup with either new or old 
> browsers. It's all about new documents authored to conform to HTML5 
> being able to work reasonably in pre-HTML5 browsers.

My understanding of  "Degrade Gracefully" is a concept that has been around as 
long as the
Web. The concept is that instead of providing a syntax error the browser
attempts to interpret and display the content.

So, old browsers will most likely degrade gracefully rather than report a syntax
error, when they start seeing HTML5 code they do not recognise.

There isn't much point in specifying what you want an old brower to do with 
HTML5. It will be upto authors to decide when to implement a feature.

>> After sending my email, I felt I should have added the broken html 
>> seems a good example of the degrade gracefully principle.
>> As I see it, Support existing content refers to features that may have
>> been in HTML3.2 or 4.01 or XHTML etc. specifications but
>> have been dropped in later ones.
> The intent of "Support Existing Content" is that it's about actual 
> existing *content*, not existing specifications. So it would not cover 
> spec features that have never been used or implemented, but it would 
> cover features that have never been in any spec and won't be considered 
> conforming.
> The "Support Existing Content" and "Degrade Gracefully" principles are 
> mirror images of each other. The first is about old content in new 
> browsers, and the second is about new content in old browsers.
>> My tests with the w3c compliance indicates that the quotes are 
>> required for a
>> tick/greenlight on conformance with 4.01 however, browsers degrade 
>> gracefully if
>> they are not present.
> I just tried with the W3C Validator and the following validates (note 
> the unquoted attribute values on img):
> "http://www.w3.org/TR/html4/strict.dtd">
> <html>
> <head>
> <title>foo</title>
> </head>
> <body>
> <p><img alt=foo src=foo.jpg>
> </body>
> </html>
>> I see the principles as relevant to the drafting of HTML5. So, 
>> providing an
>> example of the implications of dropping a feature could be useful.
>> eg. if vspace or alt was dropped browsers would still need to support 
>> this feature.
> Well, we'd have to define what "dropped" means. Dropping them from the 
> conforming language would be irrelevant to browser support. Dropping 
> them from the supported language would mean, in these two particular 
> cases, that browsers would still have to implement it but without a spec 
> to go on.
> Regards,
> Maciej

Marghanita da Cruz
Phone: (+61)0414 869202
Received on Tuesday, 18 September 2007 07:27:15 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:44:21 UTC