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

Re: Versioning re-visited (was : mixed signals on "Writing HTML documents", tutorial, etc.)

From: Raphael Champeimont (Almacha) <almacha@animala.tar.cc>
Date: Wed, 27 Jun 2007 12:41:53 +0200
Message-ID: <46823EF1.6040400@animala.tar.cc>
To: Henri Sivonen <hsivonen@iki.fi>
CC: "Philip Taylor (Webmaster)" <P.Taylor@Rhul.Ac.Uk>, Ian Hickson <ian@hixie.ch>, Philip & Le Khanh <Philip-and-LeKhanh@Royal-Tunbridge-Wells.Org>, HTML Working Group <public-html@w3.org>

It seems to me that many people see validation as some mark of quality 
about their writing. They want to make possible for readers to verify 
that the author "have taken the care to create an interoperable Web 
page" (to cite the W3C HTML Validator). When HTML6 is out, these authors 
still want to show that they created a conforming HTML5 document, even 
if the features they used have become deprecated in HTML6.

This way, validation is just some mark of "honour", which maybe 
encourages authors to write documents according to the specs...

For me this would be the only valid reason to use versionning, UAs don't 
use HTML versions for HTML parsing, a document's HTML version only 
matters if one wants to say "I wrote this document correctly, respecting 
the standard in effet when I wrote it" and this will be true forever.

At the same time, there is no real "cost" to use versionning, the real 
problem would be if UAs started parsing HTML differently depending on 
the version, but in that case the last HTML spec may require UAs to 
parse old HTML documents using the new rules.

Raphael Champeimont

Henri Sivonen wrote:
> 
> On Jun 21, 2007, at 18:57, Philip Taylor (Webmaster) wrote:
> 
>> I have been
>> creating web pages ever since HTML 2.0 (many on this list
>> will have been creating them for far longer), and every
>> page that I have validated in the past will validate today,
>> because the specification against which it is written is
>> enshrined in its DOCTYPE directive.
> ...
>> The goalposts are fixed, for perpetuity.  A valid document
>> will remain valid, an invalid document will remain invalid,
>> and a validator will never give different answers concerning
>> a particular instantiation of a document.
> 
> You make it seem like validity in and of itself was more important to 
> you than how the document fares in its practical environment today.
> 
> To me, it makes sense to check the document one is currently authoring 
> against the current conformance criteria. However, people advocating for 
> explicit version haven't explained why they want to take old documents 
> and see if they still validate against the then-current version of HTML. 
> Old documents are water long since under the bridge. And UAs need to 
> cope whether or not the old documents are conforming to any conformance 
> criteria.
> 
> If you are yourself the author and you have the intent of updating the 
> document, you are pretty much in the same situation as you'd be if you 
> were writing a new document from scratch: You should observe the current 
> conformance criteria. If you are the author but you have no intent of 
> updating the document anyway, why would you bother validating? If you 
> are not the author, why do you care whether someone else's document 
> conformed to then-current conformance criteria when it was written?
> 
> As for changes in the theoretical HTML6, there are three cases:
> 1) The future HTML6 WG goes nuts and makes an incompatible language.
> 2) HTML6 adds a feature using the parts of the syntax that were 
> non-conforming in HTML5.
> 3) HTML6 deprecates or obsoletes a feature.
> 
> Let's look at each one:
> 
> 1: Either the future HTML6 WG introduces a version discriminator or the 
> market ignores the output of the future HTML6 WG.
> 
> 2: A document authored according to the HTML5 rules stays conforming. No 
> harm done.
> 
> 3: If the future HTML6 WG deprecates or obsoletes something that we got 
> wrong, they'd better have a pretty good reason to. (Otherwise, they will 
> have gone nuts.) If something is so bad that it deserves to get 
> obsoleted or deprecated, shouldn't conformance checkers notify authors 
> instead of silently keeping the goal posts fixed for perpetuity?
> 
> --Henri Sivonen
> hsivonen@iki.fi
> http://hsivonen.iki.fi/
> 
> 
> 
Received on Wednesday, 27 June 2007 10:43:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:01 GMT