Re: Bug 7034

Sam Ruby, Tue, 23 Mar 2010 22:34:46 -0400:
> On 03/23/2010 10:08 PM, Maciej Stachowiak wrote:

>> To give a specific example, I would like to consistently avoid
>> presentational markup on the webkit.org, but I do not want to add an
>> xmlns declaration to every page.
> 
> Two things:
> 
> [chair hat off]
> 
> (1) that goes into the territory that I previously referred to as 
> YAGNI, unless you also want to explicitly close all open elements and 
> consistently quote your attributes.  My intent was only intended to 
> propose a "no motor cycle helmets" version and a "with motorcycle 
> helmets" version.  The combinatorics get much more complicated 
> otherwise.
> 
> [chair hat on]
> 
> (2) I want to make it clear: you (personally) would be willing to add 
> some bytes to your pages, but you personally could not live with 
> those bytes starting out with the characters x m l n s, and you 
> personally find this, ... what?  Offensive?

May be Maciej's issue is that you suggest having a mode (the normal 
mode ?) which allows many things that Maciej would like to be warned 
about. While OTOH, if he adds the xmlns="http://www.w3.org/1999/xhtml", 
then the requirement becomes to use XHTML syntax, which would be 
stricter than Validator.nu's current modes and perhaps "too much". ;-)

I personally do not find that I need to be warned about my use of 
<strike>, and would instead be very happy that I could validly use it 
without at the same time having to use a doctype that doesn't trigger 
standards mode. I would also like to be able to use NCRs without 
semicolons. ;-)

I personally disagree with the definition of what is semantic or not in 
XHTML 1.0 strict. But not enough to not be able to accept that 
xmlns="http://www.w3.org/1999/xhtml" signifies a best practise that is 
somewhat equivalent to a XHTML strict doctype (I suppose that is what 
Same has in mind). I also enjoy not closing <p> elements ... 

I have no clue what I would be using most - documents with 
xmlns="http://www.w3.org/1999/xhtml" or without 
xmlns="http://www.w3.org/1999/xhtml". May be I would author some docs 
with the xmlns present, and just remove it when the doc was ready - so 
as to not trigger validator errors when online.

But I feel that Sam's proposal fixes the most pressing problem in the 
current state of affairs. Namely the illogical choices we have to make: 
If we pick one doctype, then we can't use that element, or that 
attribute  - at least not in that context .... OTOH, if you pick 
another doctype where you can, then you get quirks mode ... Etc. The 
xmlns string seems to trigger a much more logical choice. And, not the 
least: it is a choice! And it is unrelated to quirks mode.

When it comes to HTML5, as it stands, then I find that it doesn't fix 
the problem of illogical choices: It creates warnings about @summary. 
And using the axis attribute is an error ... And so on. Which has made 
myself (and seemingly some others) look into how I can solve my 
problems via DOCTYPE editing instead [1]. 

[1] http://lists.w3.org/Archives/Public/www-validator/2010Mar/0027
-- 
leif halvard silli

Received on Wednesday, 24 March 2010 06:59:46 UTC