W3C home > Mailing lists > Public > public-evangelist@w3.org > May 2005

Re: Valid XML

From: Jim Ley <jim@jibbering.com>
Date: Mon, 23 May 2005 00:24:51 +0100
Message-ID: <030b01c55f25$744ab950$0200000a@Snufkin>
To: <public-evangelist@w3.org>

"Karl Dubost" <karl@w3.org> wrote:
>Jim, you are waking up in the morning, you are maybe taking a coffee.  Your 
>cup is falling over you and you have coffee all over you. Should  I kill 
>Jim Ley, Should I suppress the cup? Should I not try to drink  coffee in 
>the morning because I'm clumsy. :)

I'd go for killing me, it would be very cathartic for a great many people 
I'm sure, or drink orange juice from a kids juice cup, not only do you get 
all the same liquid as the coffee, but it's healthier and you're safe from 
being burnt by it.

>The work which has been done for QA Specification Guidelines [1] is  one 
>part of this  attempt to improve things.

The QA spec guidelines are excellent, obviously there are some parts I 
disagree with, however they do not help with broken legacy specifications, 
and continued evangelism for these broken specifications hurts the entire 
evangelism process.   Too often is the response from the evangelism 
community that invalid and non well-formed XML or XHTML is okay because it 
works in both browsers (or similar)   I know you're not going to say exactly 
that, you'll say that the documents do need to be well formed and valid, but 
I'm looking for arguments of why, as the argument of why it's okay to use 
XHTML 1.0 as text/html - "it works in both browsers" is exactly the same as 
the one used by people arguing for invalid content.

>So Jim and others, if your point is to say: W3C is not perfect and  there 
>are issues in specifications.

No, my point is that it's not perfect, and we have to recognise the 
imperfections, and not continue banging our heads against the broken bits 
pretending they're good.  XHTML 1.0 is broken when served as text/html, 
let's accept that, if there weren't sensible alternatives then it would be a 
good compromise, but there is a more than acceptable alternative, and 
focusing the evangelism arguments on that would be much more sensible.

>Le 05-05-20 à 14:16, Jim Ley a écrit :
>> Bjoern was just asking for similar help with XHTML guidelines. 
>> Arguments seem rather thin on the ground other than "well it works  in 
>> both browsers so all is okay" which is something I can't agree  with.

>:) It's where we have a strong disagreement.

So you're saying that "it works in both browsers" is a perfectly acceptable 

I disagree, I would say that "it follows where possible a standard", "it 
works in the tested user agents", "it doesn't rely in error recovery for the 
majority of its features", "it's as predictable as we can be in what future 
user agents will do with it" etc.  are acceptable arguments, just that it 
works in Internet Explorers and it's copycat admirers is not something I 
believe we should use as an argument, it doesn't move anything forward, 
invalid crazily nested crap works in those.

>>> Do you know a user agent which is unable to do c) ?
>> Yes, Internet Explorer 6.0, I gave the file (with corrections) in  the 
>> previous example.

>So you are saying that IE 6.0 doesn't display (render) XHTML 1.0  served as 
>text/html ?

I am indeed saying that, which is why it's not a good idea to evangelise to 
content authors the use of XHTML 1.0 - I would also say that it probably 
means that we should evangelise to users that they use another browser that 
can render more content types, but that's a seperate issue to this thread.


Received on Sunday, 22 May 2005 23:25:01 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:16:19 UTC