Re: Statement why the Polyglot doc should be informative

Michael[tm] Smith, Mon, 5 Nov 2012 13:03:54 +0900:
> Leif Halvard Silli, 2012-11-04 19:47 +0100:
> 
>> 1. Polyglot Markup describes a flavour of HTML5 that plays nice
>>    with XML  pipelines, something you have asked for; [1] 
>> [1] http://twitter.com/hsivonen/status/263696331141431296

> 
> That tweet is not at all Henri saying he asked for "a flavor of HTML5 that
> plays nice with XML pipelines".
> 
> What it says exactly is "Again heard the claim that HTML5 doesn't play nice
> with XML pipelines. Clearly, I need to advertise
> http://about.validator.nu/htmlparser/  more."

OK. So let me guess. He reacted to a complaint that said that it was 
bad that HTML5 "promotes" not XML well formed HTML. But he, still, this 
way, recognized the need for playing nice with XML tools. Polyglot 
Markup would be one method for "playing nice".

> That seems instead to be about the ability to put an HTML parser into an
> XML pipeline as a drop-in replacement for an XML parser -- as validator.nu
> does -- so that you can feed it non-well-formed (non-polyglot) HTML
> documents and still have everything work as expected.
> 
> So it has nothing at all to do with "a flavor of HTML5 that plays nice with
> XML pipelines".

Yeah, Henri effectively said "you don't need specific format, because I 
instead have a specific parser". So I am sorry that I did not instead 
say that he recognized the need for playing nice with XML tools. 
However, no one can deny that polyglot markup would play nice with such 
tools.

Btw, such a tool that deployed a parser like Henri’s, could very well 
be a WYSIWYG editor. And, to a degree, some WYSIWYG editors already 
work that way, as they often allow you to switch format/doctype by the 
touch of a button

The problem, though, is that I have to date not found a single WYSIWYG 
tool that implements Polyglot Markup a 100%. For example, Blue Griffon 
produces pretty polyglot markup when it produces XHTML5. With one 
exception: It *does* insert that XML encoding declaration. And also, it 
*does* ask me to pick encoding ...  and a format. (Note that the 
particular problem of the XML declaration could be said to a very minor 
issue - as it is [in principle, but not quite in practice] only a 
syntactic cruft issue ,thus it be solved by allowing it in HTML ...)

For "end authors", that use various editors depending what is best for 
the task, it can be quite frustrating and irritating that editing a 
file in one editor can cause the file to be un-openmable in another 
editor. We are only interested in the semantics and care little about 
the format. Thus, we are not very prepared to go hunting for Henri’s 
parser. We want something that works everywhere. And if if WYSIWYG 
tools standardized on polyglot markup, we would get exactly that.
-- 
leif halvard silli

Received on Monday, 5 November 2012 12:01:54 UTC