Re: TAG Decision on Rescinding the request to the HTML WG to develop a polyglot guide

On Tue, Jan 22, 2013 at 9:51 AM, Daniel Glazman <daniel@glazman.org> wrote:
> (cc:ed to DOM 4 editors)
>
> Guys, I tried to make BlueGriffon output polyglot html5 for future
> 1.7 and I am hitting a serious issue: I am unable to do it for one
> single reason...

I think users of BlueGriffon would get more value if you focused on
the editing features and didn’t sink time into polyglot.

> Rendering engines are removing support for
> Document.xmlVersion, Document.xmlStandalone and Document.xmlEncoding
> that are deprecated in DOM 4. I have then *no way at all* of
> knowing if a document was parsed with an XML declaration or not
> from script, using the DOM.

Why do you need to know? BlueGriffon already can’t round-trip
whitespace between attributes. Why care about round tripping the XML
declaration? Both are serialization artifacts.

> The XML declaration is not reachable as a PI in the OM.

It would be a domain modeling error for it to be in the OM, since it’s
not part of the infoset. It’s just a serialization artifact.

> Since an
> editor like mine can edit all flavors of html, it still needs to
> output the xml declaration for xhtml

Not if you always output XML as UTF-8, which you should, since UTF-16
makes no sense for interchange and UTF-8 is the only other encoding
guaranteed by XML to be supported.

I think the best solution would be to ignore polyglot and not support
it, though. Then it’s OK to emit the XML declaration in XML.

> and for html5 if the polyglot
> option is not chosen.

You mean never for HTML?

> I guess this is an
> architectural issue for the TAG, right?

I sure hope not.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Tuesday, 22 January 2013 08:16:52 UTC