W3C home > Mailing lists > Public > public-html@w3.org > November 2012

Re: Polyglot Markup Formal Objection Rationale

From: Smylers <Smylers@stripey.com>
Date: Mon, 5 Nov 2012 12:37:37 +0000
To: public-html@w3.org
Message-ID: <20121105123737.GE2391@stripey.com>
Jirka Kosek writes:

> On 4.11.2012 15:59, Lachlan Hunt wrote:
> > 1. Normative vs. Non-Normative
> > 
> > Specifications should be considered normative when they seek to
> > define implementation and authoring conformance criteria.  Documents
> > that merely seek to describe authoring practices or provide tailored
> > information to a particular audience about content which is
> > normatively defined elsewhere, should be non-normative.
> Polyglot is nothing more then just profile of HTML5 which defines
> subset of syntax. There is no reason why definition of such profile
> shouldn't be normative.

Yes, it makes sense that the term is normatively defined, and that the
Polyglot spec is the place that does that.

But surely the definition of polygot mark-up is simply a statement
saying something along the lines of[*1] a document is conforming
polyglot if it conforms to both the XML and text/html requirements of
HTML5 and has the same meaning in both serializations -- that is, it's a
definition of the principle, by reference.

All the details and implications of what that means are simply applying
the normative requirements of the HTML spec, so they aren't themselves
defining anything. As such I find it misleading and confusing for them
to claim to be normative; they are helpful guidance, but they don't
actually add anything in terms of requirements, since they don't add any
additional requirements on top of those required by reference as part of
the definition of polyglot mark-up.

> If you for some reason decide to conform to such profile, it's much
> better if such profile is normatively defined.

Yes. Jirka, would a normative definition like that satisfy you?

The definition of the term "polyglot markup" is in a section explicitly
marked as non-normative in the current draft spec, despite being linked
to from elsewhere in that document as a definition:
I think it's confusing that this definition _isn't_ normative, and I
don't understand what linking to a non-normative definition means, or
how there can be normative requirements for creating something which
doesn't itself have a normative definition.

Incidentally, the latest working draft, dated October 25th, has a red
box pointing at the editor's draft being more up to date:

But the editor's draft it links to is dated July 9th:

> > Such duplication of normative definitions has the potential for
> > introducing unintentional conflict between the two specifications.
> > By ensuring that the polyglot guidelines remain non-normative, then
> > it is clear that, even in the case of such a conflict, the HTML5
> > specification's normative requirements take precedence over the
> > guidelines' non-normative descriptions.
> Well this problem (if ever exists) can be easily solved by adding one
> sentence to Polyglot which will say that in the case of conflicts
> HTML5 has precedence.

Yes. In which case, the Polyglot spec by its own admission would no
longer be canonical for those definitions and requirements. So it would
be bizarre, and confusing, for it to be simultaneously claiming its
requirements are normative and that it is out-ranked by the HTML spec.
In effect, that would make its descriptions non-normative. So it would
be less confusing, and more accurate, not to claim to be normative for
those parts.

> But as Polyglot is subset of HTML5 this is quite unlikely,

I disagree. Given how many mistakes have been found in specs in general
(not saying anything about the Polyglot spec) it would seem unlikely for
_any_ spec to be published without any mistakes in it.

> and if there is such mismatch present I hope that corresponding bug is
> filled.

Of course. But checking such reports, fixing them, and publishing
updates takes time. Until that happens there would be two conflicting
specs, both claiming to be normative over something.

> And although I can be considered as very XML-biased I don't consider
> Polyglot as a preferred syntax for general case. I would recommend it
> only for very special scenarios. If you need to process HTML5 as XML
> it is easier not to put additional burden on content producers and use
> library like http://about.validator.nu/htmlparser/ to turn HTML into
> XML.

I think it would be useful if the Polyglot spec's introduction said
something along those lines.


New series of TV puzzle show 'Only Connect' (some questions by me)
Mondays at 20:30 on BBC4, or iPlayer: http://www.bbc.co.uk/onlyconnect
Received on Monday, 5 November 2012 12:38:02 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:45:58 UTC