Re: NU's polyglot possibilities (Was: The non-polyglot elephant in the room)

On Fri, Jan 25, 2013 at 5:50 PM, Alex Russell <slightlyoff@google.com> wrote:
> On Jan 25, 2013 7:36 PM, "David Sheets" <kosmo.zb@gmail.com> wrote:
>> Do you feel that the polyglot document does not help publishers
>> understand the (X)HTML syntax?
>
> This is the core question worth answering, and as far as I can tell from the
> responses here, I have no reason to think that polyglot markup aids
> understanding. It lacks the signage that clearly html or XML documents
> provide and without that clarity, its value seems washed out beyond
> recognition.

Polyglot documents are valid HTML documents with additional constraints.

How does this decrease clarity more than lenient (or conformant) HTML parsing?

>> I believe that the polyglot document serves precisely this purpose.
>>
>> Do you feel that the polyglot document hurts long-term viability of
>> the standards?
>
> I don't know, but it does impose constraints without demonstrated value.

It imposes constraints on document representations which are harmless.
Do you acknowledge *any* demonstrated value from polyglot? Or do you
feel that every argument I have made is somehow lacking? You have
offered no standard for "demonstrated value" despite my repeated
request.

What constraints does it impose on the standard? Why are those
constraints, if any (waiting...), undesirable?

>> I believe that the polyglot document decreases fragmentation and
>> guides spec authors to more sane rules.
>
> We disagree.

You believe it increases fragmentation? It is does not effect fragmentation?

You believe it does not encourage sane rules? It has no effect on sane rules?

What specific issues do spec authors have with the existence of polyglot?

>> Do you feel that the unelected, top-down structure of HTML
>> standardization should be given greater leeway to further fragment
>> implementations and introduce special cases? On what grounds?
>
> Yes. Because HTML is what gives the web value.

My understanding is that a loosely-coupled, evolvable system of
interoperable standards, hyperlinks, a uniform resource identifier
syntax, and extensible formats and protocols are what give the web
value. With only HTML markup, we'd have a pretty boring system. I
think these are some general principles of the technical architecture.
HTML uses these principles and is very important but is far from the
whole story. All deployed web standard software (servers, browsers and
XML processors alike) contribute value. Should we encourage
monoculture or allow diversity?

What specific, real costs does polyglot impose on HTML's development?

Do you have a replacement for XML you would like to offer? Do you
believe XML should be deprecated?

What harm will the web experience if polyglot becomes a REC?

Without an enunciated understanding of both costs and benefits of
polyglot, I fail to see how you or anyone else (myself included) can
render a rational positive or negative judgement on its suitability
for REC. Right now, I have several compatibility and simplicity items
in the "pro" column and 1 questionable syntactic item (HTML tags
should allow "/") in the "con".

Please, help me compile a comprehensive accounting of specific pros and cons.

Thanks,

David

Received on Saturday, 26 January 2013 04:08:56 UTC