Re: Precision and error handling (was URL work in HTML 5)

"\"Martin J. Dürst\"" <duerst@it.aoyama.ac.jp>, 2012-10-02 16:37 +0900:

> Getting back to HTML5, is there anybody except maybe the spec editor(s)
> themselves who knows all the parsing rules by heart and therefore can
> productively make use of them?

You don't need to know the parsing rules to understand how to produce
conforming HTML documents. The spec provides document-conformance
requirements for authors/producers, separate from the parsing rules --
which are for UA/consumers, not for authors.

> What would you give as an advice to somebody who wants to "productively"
> edit HTML5 source?

Follow the document-conformance requirements provided in the spec, or in
some reference, how-to, or tutorial that follows the spec.

> a) Just write what you think might work, HTML will do what you mean (DWYM).

I don't think anybody has seriously suggested that's a good idea. It
certainly not what they HTML5 spec itself suggests.

> b) Follow a few very simple rules (mostly just XML-like syntax and
> structure), and easily stay on the safe side and don't go near the cliff,

The document-conformance rules provided in the spec are in fact pretty
simple, despite the fact that some people seem to have in interest in
suggesting they're not. And these rules existed long before XML did. For
example, you've never been able to do <script/> in an HTML document. The
fact that XML created self-closing tag syntax for XML didn't change HTML.
And the fact that you don't need to use end tags for the p element in HTML
has always been a fact. The arrival of XML did not change that.

People put all quite a lot of pages up on the Web before XML came along,
and people in those years didn't seems to have considered HTML authoring
rules to be very complicated.

> and frequently use a validator or similar checking tool.

Yeah

> Second, the question is how friendly the (currently main) HTML5 spec is to
> the users.

I think we could ask that question about lots of our specs. I happen to
think the HTML spec is a lot more friendly than most. Certainly, Hixie's
put a lot of effort into trying to put useful guidance for authors into it
wherever appropriate. At the same time, we're not here to make our specs
into tutorials. For better or worse, the main target of the HTML spec is UA
implementors, and it goes to great length to document the UA conformance
requirements as thoroughly and precisely and unambiguously as possible. Any
spec that does that is inherently going to be a pretty intimidating read to
a lot of readers.

> See Noah's point about making the author spec the main HTML5 spec.

I'm not sure what it would mean to make the author spec the "main" HTML5
spec.

  --Mike

-- 
Michael[tm] Smith http://people.w3.org/mike

Received on Tuesday, 2 October 2012 08:11:42 UTC