- From: Geoffrey Sneddon <foolistbar@googlemail.com>
- Date: Fri, 23 Nov 2007 16:09:58 +0000
- To: Dean Edridge <akaroa74@woosh.co.nz>
- Cc: Maciej Stachowiak <mjs@apple.com>, Karl Dubost <karl@w3.org>, "public-html@w3.org Tracking WG" <public-html@w3.org>, Roger Johansson <roger@456bereastreet.com>
On 21 Nov 2007, at 11:31, Dean Edridge wrote: > > Maciej Stachowiak wrote: >> >> >> On Nov 20, 2007, at 10:21 PM, Karl Dubost wrote: >> >>> Some issues arose during the F2F and corridors discussions. >>> >>> * Shall the syntax style be stricter than the one recommended by >>> HTML 5 specification. >>> example: >>> <p class=intro>Readable Markup >>> <p class="intro">Readable Markup</p> >> >> It may be a good idea to recommend some style guidelines. However, >> I'm not sure this should necessarily be mixed with an intro for >> authors. For example, when it comes to quoting attribute values, >> HTML coding styles might recommend either always quoting or always >> omitting the quotes when possible; which is preferred is largely a >> matter of taste. >> >> I'd draw the comparison with if statements in C. Some insist on >> using braces even around single-line if conditions, like so: >> >> if (foo) { >> bar(); >> } else { >> baz(); >> } >> >> Others insist on omitting braces whenever possible: >> >> if (foo) >> bar(); >> else >> baz(); >> >> As far as I'm concerned, either coding style is valid, so long as a >> given code base is consistent. The same goes for quoting attribute >> values. It could be argued that skipping unnecessary quotes makes >> markup less noisy and therefore easier to read. >> >> In ay case, even if an introduction for authors recommends a >> particular coding style, it should draw the distinction between >> coding style conventions and actual document conformance >> requirements, just as Kernighan & Ritchie's famous book does the >> equivalent for C. >> >> Regards, >> Maciej >> >> >> > Why would it be a good idea to encourage people to omit quotes > around attribute values, when this method would only be suitable for > *one* of the *three* serialisations of (X)HTML5 ? > > application/xhtml+xml (XHTML) > application/xml (XHTML) > text/html (HTML) The serialisations sent over application/xhtml+xml and application/xml are wholly identical. There are only two serialisations: custom SGML- inspired text/html, and XML. > Surely it would be better to have just the one method that works > with all serialisations and get everyone to use that. > Otherwise we end up with web tools (WYSIWYG editors and CMSs etc) > that only generate markup suitable for text/html. But there are other differences between HTML and XHTML. xml:lang and lang only exist an XML and HTML documents respectively. |meta|@charset doesn't exist in XML documents (as it would mean redefining XML's character set detection, and would prohibit verbatim XML parsers from being used). |noscript| doesn't exist in XML documents (due to it needing to have an effect on the parser). To make conforming HTML documents a (tiny) subset of XML, a great deal of documents would become non-conforming HTML 5, requiring them all to be rewritten: If we make moving to HTML 5 too big of a change, will people bother? Lastly, there are some differences that make it impossible to make conforming HTML document a subset of XML: |blockquote| cannot be nested within a |p| for example. The only way to avoid this would be to require a third parsing mode incompatible with current HTML UAs in HTML5 UAs (which goes against the degrade gracefully principle). -- Geoffrey Sneddon <http://gsnedders.com/>
Received on Friday, 23 November 2007 16:10:20 UTC