Live blogging the first morning session . . .
This is a full day workshop about educating web professionals. People involved in the presentation include Chris Mills, Stephanie Troeth, Aarron Walter, and Bill Cullifer. Others involved include Dave Shea, Derek Featherstone, Jeffrey Brown, Glenda Sims and Leslie Jensen-Inman.
Chris Mills from Opera explained that we are all here because we care about training web professionals. He’s the standards evangelist at Opera and was instrumental in developing the Opera Curriculum.
The speakers introduced themselves and then there was time spent finding out about the people attending and why they are here.
Steph talked about the fact that although it’s easy to implement a web page, it’s difficult to do it well. Aarron mentioned that education needs help keeping up. Industry needs to reach out. Adequate and appropriate materials were also mentioned.
Available right now is the Opera Curriculum. Coming soon is WaSP Interact. This will be public March 16 and will be the WaSP Curriculum Framework. These two projects dovetail in that the courses are based on competencies, with related assignments and evaluation rubrics for each assignment.
They handed out a pamphlet full of example assignments from the WaSP Interact curriculum. There was a markup assignment, a presentation assignment, a scripting assignment and an accessibility assignment. Each assignment contained a description of the assignment requirements for the student’s use, and sample answers. There were other assignment ideas for related work, plus a list of books and articles to assign related to the assignment.
Mike Smith on Markup
Mike Smith talked about markup. Semantic markup is markup that encodes meaning into content. Semantic markup transforms a document into an information source. The information becomes usable in unanticipated ways when the structure is reusable. In addition, semantic markup is portable in the sense that it isn’t specific to a particular device but works across a range of devices and contexts. Accessible markup does not exclude users of any particular class. The same techniques that make markup semantic also make it accessible. He talked about using class names and id values well so that markup is maintainable.
There are two syntaxes for storing documents: text/html or XML syntax (XHTML). Don’t serve machine-generated XHTML with a text/html MIME type because it may not work in every situation. And, don’t use self-closing tags served as text/html. HTML doesn’t enforce much structure on documents. Therefore, authors need to build structure in consistent ways, e.g., make good use of heading elements and div elements. Use other structural elements such as microformats and RDFa. Do it earlier rather than later. Some examples of adding semantic meaning that might be useful include adding title attributes to <a> elements to help people decide whether or not to click on a link.
Smith talked about the controvery over the use of the alt attribute with the img element. He mentioned that an image can be used for different purposes, and the alt text should be appropriate for the purpose the image serves in a particular context. Good alt text writing is a major needed skill. He’s on the side of the argument that says alt text does not need to be a required part of the schema but that WCAG should list alt text as a requirement. Interesting way to approach the idea and one I need to ask him more about later.
Smith talked about definition lists. He said they are for associating ideas and don’t necessarily have to be lists, as such.
In HTML 5 <em> will be emphatic stress and <strong> will be strong importance, if the current draft becomes a reality. He also mentioned that in HTML 5 <i> and <b> may be given some sort of semantic meaning as opposed to the strictly presentational use they now have. This is not a done deal yet either. He said the HTML 4 spec that said to use <cite> for names of people is wrong and <cite> should not be used to markup names of people.
He said SVG is going to be massively important in the future.
I don’t know if you or Mike Smith got this wrong, but self-closing tags do work in HTML sent as text/HTML… HTML 5 explicitly allows it. It is just ignored.
Lars Gunthers last blog post.. Jag skall tala i Frankfurt
Lars, I heard Mike tell us two different times not to use self-closing tags, so I know I got the quote right. As for whether it will validate or be harmless in HTML5, I don’t know.
Lars, What Virginia wrote here does in fact accurately reflect what I said. I did say “Do not use self-closing tags.”
The context for that statement was slides 59-61 of my presentation at http://www.slideshare.net/sideshowbarker/michaeltm-smith-ed09-presentation
The statement was in the context of best practices, and not really around whether self-closing tags are allowed in HTML5 or work in text/html. But I cited a specific case where a self-closing tag is in fact not allowed either in HTML4 or HTML5, and will not work as expected in all browsers. The example was . And earlier in the presentation, I gave a similar example:
Both HTML4 and HTML5 require an end tag on all and elements, so both of the above examples are actually non-conformant. Unfortunately, XML parsers have no way to check for the end-tag-required constraint — so if you’re using an XML toolchain to author/generate/validate content, there’s no way to check for and catch an end-tag-required constraint violation.
Yeah, I know I could say, as a best-practice statement, “Don’t use self-closing tags for end-tag-required elements”. But I frankly expect that there are many authors who will either never properly learn (or never properly be taught) which elements are end-tag-required elements, or who will pretty quickly forget it after learning it.
So I think it’s a minimal best practice to teach students, “Don’t use self-closing tags in text/html. At all. Ever.” An additional best practice to teach beyond that is, “Do know which elements are end-tag-required elements.”
At other events when I’ve discussed the problem with expecting student to remember which elements are end-tag-required elements, it’s been suggested that the burden on instructors and students of teaching/learning what the end-tag-required elements are is no greater than the burden of teaching/learning what the void elements are (that is, what HTML4 refers to in certain cases as an “empty element” — link, meta, img, input, br, etc.).
My answer to that it’s far less dangerous for an author to make the mistake of putting an end tag on an void element — for example, — than it is for an author to omit an end tag from a end-tag-required element such as .
So while I would say that another best practice is to teach students, “Do know which elements are void elements.”, I would also say that you should not always expect author to remember that. I do recognize that teaching a “Don’t use self-closing tags in text/html. At all. Ever.” best practice to authors who aren’t going remember which elements are void runs the risk of them putting end tags on void elements as well as non-void one. But as I mentioned above, I think it’s less dangerous for them to make that mistake than it is for them to make the mistake of omitting an end tag from an end-tag-required element.
–Mike
Michael(tm) Smiths last blog post.. Carakan
It looks like the comment system doesn’t escape markup in my comments. My bad for not using entities when I wrote it. Anyway, the text should have read:
The example was <script src=”functions.js”/>. And earlier in the presentation, I gave a similar example: <a name=”abstract”/>
If your browser render all the text after the 3rd paragraph of the comment as a hyperlink to nowwhere, it’s because of the <a name=”abstract”/> instance.
The other parts of the comment where the markup got eaten was where I wrote:
Both HTML4 and HTML5 require an end tag on all <script> and <a> elements…
and:
My answer to that it’s far less dangerous for an author to make the mistake of putting an end tag on an void element — for example, <img src=”foo.jpg”></img> — than it is for an author to omit an end tag from a end-tag-required element such as <script>.
Michael(tm) Smiths last blog post.. Carakan
Thanks for noticing what was going on with your first comment and coming back to clarify.
The reason this came to my intention is that students were confused. An early draft of the HTML 5 spec disallowed the self-closing syntax on void elements. I saw some implementations of php’s nl2br using regexp, since nl2br outputs <br />
I.e. I came into this issue from the other end. I have found it very easy to teach students that there are basically some elements that you can use the self-closing syntax on: img, br, hr, input, link and meta. Rule: They are always void.
Some elements are sometimes void, and could be self-closed in true XHTML (script, a). To avoid mistakes I recommend them to never use the self closing syntax on these – ever…
IMHO opinion that is easier to teach.
What they find hard to grasp is when elements are implicitly closed. E.g. this is easy to get:
<td>foo<td>bar<td>
But when a paragraph seemingly contains a table? IE 6 and 7 gets that wrong (Acid2) to add confusion to the issue…
Lars Gunthers last blog post.. Jag skall tala i Frankfurt