[whatwg] How to make HTML5 easier to teach (Was: several messages about HTML5)

[Keryx Web]

Ian Hickson wrote:
> On Sat, 24 Feb 2007, Keryx Web wrote:
>> If I as a teacher would want anything from a spec, it would be that it said
>> "this is wrong" and "this is right" in such a way that it becomes an
>> instrument in grading. ---  The
>> technology must enforce it upon them.
> 
> Could you elaborate on this, maybe showing some concrete examples of what 
> you mean? I'd love to make the spec easier to use for education.
> 

Speaking from __my__ experience, and the experience of those (too few) 
colleagues that I've met in Sweden who teach standards based web 
development, it is hard too make the student understand that something 
is wrong if he/she "get's away with it".

In my case that has led me to demand that all student's who want more 
than the lowest grade must use a strict doctype. The only problem with 
those doctypes, that are not intuitive, is that inline-level content is 
not allowed as children of the form-element. As soon as I explain "well, 
that's the rule..." most students do it right.

Since there is a ton of examples claiming to teach best practice that 
uses the transitional doctype, I must carefully enforce the 
__philosophy__: (S)eparation of content, design and behavior, 
(S)tandards compliant code, (S)emantic code. In this process terminology 
does matter. I always correct a student that says (designing with divs". 
("You are not designing with divs, you are designing with a CSS grid!")

The current tripartite doctypes are a tool. One can explain how things 
are "deprecated", that is allowed to ease the transition of existing 
material, but really not good to use, and that with the preferred strict 
doctype layout without CSS is practically impossible.

I would like the spec to clearly state what is allowed for backwards 
compatibility only and what is the preferred way of marking up content. 
I would like a spec that clearly says that some ways of marking up 
content is detrimental to accessibility and perhaps also usability. E.g. 
frames, including the iframe, or tables used for layout. You would not 
believe how many colleagues of mine who actually teach that frames are a 
good thing. My nephew, who studies i a nearby city, even had frames as a 
required feature of his work!

I would also like to see that the conformance checking software would 
give me a graded output:

1. Warning - technically allowed but bad practice
2. Error - not according to the spec
3. Fatal error - most browsers will not be able to parse at all

Putting a students work through a validator (be it HTML, CSS or WAI) and 
showing that it is not only my personal view, but hard facts, that 
influence my grading, makes the students feel more comfortable in what 
they are expected to learn and that they receive fair grades.

Of course some things, such as interaction design and aesthetics, are 
impossible to grade with a machine. And yes, I know that it is possible 
to write valid code that is unsemantic, so one must look at the code 
manually as well.


Lars Gunther