Re: The right usage of tags

On Tue, 28 Mar 2006, Karl Dubost wrote:

> IMHO, it's not really a question of tag names and/or attributes.
- -
> Applying semantics to a text means *** authoring *** this text.

Agreed.

> We can create all elements on earth with rich semantics, if we don't give a 
> way for people to do *** Rich authoring *** (aka ways of interoperable 
> templating plus UI langage for richer interface in browsers), we will not be 
> able to improve the right use of semantics on the Web.

I think the obvious implication is that markup should be simple enough to 
be understood and gladly supported by people who create authoring tools. 
This means that less is more.

On the other hand, many authoring tools _have_ structured tools like a 
menu for inserting elements, as a logical rather than physical menu.
Yet the font selection tools are there too, and so are the buttons for
bolding and italics, etc. This is the main difficulty, I'd say.

> An HTML 4.01 authoring tool (not defined) should not propose (except if you 
> want a source code view) something like.
> 	"Add a blockquote"

The example is not very good, since "blockquote" can be understood as a 
prose name, short for "block quotation". So the function is really 
logical.

> But could propose
> 	- "Add a citation" with an interactive form where you can enter the 
> text, the author, the source, the origin of the quote (URI, URN, etc.)
> 	- Cut and Paste from another application ("copy this as a quote")

Agreed, but it's not a citation but a quotation. Half of people seem to 
get this wrong. I did too; the Finnish word "sitaatti" (from Swedish
"citat"), the German word "Zitat", etc., mean quotation, not citation.

Incidentally, this means that an authoring tool should preferably be 
localized, so that the author can work in his native language, or his 
preferred language and see menus and explanations in an understandable (to 
him) way. More importantly, it needs to be _properly_ localized if it is 
localized at all, and this is harder than most people think.

The general idea is essential, though. Structured authoring should be made 
as easy as possible. Layout-oriented authoring has become very easy 
(actually, too easy in many ways), and something similar should be 
developed for structured approach.

Of course tools are distinct from specifications, but they should be 
developed at least partially in parallel. Besides, specifications should 
be realistic: is it imaginable that simple tools will be created for
using such-and-such markup and that people will use it?

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Tuesday, 28 March 2006 20:37:40 UTC