W3C home > Mailing lists > Public > www-html@w3.org > February 2004

Re: Equivalency of caption, h, and title.

From: Orion Adrian <oadrian@hotmail.com>
Date: Wed, 11 Feb 2004 16:54:51 -0500
To: lachlan.hunt@iinet.net.au
Cc: www-html@w3.org
Message-ID: <BAY1-F22TI5zzubChbt0000d2ce@hotmail.com>

And all three of those examples are the names or identifiers (normal English 
use) for those objects.  The name of the section happens to be presented as 
a heading, the name of the table or image happens to be displayed by the 
object.  And the name of the document happens to be listed in the title bar. 
  But semantically, they are all the same thing.  To me that is the 
important part.  That I be able to search the semantic meanings of things, 
not just their implicit presentational information.

The defintion of caption (http://dictionary.reference.com/search?q=caption):
1. A title, short explanation, or description accompanying an illustration 
or a photograph.
3. A title or heading, as of a document or article.

The definition of heading 
1. The title, subtitle, or topic that stands at the top or beginning, as of 
a paragraph, letter, or chapter.

I didn't even bother to look these up before I made the suggestion, but it's 
nice to see that they at least mention the word title in the primary 
definition of both caption and heading.

Moving on; what XHTML really needs is to incorperate a metadata system.  But 
something slightly more robust than the existing system.

Looking onwards with metadata file systems there is a lot of information 
that is going to be there to be accessed and queried on.  So let's 
incorperate an XML based Meta language (this could be RDF, but from what 
I've seen it doesn't really work).  Let's use this metadata language to 
markup objects.  This would give us a universal system - a modular system.  
Then we could use this markup language to markup all different types of 
objects - forms, tables, sections, images, audio, movies, 3rd party objects. 
  Now that each object type is wrapped in this metadata, I as the content 
author don't have to write it.  Also it introduces consistency in the 
object's metadata is attached to the object and not the thing that refers to 
the object.

Once the metadata system is in place, now the following elements and 
attributes disappear.  title, caption, h, h*, and @title.  There are 
probably others, but I'm going to leave that for somebody else for now.

Now I can search within documents based on metadata.  I can extract a 
specific object based on it's metadata from a search engine without having 
to worry about the rest of the document.

Now the reason that RDF doesn't work for me is that it's a document that 
talks about resources.  The data isn't embedded with the object itself - at 
least initially.  I'd like to see metadata embedded similarly to the way 
that the head element works now.

<html xmlns="..." xmlns:m="metadata_URI" xmlns:t="table_URI" ...>
    <m:metadata profile="document_URI">
        <m:property name="3rd party prop">value</m:property>
            <m:metadata profile="text_URI">
                <m:title>Section Title (aka heading)</m:title>
                <m:summary>Summary of this section</m:title>
            <m:metadata profile="table_URI">
                <m:title>Table Title</m:title>
                <m:summary>Table Summary</m:summary>
                <m:property name="3rd party prop">value</m:property>

As you can see above the metadata system removes all of the elements of HTML 
that are simply metadata for actual content.

The above not the only way to do it.   Another way to do it would be to have 
a content element in the metadata element that would specify either the 
content or a link to the content.  Either one is acceptable.

It is my opinion that metadata systems will become such a force in computing 
in the few years to come that HTML must update the standard in expectation 
of those new systems.

As a side note, I would like to see more of the elements of HTML brought 
into their own namespaces like XForms has.  XTables would be nice.  Tables 
have a lot more potential than HTML allows them to have and an independant 
spec would allow them to expand.

Orion Adrian

Keep up with high-tech trends here at "Hook'd on Technology." 
Received on Wednesday, 11 February 2004 16:55:08 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:07 UTC