Background
- HTML and fixed styles
- CSS
- Dave Raggett's CLASS attribute
Web Culture
- Web culture: rough consensus and running code
- In other words, quick hacks
- Web browers became an undergraduate project.
XML and Engineering
- SGML (and later XML) used in engineering:
- E.g. nuclear submarine repair
- (and bird watching trip reports)
- Engineering for precision and reliability.
The Semantic Web
THE SEMANTIC WEB
The Semantic Web
REQUIRES...
The Semantic Web
UNAMBIGUOUS REPRESENTATION OF
The Semantic Web
MACHINE- DISCOVERABLE
The Semantic Web
SEMANTICS.
The Semantic Web
The semantic web requires
unambiguous
representation of
machine-discoverable semantics
Building The Semantic Web
So we need to add meaning to Web pages.
- When the old pond
- gets a new frog
- the hype begins
Microformats
A Microformat is a named markup idiom.
The most common is to use the HTML CLASS attribute
to add rough meaning.
Example: geo
<div class="geo">
<abbr class="latitude" title="37.408183">N 37° 24.491</abbr>
<abbr class="longitude" title="-122.13855">W 122° 08.313</abbr>
</div>
Ambiguity
THE DISCOVERY OF TIN IN BRITAIN
Ambiguity: examples of title
- hcard: for job title
- so does hresume
- xoxo (a microformat for outliners) for an item title
- workofart uses it for the title of a work of art.
- Others? In law, a title is a right to property
Reminder:
The semantic web requires
unambiguous representation of
machine-discoverable semantics
Machine-discoverable
Which microformats does this document use?
What inferences can we make from the markup idioms?
Let's add markup to the document head to point to definitions.
Unambiguous representation
How do we avoid talking about Mrs Aerial Photography?
Validation How?
Two approaches:
- add explicit namespaces to elements and class values
- pick out the idiomatic markup into separate document fragments
Automating Validation
Either approach (namespaces, fragments) needs automation.
Today, people use a technique called...
Is this approach good?
THIS IS
Why?
- Hard to write
- Hard to maintain
- Hard to extend
- The wrong people for deployment
What to do?
PONDER
MDL
The solution: a little language to define a microformat,
and from which XSLT could be generated automatically
MDL Purposes
- Add disambiguation in place
- Extract fragments for validaation
- Generate HTML from fragments
- Allow recognition of the idioms
- Support formal documentation
MDL Disadvantage
MDL has some disadvantages
- Complex when microformats overlap
- It doesn't exist.
Discussion