Comments on DI Select

I wanted to add my own personal comment to the DI Select review.

Recent trends in content adaption have been based on building self 
describing content, where the user agent can understand what the role 
and characteristics of section of content, and hence adapt it to a 
specific user needs. This trend allows the user agent adapting and 
rendering versions of the content  for user needs and scenarios that the 
author had never even considered.

DI Select does not require describing the content in order to adapt it 
to author specified scenarios.

With DI Select the author encapsulates each alternative based on boolean 
trigger that describes a user scenario and device that Author considers 
as most important. Other  profiles, such as are not in the author's 
"target audience"  will be not be included for special adaptations.

My fear is that it will discouraged authors from describing content 
attributes (though use of roles or other descriptors) as a method to 
promote content adaption. The author may not see it as necessary, nor is 
it promoted in the specification (even if it is possible).

These expressions will be hard for a user agent to reverse engineer, and 
understand what properties of the content made an alternative suitable 
for a user device or scenario. Therefore, information about the content 
required for adaption will be lost, making adaptive interfaces for 
marginalized groups the poorer.

I would like to see a two stage methodology, where the content is not 
encapsulated, or directly linked to the controlling expression, rather,  
descriptive properties of the content bind a section of content to a 
controlling expression. 

All the best

Lisa Seeman

Received on Wednesday, 8 June 2005 18:00:50 UTC