Re: Role Taxonomy - some comments

At 4:33 PM +1100 12/6/05, Charles McCathieNevile wrote:
>Some random comments on the spec version 


>2. Rather than assigning a numeric importance to each heading level, 
>why not just define them in terms of being more or less important 
>than another, with a transitive property? This is likely to be more 
>accurate, and allows for extensibility since it is relatively easy 
>to insert something in a list.


We should offer help with pruning/summarization. But a linear scale
of importance has fidelity issues.

There is also relative importance in terms of dependencies: "To
understand A, it is important that you understand B"


Received on Tuesday, 6 December 2005 15:23:44 UTC