Re: Senarios for feature matching

[Graham Klyne <GK@xxxxxxx>]

Al,

I broadly agree.

The point of my posting was to raise the issue:  should we be considering
such levels of flexibility (hence potential complexity)?  And to try and
identify the trade-offs involved.

#g



At 10:14 08/04/98 -0400, Al Gilman wrote:
>
>One scenario for feature value comparison (feature expression
>reduction) is that the registration contains:
>
>* A grammar for feature expressions.
>
>* [An alternate grammar for the representation of said feature
>expressions in URLs]
>
>* The base URL for a public expression evaluator for expressions
>dealing in this characteristic.
>
>This is the simplest way to support an operationally-defined
>feature class across the whole Internet.
>
>This give the maximum performance on expressivity and reach
>scales.  CPU and wall-clock time to complete negotiation will
>be better for other techniques.
>
>Some effort will be required to police the consistency of
>expression evaluations and continuity of service of the
>expression evaluation resource.
>
>Al Gilman
>
------------
Graham Klyne