I'm wondering whether it might be simpler, at least in the near
term, to let consumers subscribe to selected sources rather than to
topical categories. That pushes the question of message
authoritativeness / jurisdiction /credibility out of the CAP
infrastructure and into the larger field of inter-agency and inter-
jurisidictional coordination, where it more properly belongs.
Taxonomies tend to be culturally loaded and can never be guaranteed
to be complete. Thus there's a real risk of "categorical
disconnects" leading to missed alerts either because of differing
interpretations of categories or of unforeseen events that don't fit
our preconceived categories. Maybe someday we'll have a reliable
taxonomy of the unexpected, but right now a degree of deliberate
imprecision seems to be the best we can do... and I sometimes wonder
whether even that is more helpful than it is risky.
- Art
On Jul 12, 2009, at 7/12/09 11:58 AM, Hannes Tschofenig wrote:
I should provide a bit more feedback about the background to my
question.
If you only set the value in the category field for the purpose of
human
consumption then there is not really an interoperability issue.
Now, with the work on http://tools.ietf.org/html/draft-rosen-sipping-cap-03
we wanted to define an event package for SIP that allows you to
"subscribe"
to certain type of events: you might indicate something like
location and
the type of events you are interested in.
Now, the semantic of the category field suddently matters. With the
individuals-to-citizen emergency services we tried to come up with a
description of the emergency services categories, see RFC 5031.
Ciao
Hannes
_______________________________________________
earlywarning mailing list
earlywarning at ietf.org
https://www.ietf.org/mailman/listinfo/earlywarning
