Graham Klyne (GK@ninebynine.org)
Sun, 08 Apr 2001 09:12:35 +0100
At 04:49 PM 3/26/01 -0500, Keith Moore wrote:
> > I think we are focused on "today" and offering hints about today's
> > semantics. Predictions about the future usage styles and their labels
> > are a bit beyond the scope of this effort.
>
>I agree that we shouldn't make such predictions. but this still begs the
>question of even how to represent today's usage styles without imposing
>present-day assumptions on the future.
To answer the last bit first, the context labels are extensible (through
registration), and some effort was made to draft the spec in such a way
that the label used was not of critical importance to message handling
semantics; i.e. if a new context label is not understood by an old
client, the message can still be delivered and processed, and similarly if
an old context label falls into disuse that will not be a cause for failure
or non-interoperability.
The goal here is for cooperating clients to convey information that allows
the user experience of message presentation to be improved, rather than to
enable all-new functions.
I would hope that the simple, minimalist approach to message-context
semantics will make its use amenable to evolution, following the styles of
user interaction supported by email protocols.
I think your observation makes one point clearer to me: the bar for
registering new message-context values should be set fairly low, to
facilitate evolution.
...
I think Jutta has it about right, in suggesting that the content of a
"short text" message is likely to of transient interest to the recipient.
#g
-------------- Graham Klyne GK@NineByNine.org
This archive was generated by hypermail 2.0b3 on Sun Apr 08 2001 - 12:26:49 IDT