At 02:28 PM 5/15/00 -0400, Doug Chamberlin wrote:

>At 5/15/00 02:20 PM (Monday), Bill Karwin wrote:
>>I think this would be far more useful if we extend the EVENT mechanism to be
>>able to return data instead of just the event name. But we all wanted that
>>capability anyway.
>
>That's exactly where my thinking was headed.
>

Perhaps you could propose an alternative scheme that we could
examine.

The primary problem with passing data is timeliness. In almost
all cases, the data associated with an event has to be assumed
stale, which requires starting a transaction to get current value
anyway. There are secondary problems of queing and server side
resource consumption by laggardly clients, collapsing of events,
out of order delivery, etc.

But, please, go ahead and sketch an alternative scheme.

Jim Starkey