emix — archive
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
— [Date Index]
| [Thread Index]
| [Month Index]
| [List Home]
Definition of a "Price Interval"
Rish, Phil,
Based on your comments, here are my thoughts.
·
Some DR or price programs have "scheduled
events" versus "unscheduled events". By that I mean, I can know
that every day at 4pm I will get the next 48 hours of estimated hourly prices,
and that every hour I will get an update on the next 12 hours with the next two
hours fixed. So, if those are events, then they are scheduled events like
getting the mail delivery. Even DR programs may have fixed time intervals (or
patterns) when called. And DR programs are scheduled (called) maybe day ahead.
So, we have the case of everything fixed and known ahead, as for RTP, and
alternatively, unknown time intervals (varying lengths) scheduled ahead of time
by some notice (could be very short notice). Does that leave out any DR program
or RTP program?
·
Seems for DR events, we need a flexible time
interval with contained event parameters (per Rish's original email). But
do we need this structure if we know the time intervals ahead of time? Do we
always need end time? We should have a format that allows reduction to only
giving start time, no end, and only giving the start time of the first
interval, and not of subsequent (maybe giving the constant time interval
length). But maybe the only advantage is conserving bits (and this is not worth
much)?
·
There should be a program identifier at the
start of a message so I can choose to pay attention or not. Or I guess you get
the message (pull/push) only if in that program.
·
Then a message type, since a program might have
different basic messages, i.e., the notice of event vs. the actual data, or the
48 hour price data vs. the 12 hour data.
Sorry if these are things discussed before...
David
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
— [Date Index]
| [Thread Index]
| [Month Index]
| [List Home]