get queued messages as a ping?

[sondericker]

On page 49 of the standard it states that a Get STATUS_MESSAGES command of type STATUS_NONE shall be used to establish if a responder is alive on the link and is still talking. I have always used a Get QUEUED_MESSAGES command for this as it has the effect of also collecting things like a queued message generated when the DMX address has been changed at the responder. This saves me the extra step + traffic of having to ask for the queued messages separately.

Is it important that I follow the spec to the letter in this case where I really just want to be occasionally ping my responders? should I change my implementation?

[sblair]

Ideally you determine whether a device is still on the network through the normal use of QUEUED_MESSAGES so that you routinely pick up the Queued messages and any Status messages that are pending.

The reason the STATUS_MESSAGE with STATUS_NONE exists as there were some that had a need/desire to be able to routinely determine if a device is online but didn't necessarily have the ability or desire to be handling the Queued/Status messages coming back. So we defined a specific way for people that want to do that.

My personal position is that ANY message that generates a response, even a NACK, lets you know the device is online. Ideally you should be using the normal QUEUED_MESSAGES poll so that you are getting that info back and the STATUS_NONE mechanism is only for those rare exception cases.

[prwatE120]

John

your use of GET:QUEUED_MESSAGE must allow for the fact that I (as a responder) may not actually support this PID, as it is not one of the mandated PIDs.

As Scott has indicated, even the NACK you would get from me in this case indicates I am still in existence.

regards

Peter Willis