llMessageObject(key id, string message)

I, for one, have been waiting on promises of improved object to object communications for over TWO years. If you have too, please VOTE and COMMENT on the Jira issue below. Let's see if we can't get this reviewed and something finally done about it!

https://jira.secondlife.com/browse/SVC-310

One question. Once you create such object communication system and take it into your inventory. Then rezzing it again, you should find the object key change and be unavailable. What do you do?

Re-register the object. It's no different a problem than that faced by existing object to object email systems that may use http/xml-rpc or just notecards.

How would you re-register it?

And about the proposal:
Isn't the latest llRegionSay() enough for that purpose? With some tinkering, it can be used for that purpose too...
Or are you aiming for a private, one-way communication command between two objects, and two objects only? Sure it would be nice and useful, but there really is the need for it...?

I've been whining about wanting this feature too, but at least to me, it's much more about hoped-for performance advantages than actual functionality, which is equivalent to llEmail() for every purpose I've considered. The concern about llEmail() is that the implementation actually seems to be something like sendmail/smtp based on all the header info that the recipient gets, and that must be a heavy load on LL's servers and networks compared to any sensible implementation of the existing object-to-agent llInstantMessage(), which the proposal contemplates extending to object-to-object.

For in-sim comms, llRegionSay()'s implementation should be pretty lightweight--even less demanding than the llShout/-Say/-Whisper, since it never has to cross sim boundaries--but it will wake up every script in the region listening on the channel, at least long enough to filter out the message. Usually that's no big deal, but it's at least inelegant and potentially piggish when individual object-to-object comm is all that's intended.

So to my understanding, this really should be more a priority of LL than of scripters, but somehow we need to communicate the degree to which we'd use it, for LL to judge the potential benefit.