Best way to get/discover UUIDs for system comms???

Hi all,

I have a system under developer with multiple communicating objects. I want the communications calls to be able to cater for varying distances between objects (i.e. abstract this from working code), which I've been able to do with assistance (tks to Christopher Omega). However writing generic code which can handle LINK/CHAT/EMAIL (depending on the distance between objects & whether they are linked) requires the object UUID as lowest common denominator (e.g. EMAIL required UUID).

So what's the best approach for a system to obtain UUIDs when required. I had been going down the use of sensing when required "sensor", however this has negatives of (a) need to move objects close <100m for this step first at initialisation, (b) requires manual intervention like this again if there is an issue and one of the objects is reset and then has to re-discover the other object's UUID, and (c) is problematic if you want to rez 2 instances of a system within 100m of each other (without doing a special build per instance with tweaked names).

I'm now thinking that hardcoding of UUIDs into a notecard which every object gets may be best. An auto-updater script/approach could then help automate the distribution of updated notecards (e.g. when one object in a system's UUID has to change).

Comments?

PS Will attach my table below

Ummm - can't work out how to attach the image of my spreadsheet - I thought I'd done this previously - have attachments been not allowed now?

I think http://w-hat.com/objdns might be exactly what you're looking for.

Wow - never knew this existed Ben (2 great finds within 10 minutes - the first was "lslint" )

Is this service (i.e. objdns) widely used by scripters? I guess I'm interested in how available/robust it is and some impression of whether it'll be around for months/years to come? I'd be quite open to supporting a defacto standard approach here however, don't get me wrong.

I'd still be very interested in hearing what the most common native-LSL approach for obtaining/discovering UUIDs that core LSL developers use? (see my original post - e.g. manuall update into a notecard, versus automated sensing based on object name etc)

Further thoughts after some received advice (use of DNS server type object) - any comments welcome

Recap - I'm trying to satisfy:
(a) multiple instances of a system/game can exist and
(b) that the same or different owner/hosters can have multiple instances
(c) ability for the owner or the system to manage rerezzing of the system/game (i.e. no involvement from me necessary)

I think this implies I'll need a service that supports a "getUuid(objectName, domainName)" where the domainName is like a name specific to that instance of the system/game.

Options/decisions would be:

[1] How objects are registered:
====================
a) Self registration to DNS server (e.g. register(object, domainName, UUID))
- ADV: Automated registration, no uuid notecard distribution required
- DIS: If DNS server UUID changes need to update each object's "DNS Server" entry (i.e. manual effort for the system owner)
- DIS: If DNS server resets need to manually trigger objects to reregister (i.e. manual effort for the system owner)

b) A "DNS/updater" service sends out UUID notecards to objects in the system/game
DIS: Owner has to manually update the UUID notecard and trigger the "distribute" mechanism to send it out to system objects


[2] Who owns the DNS server / how many are there:
===================================
a) One central DNS object owned by me
- DIS: Need to make sure I keep the same land (which I'm not sure of)
- DIS: If UUID does change need to somehow get each object in each owner's system updated manually with the new DNS UUID address

b) Each game/system has their own DNS object within their domainName
- ADV: Up to the owner to ensure contistency of the availability of the DNS server
- ADV: Could then not need the DOMAIN parameter when registering (see above)
- ADV: This DNS object could be the same object used for a script updater function (i.e. the script updater would need to know object UUIDs, which the DNS object would have)

Ummm, based on this I think I'm tending towards choosing 1B & 2B as the options. (?)