|
Everything quantum-x said is correct... and they use strack to convert linking codes for site A to linking codes for site B.
And, a different code for different sites makes sense, it is not "stupid" or "pointless". The reason they exist is that NATS has to handle the join forms and tracking for all sites, not just for one. So, it needs to know which site it is supposed to handle right now. Of course now everyone is going to say "but you know that domain"... and yes, you are correct, we know the domain. But using the domain to lookup which site it is (especially because it can have subdomains and such) is considerably slower than using a numeric id, which we deceided to do when we started NATS.
The unencoded linkcodes still use small strings to identify them which does not have a big overhead on top of the normal numeric id.
Hope this explains it.
__________________
"Think about it a little more and you'll agree with me, because you're smart and I'm right."
- Charlie Munger
|