Posted: May 18, 2010 10:32 pm
by Calilasseia
The reason why thread splitting was implemented over at RDF was because the database structure, and the construction of indexes on that structure, scaled upwards poorly once the database grew beyond a certain size. If LIFE has figured out what the underlying problem is, and his fix allows threads of arbitrary length without the search function taking enormous amounts of time, or bringing the server's processor to its knees, then the need for this procedure ceases to be. Now I happen not to be aware personally of any fixes that have been made facilitating this, but if our server admin is confident that this problem has been overcome, then I'm not going to argue. Not least because we have a server admin who, unlike a certain personage I and many others here could name, knows what he's doing. :)