Assess runtime performance for common names like Y. Wang about reciter

[jl987-Jie]

Receiving following error with Lucene:

Document contains at least one immense term in field="AFFILIATION_UNTOKENIZED" (whose UTF8 encoding is longer than the max length 32766)

when running ReCiter with yiwang as input.

Update:
Trimmed affiliation information to have maximum of 32766 characters long.

[jl987-Jie]

ReCiter starts slowing down after reaching into the first thousand articles. Could take a long time to cluster all those 70,000+ articles. (When ReCiter reaches the 3000th article, it's taking ~1 second to assign the article.)

[paulalbert1]

So, that means it would take ~18 hours to run for Yi Wang. That's too long. As they say, let's not let perfection be the enemy of the good. Don't forget about the other 98% of authors who have reasonably unique names!

Even if this figure could be improved by 100%, we would need some sort of fall back plan (search by full name). And if that means that for such people, we don't get every last publications, so be it. Can you predict how long it takes to compute publications for someone with 20 candidate articles vs. 100 vs. 1,000 vs. 10,000? I assume it is exponential?