Interesting. I see in that discussion that E.O. already piped in about the issue with regards to bots (ie, DeepBot). While the reasoning for the limitation seems sound, the alternative is much worse. They seriously should reconsider ways of imposing a limitation, such as requiring an oauth token for anything beyond the first 100 names retrieved and when the oauth token is used to retrieve a list from a streamer they don't have broadcasting rights with, set throttling limits (one request of up to 1,000 names per 30 seconds) or from a stream they do have broadcasting right for, something like 2,000 names per 15 seconds. That would lighten the use of resources while maintaining functionality.
As I said before though, would be better to only make a call per viewer when it's been locally detected to be over a month since the last check, rather than pulling the information and then checking, as many calls would end up being unnecessary.
