ISFDB talk:Non-SF Cleanup

Looking at the current list of targets, I notice that the following keywords:

• Digimon
• El Autobus Magico
• Magic School Bus
• Pokemon
• Powerpuff
• Sabrina the Teenage Witch
• Shrek

may target borderline legitimate speculative fiction as well as. I wonder where we should draw the line? My thinking is that purely animal books for very young children (as opposed to books for older kids by, say, Brian_Jacques) are not really speculative fiction while Pokemon, Shrek and up are. Thoughts? Ahasuerus 08:13, 30 May 2006 (CDT)

I'm using the list to add heuristics to Dissembler. I'm assuming that Dissembler isn't going to be generating 100 percent of our new content - so if Dissembler is a bit overly aggressive on this topic it won't cause great harm (it only adds content - it doesn't remove any content), as someone will be free to add it manually. So the real potential harm comes from humans (or *extremely* intelligent AIs not under our direct control), who might pedantically follow the above list and delete relevant titles. Can we articulate when the above titles are okay? Or all they always okay? Should they be moved to their own cautionary section? Alvonruff 08:22, 30 May 2006 (CDT)

Well, the problem with data mining Amazon et al for children's series data is that it can result in a lot of false hits due to overlap with similarly flavored toys, videos, calendars, and other types of merchandise. It's useful to have these keywords listed here for human editors to review, but I am not sure that Dissembler's heuristics are advanced enough to take advantage of this list. Perhaps we could return to the issue once we have done more cleanup and have come up with more specific rules? Ahasuerus 09:52, 30 May 2006 (CDT)