Kevin,
I am also interested in a deeper explanation of what you mean by, "The easy way to do that is just to add a trigram index and search for similar strings, and forget about full text search." Because I need to make a decision about whether to use full text search or use other pattern matching facilities such as LIKE and/or regular expressions. For me, the reason I don't just default to full text search is the documents are relative small (i.e, HTML <= 128K) and number fewer than 10,000 so I'm not sure if the effort expended to learn the ins/outs of full text search will be beneficial to my use case.
Regards,