Going over the comments in SOLR-1316, I seemed to have lost the forrest for the trees. What is the benefit of using the spellcheck based suggester over something like the terms component to get suggestions as the user types?
Maybe it is faster because it builds the in-memory data structure on commit? Seems like the terms component is pretty fast too. I'd appreciate any additional insights about this. There are so many "solutions" to auto-suggest for Solr, it's hard to decide what approach to take. Cheers, Tim
