This talk gives an introduction to an often used but little thought about feature of fulltext search systems: did-you-mean (aka spellcheck). Whenever you misspell a query this useful feature kicks in and and will show you the query you really meant - or won't it?

We will highlight potential requirements to a did-you-mean feature and identify their difficulties. We will discuss the implementations available in Solr and Elasticsearch. But there are also alternative algorithms and data structures used in real world search projects which we will show. What is the difference between all these approaches and which one should you use when?