What 127.5 million forms can tell you about the state of front-end input validation on the Web
Summary
A research-focused analysis of front-end input validation patterns collected from 127.5 million forms. The author uses a web crawler dataset and an SMT-based constraint solver to study the prevalence and quality of HTML5 pattern attributes, revealing that patterns are common but often redundant or buggy, and that semantic misunderstandings lead to weak security guarantees.