I explored the power of hyperlinks beyond traditional web pages, emphasizing their potential to connect apps and websites more seamlessly. Current linking methods, while fundamental to the web's interconnected nature, fall short of the ideal. I advocate for expanding the capabilities of registerProtocolHandler to access native schemes system-wide, enabling websites to handle various content types and register as system file handlers. Furthermore, I propose a richer vocabulary of link actions beyond 'VIEW', such as 'PICK', 'SAVE', and 'EDIT', to better define website and app capabilities, similar to Android Intents and Siri's functionality. Tools like Comlink inspire this vision by simplifying cross-app communication, paving the way for link-driven function discovery and a more integrated online experience.
This post revisits the UX issues that contributed to the demise of Web Intents. We never adequately addressed the wide array of potential actions and data types, leading to a generic and confusing user experience. Handling data return from a long-running operation in another app proved problematic, especially if the initiating app closed. The lack of an explicit API for expecting returned data, similar to Android's startActivityForResult, further hindered the development of intuitive UI affordances. The open nature of Web Intents resulted in a proliferation of schemas and protocols, making standardization difficult. Users were forced to choose an app for every action, and developers lacked control over how apps launched, leading to inconsistent experiences. Finally, the absence of reliable fallbacks for unsupported intents eroded developer confidence and potentially left users stranded. In short, a combination of UX shortcomings ultimately led to Web Intents' downfall.
