When we reuse code on the web, we tend to do so by downloading a library and then including that in our own code. These libraries give us objects which we then use to get other things done. Most code reuse on the web is of libraries that are, not surprisingly, generic and multipurpose.
When we want to ship data across the internet, especially APIs from service results and so on, we tend not to ship any behaviours or code for parsing that data along with it.
What if we could ship objects on the web, extending them in a distributed way? Would that be useful or, because code is so fragile, would it be an unmaintained, indeed unmaintainable, nightmare?
What if there were a wiki that people could contribute to for scraping services and so on? One big problem with code on the web is that in order to be safe, we end up with browsers not permitting redirects to data:, URLs, and CORS denials breaking our frame tools. You can't really get much done in an environment like that, which is why people try to do those things in phablet apps instead. This is pretty bad for the open web.