> Imagine all e-shops would implement the same public API Very idealistic, but if you have ever heard the acronym "EDI" then it will become quickly apparent ho

> Imagine all e-shops would implement the same public API Very idealistic, but if you have ever heard the acronym "EDI" then it will become quickly apparent how incredibly fraught it is to try and map ones "local" taxonomy onto a "global standard" taxonomy. Let's just take https://schema.org/brand https://schema.org/brand -- should this property be the legal brand (Sony Interactive Entertainment Inc.), the marketing brand (Sony), some URI that everyone agrees upon and can readily look up from their desktop? What is the verification process a fivrr contractor can use at the end of their php work for Middleville Cakes and Cookies to know that the result conforms to the global taxonomy and thus will be consumed correctly by the aggregators and price comparison engines mentioned? I do recognize this may have a tone of "pfft, what a terrible idea" but I promise I don't mean it in such a dismissive way. I was similarly thrilled back when I learned about the semantic web, too. But having been down in the muck for as long as I have, trying to implement loosey-goosey specifications, has made me realize it's not the idea that's flawed, it's the assumption that all actors have good intentions, perfect knowledge, and infinite time to conform to perfect standards. And don't even get me wound up about the semantic law markup / legislation as code ... > with no need for web scraping. That will only ever be true in some mythical future where the HTML contains only presentation structure and is populated by the aforementioned public API. There are countless examples of some parts of the data being statically embedded in the HTML that are not part of the API response