Microformats – building blocks for data-rich web pages
11 points by rau
11 points by rau
This is a blog front page, the latest entry was in 2022.
I think we can all agree that the Semantic Web dream is dead now. Not only is the core idea (labelling your content so that it's easier for machines to classify) moot with the advent of LLMs, the very concept feels less like a dream and more like a nightmare for many people.
I don't think it's moot? Just because an LLM can guess we often doesn't mean knowing for sure is useless. This is why APIs etc still exist. Microformats provide a standard set of vocabularies for building APIs into the webpages you are making anyway
I got the impression that microformats were mostly irrelevant from the start (two-ish decades ago), because (from the client’s or user’s point of view) it was always easier to extract the information with some basic natural language processing (or just simple DOM scraping) than it was to persuade people to add markup that wasn’t necessary for the purpose of the web page. Microformats were based on the idea that there might be a smooth scale from appearance-oriented XHTML to XML APIs rendered with XSLT, and they thought a popular intermediate point might be more-informative markup without full-fat RDF. Instead xhtml was replaced by html5 (because xhtml was too brittle) and xml was replaced with json, and data-rich web pages are now developed with a completely different architecture based on javascript front-ends querying json APIs.
I checked my link log which reminded me that in 2009 I saved a link to a warning about the real cost of microformats.
I would be interested to hear of cases where the semantic web and linked data have been successful, ie, helped to produce simpler implementations or more straightforward interoperability.