I do think it is a bit disingenuous that the author compares this to a classic web application and its data needs, when the needs here are so trivial. It is very well designed for what it does (the security needs of such a checker almost dictates the design), but is a model usable by very few applications.
As an aside, the article talks a bit about Brotli and it's worth noting that Brotli is nothing more than LZ77 with a dictionary pre-seeded with a 119KB static dictionary of commonly seen web text. It is of course going to be fantastic for compressing an HTML document, where much of the content is verbose and common, but would do nothing above gzip for the hash result data. I would be surprised if it yielded a single byte of savings in that case. Brotli is supported by most browsers as it was snuck into the WOFF 2.0 standard, so browsers that support the new web font standard automatically have to support Brotli.
As an aside, the article talks a bit about Brotli and it's worth noting that Brotli is nothing more than LZ77 with a dictionary pre-seeded with a 119KB static dictionary of commonly seen web text. It is of course going to be fantastic for compressing an HTML document, where much of the content is verbose and common, but would do nothing above gzip for the hash result data. I would be surprised if it yielded a single byte of savings in that case. Brotli is supported by most browsers as it was snuck into the WOFF 2.0 standard, so browsers that support the new web font standard automatically have to support Brotli.
https://dennisforbes.ca/index.php/2016/01/28/eat-your-brotli...