My idea to significantly speed up web browsers with a HTML “checksum” attribute

4 points

11 years ago

https://code.google.com/p/chromium/issues/detail?id=554981&thanks=554981&ts=1447349693

Downloading numerous file dependencies can slow down the web browsing experience. These include frameworks such as Polymer, Angular, jQuery, jQueryUI which are often huge, comprise tens of separate HTML, JS, and CSS files, but are however duplicated across multiple servers, often out of necessity.

CDNs partially solve this problem but there are still numerous competing CDNs, as well as places (e.g. China) where ones' choice of CDN may not work well.

I propose an HTML attribute "checksum" where one can specify the checksum of a file, e.g.

    <link rel="import" src="/js/polymer/polymer.html" checksum="44194508fca6f3187c5db3602c72fb03bb6055b8">

    <script src="/js/jquery.min.js" checksum="43dc554608df885a59ddeece1598c6ace434d747"></script>

    <img src="/images/some_huge_image.jpg" checksum="cb3c80c5d6f14bcd5a134399253c0566951af38e">

Web browsers should then keep track of the SHA1 checksums of all cached files, and if a file with a matching SHA1 checksum is found, REGARDLESS OF ITS ORIGIN, the file is not requested again from another server.

This would further speed up the web experience as we move forward with things like Angular, Polymer, Webcomponents, and other things that are going to have a huge, huge number of file dependencies.

7 comments