[BUG]: time-to-read userscript activates on non-article websites

[bittricky]

Userscript

• time-to-read.js

Problem:

The time-to-read userscript currently activates on many websites that don't contain article content or long-form text. This causes unnecessary script execution and occasionally displays reading time indicators on inappropriate pages.

Potential Causes:

After looking throught the code, I've found several factors contributing to this issue:

1. Overly Broad URL Matching:

   • Using @match *://*/* in the userscript metadata means the script loads on every website
   • While the script does have exclusion logic, it only blocks specific URL patterns rather than identifying article sites

2. Generic Content Selectors:

   • Fallback selectors like main, .content, and generic article tags match elements on many non-article websites
   • The "catch-all" pattern { domain: "*", contentSelector: "main" } is particularly problematic as nearly all websites have a <main> element

3. Liberal Text Block Detection:

   • The findLargestTextBlock() function selects any block with sufficient text density
   • Many sites with long navigation menus, documentation, or code examples can trigger this detection

4. Minimal Word Count Threshold:

   • Current MIN_WORD_COUNT of 100 is too low for reliably identifying article content
   • Many non-article pages (product listings, documentation indexes) can exceed this threshold

Steps to reproduce:

1. Visit sites like GitHub repository pages, documentation sites, or e-commerce product listings
2. Observe that the reading time indicator appears despite no article content being present
3. Specific examples:
   • Shopping cart pages on e-commerce sites
   • GitHub repository main pages
   • API documentation home pages
   • Social media feed pages

Proposed Solutions:

1. Improved Site Detection:

   • Implement a more robust content detection algorithm
   • Consider checking meta tags (e.g., <meta property="og:type" content="article">)
   • Look for article schema markup (itemtype="http://schema.org/Article")

2. More Specific URL Matching:

   • Limit script execution to known content sites (news, blogs)
   • Add more excluded URL patterns for common non-article paths

3. Better Content Heuristics:

   • Assess text-to-HTML ratio in candidate content blocks
   • Check for article-specific patterns (date published, author byline, etc.)
   • Increase MIN_WORD_COUNT to a more selective threshold (e.g., 250-300)

4. User Configuration:

   • Add a whitelist feature where users can specify which sites should run the script
   • Implement a "training mode" where users can manually select article blocks