This is a dedicated test post for measuring the performance impact of different comment systems on our Hugo site.

Testing Methodology#

We’re comparing four popular comment systems:

  1. Disqus - Popular third-party solution with cloud hosting
  2. Giscus - GitHub Discussions-based, open source
  3. Cusdis - Lightweight, privacy-focused solution
  4. Remark42 - Self-hosted, privacy-focused with rich features

Performance Metrics#

For each system, we measure:

  • Page Load Time - Total time to fully load the page
  • JavaScript Size - Amount of JS downloaded for comments
  • HTTP Requests - Number of network requests made
  • PageSpeed Score - Desktop and Mobile performance scores
  • Core Web Vitals - FCP (First Contentful Paint) and LCP (Largest Contentful Paint)

Test Environment#

  • Hugo Version: v0.152.2 extended
  • Theme: Terminal v4.2.3
  • Browser: Chrome/Brave (latest)
  • Network: Real-world conditions via PageSpeed Insights
  • Cache: Hard reload (Ctrl+Shift+R) for accurate measurements

Baseline Measurement#

Before testing any comment system, we capture a baseline measurement of the page without comments to understand the base performance.

Why This Matters#

Comment systems can significantly impact:

  • Page load speed (critical for SEO and UX)
  • Mobile performance (often overlooked)
  • Hosting costs (for self-hosted solutions)
  • Privacy compliance (GDPR considerations)
  • User experience (ease of commenting)

Test Post Content#

This post contains a realistic amount of content including:

  • Multiple headings
  • Paragraphs of text
  • Lists (ordered and unordered)
  • Code blocks (if needed for testing)

The goal is to simulate a real blog post to get accurate performance measurements.

Expected Results#

We expect to see variations in:

  • Load time: 0.5s - 3s difference between systems
  • JS size: 50KB - 700KB range
  • PageSpeed scores: 10-30 point difference

This test post will remain unlisted and accessible only via direct URL for testing purposes.

Notes#

  • Each system will be tested separately by changing the commentSystem parameter in hugo.toml
  • Tests include both Desktop and Mobile measurements (3 runs each)
  • HAR files will be captured for detailed network analysis
  • Screenshots will document the visual appearance of each system

Feel free to leave test comments below when the system is active!