-
Using the new client-side JavaScript rendered Twitter and thinking about @danwrong and co’s stellar 2012 Twitter redesign: blog.twitter.com/engineering/en_us/a/2012/improving-performance-on-twittercom.html I’m curious how the time to first tweet compares in this new version.Permalink On twitter.com
♻️ 6 Retweets ❤️ 25 Favorites Mood +1 🙂
-
@danwrong Okay, I’ve measured it. Old “boring” server rendered Twitter (Fast 3G, empty cache): Twitter stream render: 4.13s Trends: 9.03s Who to Follow: 9.06s Finish: 11.17sPermalink On twitter.com
-
New client-side rendered Twitter (Fast 3G, empty cache): Twitter stream render: 7.21s (+74%) Trends: 7.25s (-20%) Who to Follow: 7.06s (-22%) Finish: 10.27s (-8%)Permalink On twitter.com
-
New Twitter definitely feels slower, despite finishing sooner. That stream render difference is pretty huge. It strikes me as a good thing that Old Twitter lazy loaded Trends & Who to Follow. I wonder if New Twitter’s stream render time could improve by doing something similar?Permalink On twitter.com
-
The other striking thing is that New Twitter has a more unstable load, a trait seemingly linked with client-rendering—4 different stages to load the main content: 1. Logo screen 2. Skeleton page with Masthead 3. Skeleton page with Stream container 4. Final stream content loadedOn twitter.com
-
All this said, I quite like the usability improvements in this new version (did they ditch all the awkward modals? 🎉). If nothing else, a bit slower page load might help keep me from using the site to such an unhealthy degree 😅Permalink On twitter.com
-
Switched back to Legacy Twitter today after two weeks really trying out New Twitter and putting it through the wringer. Have y’all switched to New Twitter? Did it stick for you?Permalink On twitter.com