"The New Yorker The average load time is 102.127 seconds 431 requests" Author is asking for "text-only"; this of course only requires one request. curl -

"The New Yorker The average load time is 102.127 seconds 431 requests" Author is asking for "text-only"; this of course only requires one request. curl -4o https://www.newyorker.com/magazine/2015/05/25/high-gear-current-cinema-anthony-lane | tr -cd '\11\12\15\40-\176' | sed 's/</\ &/g;s// /g;s/ *//;' #tab |sed -n '/./{/sectionBreak/,/<\/div>/p;/figcaption/,/<\/figcaption/p;}' > a.htm firefox a.htm We do not need a complex, graphical browser that autoloads resources and runs Javascript in order to request or display a page of text-only content. "Reader Mode" is great but it is not available for every website. The time spent by the browser on those 431 requests is significant. What if a simpler client was used, e.g., time printf "GET /magazine/2015/05/25/high-gear-current-cinema-anthony-lane HTTP/1.1\r\nHost: www.newyorker.com\r\nConnection: close\r\n\r\n" |openssl s_client -connect www.newyorker.com:443 -ign_eof > /dev/null curl -4o /dev/null https://www.newyorker.com/magazine/2015/05/25/high-gear-current-cinema-anthony-lane time tnftp -4o /dev/null https://www.newyorker.com/magazine/2015/05/25/high-gear-current-cinema-anthony-lane time links -dump https://www.newyorker.com/magazine/2015/05/25/high-gear-current-cinema-anthony-lane > /dev/null Guess how many seconds this one request takes when we do not use a popular graphical browser to make it