Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited


This user doesn't have any awards

About HTML5rocks

  • Title

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. HTML5rocks

    Edge is... actually GOOD now!?

    Since as far as I see from your past videos, you care a lot about making good informative and technically correct content, I would like to bring to your attention two potential errors in the script for this video. First of all, I have some doubts about the "HTML5Test (USING HTTP)" graph scores presented in the video at 7:16. Based on my own tests, I believe you might have accidentally tested Firefox and Chrome over HTTPS when you meant to test over HTTP. Disclaimer: I'm not advocating for either browser, I just try to make sense of graph presented in the video. Also, I don't know which browser versions you used and I can't test Chromium-based Edge 74.* that I think you used (I don't have it). I do not know which browser versions you tested so I tested the recently released browsers as well as (Firefox 67.0 64-bit stable, released only yesterday; Chrome 74.0.3729.157 64-bit stable; Chrome 76.0.3800.0 64-bit dev; "old" Edge 18; new Chromium Edge dev) without any special modifications. I believe you tested Chromium-based Edge 74.*, but I was not able to test it because my installation was already updated to 76.*. Also, I tested Chromium 76.* dev just to make sure the scores are the same with 74.* The browser versions are important (as well as settings, flags, etc.), but I do not know your testing strategy and do not think it would produce such a drastic score difference (the scores are pretty stable over time unless you change flags). The "HTML5Test (USING HTTP)" graph shows: Firefox - 513 <---- according to my tests, this should be 491 Old Edge - 492 Chromium Edge - 481 Chrome - 535 <---- according to my tests, this should be 481 The HTML5Test (USING HTTPS)" graph shows: Firefox - 513 Old Edge - 492 Chromium Edge - 535 Chrome - 535 That is surprising to me because: (1) I don't understand why there would be such a gap between the new Edge and Chromium (they are effectively the same thing for these tests) and (2) HTTP and HTTPS scores in Forefox and Chrome/Chromium should be different (due to security considerations). A few tests I conducted: Firefox 67.0 over HTTP - 491 (mismatch) Firefox 67.0 over HTTPS - 513 (matches your Firefox over HTTP and HTTPS) "old" Edge 18 over HTTP - 492 (matches your result) "old" Edge 18 over HTTPS - 492 (matches your result) Chromium Edge 74.* dev over HTTP - unable to test, but should be the same as Chrome 76.*, 481, then matches your result Chromium Edge 74.* dev over HTTPS - unable to test, but should be the same as Chrome 76.*, 535, then matches your result Chromium Edge dev over HTTP - 481 (matches your result) Chromium Edge dev over HTTPS - 535 (matches your result) Chrome 74.0.3729.157 over HTTP - 481 (mismatch) Chrome 74.0.3729.157 over HTTPS - 535 (matches your Chrome over HTTP and HTTPS) Chrome 76.0.3800.0 dev over HTTP - 481 (mismatch) Chrome 76.0.3800.0 (Official Build) dev over HTTPS - 535 (matches your Chrome over HTTP and HTTPS) My results match my expectations much better: Chrome 76.* and the new Edge 76.* perform exactly the same (because Microsoft mostly re-skins Chromium and drops a few non-standard components that the site does not test for and Chrome 74.* and 76.* perform exactly the same because the public announcements and release notes do not mention any deprecations (and the site does not test for newly added features) Firefox and Chrome and the new Edge have lower scores over HTTP than over HTTPS, as per Chromium and Firefox policy of deprecating powerful APIs on insecure origins. The score gap between secure and insecure origins is actually supposed to increase as more restrictions take place. Secondly, the video is the video refers to https://html5test.com/ as a "performace test" three times. https://html5test.com/ is a standards compatibility test, not performance test as video suggests three times (and never mentions word "compatibility"). I realize that you probably meant "test performance" as in "test score" and not "computational performance", but an average user probably would not know. I realize that videos are not created to split hairs over language expressions and technical details, however I feel this is an important distinction to mention. Specifically, I refer to these statements at 6:45 on the video, it states: At 7:15 the video states: At 7:28 the video states: From this a viewer unfamiliar with HTML5test.com might infer that the site tests traditional computational performance, not standards compatibility and compliance test performance. Edit: CC @LinusTech @RILEYISMYNAME Edit 2: sentence structure.