Woot! I did not know service workers had access to canvas yet. It certainly beats passing array buffers and manually implementing drawing functions. Also, I have literally never used canvas as a DOM element, only for game windows. I keep thinking, "Oh, this is the perfect spot to make a shiny particle-effects button with canvas", but then it never fits the aesthetic quite right or I end up killing it to eek out a little more performance. Still, someday...
All these new features are great and awesome like ::part() and ::theme() ;) ... but i still want to have simple fixed `thead` for vertical table scrolling without hacking around. :)
I wonder if low latency canvas could possibly support adaptive sync (FreeSync, G-Sync), so you won't end up with screen tearing when your framerate doesn't exceed monitor refresh rate.
The low latency canvas would be a very welcome feature. I however see no difference with the low latency test ¹ tried both on Windows and Ubuntu. Chrome/Chromium version 72 and 71. With experimental Web Platform features ² set to Enabled. Edit: I did actually notice one difference, that when the canvas is resized, it doesn't reset itself with lowLatency set to true! But that's about i! I also tried with a IO-latency meter ³ and there where no different with or without lowLatency setting. The current canvas IO latency is about 14-55ms, if you could do anything, witchcraft or magic, to get the IO-latency down to 1ms, I would be a very happy person! It would allow having nice things in the browser. 1) chromium.googlesource.com/chromium/src/third_party/+/master/blink/manual_tests/canvas-low-latency.html 2) chrome://flags/#enable-experimental-web-platform-features 3) pavelfatin.com/typometer/
Question: Can we make a push to add SQL back to the spec, us developers want it and the promise that it can just be built easily on top of IndexedDB doesn't seem to be true. Writing PWAs with complex interrelated data is practically impossible without SQL.... how can we go about bullying chrome/firefox/safari/w3c to add this back.... or at least get a WebSQL2 spec in the pipeline so that when WebSQL finally gets removed we've got a replacement.
How I miss this series and these two guys!
Woot! I did not know service workers had access to canvas yet. It certainly beats passing array buffers and manually implementing drawing functions.
Also, I have literally never used canvas as a DOM element, only for game windows. I keep thinking, "Oh, this is the perfect spot to make a shiny particle-effects button with canvas", but then it never fits the aesthetic quite right or I end up killing it to eek out a little more performance. Still, someday...
Sadly 3 years later still no support on safari. One still needs to create a canvas in another to the work off the main thread in safari
This made me happy, thank you.
Super interesting.... love the podcast too...been going back through all the episodes.. " ./ whats this path relative too... your face"
All these new features are great and awesome like ::part() and ::theme() ;) ... but i still want to have simple fixed `thead` for vertical table scrolling without hacking around. :)
How do you like them pixels?
Square
I wonder if low latency canvas could possibly support adaptive sync (FreeSync, G-Sync), so you won't end up with screen tearing when your framerate doesn't exceed monitor refresh rate.
It'd be interesting to support this supported in the web in general. Although, I don't know if it makes sense in multi-window environments.
Awesome
hubba bubba double bubble buffering
How does Surma know these stuff? I hate and I love him at the same time :)
It's his job :D, both these guys sometimes work on new web specifications so they do have to know their stuff really well.
Take 5? Are there bloopers?
In this case I think it was stopping to change the battery in cameras, then we had a problem with the sound etc etc. Not fun bloopers.
The low latency canvas would be a very welcome feature. I however see no difference with the low latency test ¹ tried both on Windows and Ubuntu. Chrome/Chromium version 72 and 71. With experimental Web Platform features ² set to Enabled.
Edit:
I did actually notice one difference, that when the canvas is resized, it doesn't reset itself with lowLatency set to true! But that's about i!
I also tried with a IO-latency meter ³ and there where no different with or without lowLatency setting.
The current canvas IO latency is about 14-55ms, if you could do anything, witchcraft or magic, to get the IO-latency down to 1ms, I would be a very happy person! It would allow having nice things in the browser.
1) chromium.googlesource.com/chromium/src/third_party/+/master/blink/manual_tests/canvas-low-latency.html
2) chrome://flags/#enable-experimental-web-platform-features
3) pavelfatin.com/typometer/
Question: Can we make a push to add SQL back to the spec, us developers want it and the promise that it can just be built easily on top of IndexedDB doesn't seem to be true. Writing PWAs with complex interrelated data is practically impossible without SQL.... how can we go about bullying chrome/firefox/safari/w3c to add this back.... or at least get a WebSQL2 spec in the pipeline so that when WebSQL finally gets removed we've got a replacement.
"Writing PWAs with complex interrelated data is practically impossible without SQL" NOTHING is impossible. Learn GraphQL stop crying.
Bubble buffering