Shih-Chiang Chien

In order to run WebSocket over QUIC, integrating WebSocket with HTTP2 is essential from [this discussion](https://groups.google.com/a/chromium.org/forum/#!topic/proto-quic/WIBN7HqHIY0). An early [proposal](https://lists.w3.org/Archives/Public/ietf-http-wg/2016JulSep/0615.html) has been posted to httpwg but no further discussion.

Looks like there is more progress on the IETF side. The approach of WebSocket over HTTP/2 is likely to embed WebSocket frame inside a HTTP2 frame, instead of steal the...

According to [this proposal](https://docs.google.com/presentation/d/e/2PACX-1vSk0NcS8COhQ3vX3KdzEdufy72gaWONQeybRsFF843RheCusbUzj0l5XRmxiePiqTRKlE4SQbUpK_o8/pub?start=false&loop=false&delayms=3000#slide=id.g2a6f3cad35_12_69), additional signaling channel is still required to establish QUARTC.

Sure, I plan to convert the [draft protocol](https://wiki.mozilla.org/WebAPI/PresentationAPI:Protocol_Draft#Protocol_Details) we used in our previous experimental implementation into a proper format for CG.

Exchanging SDP and ICE candidates are the missing part for bootstrapping RTCDataChannel. https://wiki.mozilla.org/WebAPI/PresentationAPI:Protocol_Draft#Establish_Data_Channel is the minimal protocol we need. Three message types to be defined: 1. send-offer: contains the full...

This Akamai's [blogpost](https://blogs.akamai.com/2017/10/i-can-haz-tls-13.html) has pretty good summary of the current status of TLS 1.3 deployment. The failure rate caused by middlebox is still too high. More information would be available...

有hotkey的確能增加方便性，但從今天下午我自己輸入的經驗來看，一張圖片載入後可能產生三種動作：1. 輸入, 2. 回報空白, 3. 驗證。當我連續完成數個輸入後若突然來個空白圖或驗證，就會讓我卡住一下，如果能讓參與者固定只進行一項動作，應該能提升整體效率。 純白空格的圖片的確不難做，不過有些空白圖上是有浮水印的，不確定是否能純靠程式辨認。