9 ms·
Having developed HTTP/2, SPDY, and HTTP/1.1 framers, I can tell you that HTTP/1.1 is by far more error prone. Length encoding (i.e. binary format) is a god-send
by dcsommer 12y ago
Having developed HTTP/2, SPDY, and HTTP/1.1 framers, I can tell you that HTTP/1.1 is by far more error prone. Length encoding (i.e. binary format) is a god-send for reducing bugs.
I'm curious what you think about my analysis of the binary vs text argument: https://www.youtube.com/watch?v=-yxQIRl6Qic&feature=player_detailpage#t=469 https://www.youtube.com/watch?v=-yxQIRl6Qic&feature=player_d...
- Dylan16807 12y agoWell I was going to give an example of a non-binary format with length prefixing but I've ended up entirely unsure what the definition of 'binary format' is. Instead I'll say that length prefixing is good and useful but doesn't necessarily provide a correctness advantage over simple delimiters. It doesn't really shine until you've already decided to allow unrestricted binary data inside of strings.
- captaincrowbar 12y agoOut there in the real world, people are going to keep putting unrestricted binary data inside of strings regardless of whether or not anyone has decided to allow it.
- Dylan16807 12y agoIf you delimit on \0, period, no exceptions or escaping, then I assure you no one will put \0 inside their strings. Same for other trivial schemes.