More importantly, this isn't fixable by versioning. A new version of JSON couldn't make JS/Racket/Java have the same number system.
-
-
Replying to @samth
This works fine in protobuf, thrift, cap'n proto, flatbuffers, etc. because they're precise about their types.
2 replies 0 retweets 1 like -
Replying to @danluu
The point of JSON is to map to the "native" types in each language. JSON's approach has drawbacks, but versioning is irrelevant.
1 reply 0 retweets 0 likes -
Replying to @samth
If that's the point, the original spec is quite odd in specifying, multiple times, "very much like a C or Java". Why say that?
2 replies 0 retweets 1 like -
also odd that it says that it's based on of ECMA-262 and only has one type of number. Relatively few languages only have one number
1 reply 0 retweets 0 likes -
IMO it seems more like someone who writes js created a thing that's convenient in js and falls over if you want (for ex) int and fp
2 replies 0 retweets 0 likes -
if it were versioned like websockets someone could tell me in the protocol they're using extension XYZ and pass me an int
1 reply 0 retweets 0 likes -
instead they tell me verbally that vars A, B, and C, which are passed as strings, should instead be parsed as ints
1 reply 0 retweets 0 likes -
I want to use a serialization format to avoid having to do exactly that.
1 reply 0 retweets 1 like -
Replying to @danluu
JSON, like XML, is half of a serialization format in the sense you want (which is why it's successful, IMO).
1 reply 0 retweets 1 like
Well, I agree there. Most successful things (PHP, Mongo) are not at all what I want :-). They're low friction, but have sharp edges
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.