Kyle Wanamaker

Correct, it looks like it's all parameters if they into the request parameters section. I haven't tried the raw body. Our OAuth implementation throws away any values it doesn't like,...

This discussion looks to be related to #123.

To me, deprecation seems like a fairly standard part of the software lifecycle. I would suggest a `$deprecated` object rather than a string. With an object, `$deprecated` could then resemble...

> > With an object, $deprecated could then resemble other objects in the spec. > > This I do not really follow :) Why would `com.my-co.lifecycle` have the need to...

> Instead of a boolean flag, $deprecated could just use a string value? The reason for `$deprecated` to be an object is for forward compatibility and for machine readability. One...

@romainmenke said: > My main concern is that regexp will be used or that we will see everyone writing custom parsers for these. I agree with this concern. As written,...

@rdlopes said: > Is it set upfront that durations will always be expressed in milliseconds? From https://tr.designtokens.org/format/#duration: > Represents the length of time in milliseconds an animation or animation cycle...

This comes back to an area where I think the spec is [stringly typed](https://cocoacasts.com/the-danger-of-string-literals-and-stringly-typed-code) when it could be strongly typed. Currently, alias is determined from a leading and trailing `{}`....

In the case of parsing or linting, I think it is much simpler for each token to carry a `$type`. Using `$type` for every token also adds security that an...

Over the weekend I did some thinking. One place I got stuck is how to process `$value` in Java when there is no `$type` to decide how to parse a...