Eemeli Aro

> I think part of the motivation for this might be to save work for translators, or to provide a level of compression for the translated results. But we must...

I believe we've confirmed that data model errors are not considered recoverable?

This algorithm does not match exactly this part of our current solution: https://github.com/unicode-org/message-format-wg/blob/07f63094e33b0a8dbe7ab55ef2384c2f6ab0bfa5/spec/formatting.md?plain=1#L488-L489 To demonstrate this, consider the message ``` .match {1 :number} {1 :number} * 1 {{star + exact}}...

> I don't agree with your reasoning. `star + exact` is not a better match than `exact + cat`. You can see this if you switch the order of the...

For any annotated expression being resolved (be its value used for formatting, selection, or as an operand or option value), _some_ function needs to be provided that can return a...

That looks fine, except that we really should not require passing the full formatting context to functions, given how it includes all of the raw input variable values and access...

That's pretty much what I ended up implementing; the [MessageFunctionContext](https://github.com/messageformat/messageformat/blob/main/packages/mf2-messageformat/src/data-model/function-context.ts) is a very restricted subset of the full [Context](https://github.com/messageformat/messageformat/blob/main/packages/mf2-messageformat/src/format-context.ts).

Probably the right prominent place for this is the formatting intro, where we need to replace this MAY with MUST: https://github.com/unicode-org/message-format-wg/blob/38fdd698e425e165047ec7d3a53e03459a8d5cb5/spec/formatting.md?plain=1#L10-L11

https://github.com/unicode-org/message-format-wg/blob/e76196481b23e6e9245923a1239282e19484efd0/spec/formatting.md?plain=1#L175-L178

@mihnita How would you represent the operand of this expression in the data model? ``` { 1.00 :x:number } ```