[MLton] should Date.{fromString,scan} allow space instead of zero?

[Daniel C. Wang]

The one way you can encourage any standard to evolve is to get a defacto set 
of common extensions agreed on. I think MLton should do its part by making 
things better when things seem broken. Consider it a protest vote on details 
that are gotten "wrong".  By turning the flag on by default we immediately 
get good feedback on if the changes are a good idea or not.

If people complain their code doesn't work the way they expect in MLton we 
can consider that as evidence the "improvement" wasn't one. If we get 
compaints that their code doesn't work as they expect in another ML 
environment that is conforming to the spec we can knock up a point against 
the spec. :)

Anyway, maybe this is moot, as perhpas this date issue is the only place 
where there is an obvious need for improvement... but if this becomes a 
trend.. I'd prefer the improvments be on by default.

Henry Cejtin wrote:
> I am arguing that the default has to be to follow the current official spec.
> If you want a flag to indicate a `better' (and I agree it would be better)
> semantics for something, then you have to explicitly take action to get it.