Jason Eisner

@nwf The above is a cute design, thanks, and I think we should probably go with some version of it. It is a more concise version of the following destructuring...

> > > [from @timvieira] Will my version of argmax be enough for you lecture tomorrow? > > > > [from @jeisner] What's the current version of argmax? > >...

Awesome. What's your thinking about names? @nwf, feel free to chime in.

A proposal: It seems to me that the pair is best described as a [val,key] pair, so we could use $key instead of argm. I wondered what y'all thought about...

Thanks for implementing! I now have a proposed change to the syntax that might be workable. > don't forget to quote the argument Hmm, I was assuming that `$key` autoquoted...

Thanks for implementing this: I'll put it (and `:-`) on the parsing assignment! I agree with closing this ticket. But what do you mean by "head destructuring will eventually replace...

But how would you even write this aggregation without the `with_key` keyword? ``` g max= f(X,Y) with_key X. ``` or for that matter ``` g with_key k max= f(X,Y) with_key...

> It's not a "keyword" it's a term 'with_key'(f(X,Y), X). Right, it's syntactically an infix functor. By "keyword" I meant that it is hard-coded to have a special meaning or...

> Sorry, I hit something that made that comment post before I was done with it. Let me fix it before you read it. Ok, you can read it now....

Note on where this design is going: Remember that the eventual design is that a .dyna file can contain both rules and scripting commands. It defines a dynabase together with...