Implement Marpa's Scanless Interface (SLIF) for writing grammars

[iarna]

I'm gonna give a go at this. It's documented here:

https://metacpan.org/pod/distribution/Marpa-R2/pod/Scanless/DSL.pod

[rns]

Glad to hear that! Most likely you know it already, but just in case: Marpa IRC channel is a good place to look for help/advice re SLIF.

[aredridel]

Ooo!

[jeffreykegler]

I'm interested in helping with advice.

Some questions: lotsawa currently does not implement Leo, does it?

Also, is there a test suite? @rns and I created one for Libmarpa. Getting lotsawa to pass some sort of test suite is IMHO a necessary first step.

[rns]

Some test cases can also be extracted from Marpa::R2 test suite.

[aredridel]

It does implement Leo, though to make it efficient, I'm going to have to binarize all the rules.

[iarna]

There is a test suite, although it's still necessarily pretty basic as lotsawa does not yet have any mechanism for getting information about a parse– it only reports success, ambiguity or failure.

[jeffreykegler]

There's no evaluation phase, you're saying?

[aredridel]

Correct -- getting tree generation and evaluation going will happen at the same time.

[jeffreykegler]

That's tied in with how much Leo you want to implement. Hopefully my new
paper will be helpful.

[aredridel]

That's part of what I was passively hoping would exist by the time I got there. but I shall dive in a bit!