For me a HUGE feature is:

"Mirah code has no runtime dependency other than libraries accessed directly from Mirah code."

Wow - Charles has done an amazing job sticking true to that. In fact so much I thought it was a core aim.

I love scala, clojure - but there is significant runtime libraries to include. for most people this isn't an issue (and it isnt' for me most of the time) but Mirah doing what it does without runtime deps, is a refreshing alternative.

Without knowing too much, cant you do that easily enough for the other two? At least scala, you could package scala-library.jar with all the other classes (I would imagine clojure wouldnt be too much extra).

Of course unless you run it through proguard or similar, its gonna me massive.

well it is still 2-3 od MB of weight. I guess you could strip it down easy enough, and package it. But if you are delivering a library - then it has to play nice with other libraries that may also want to use scala - if you really wanted you could jarjar it into its own namespace (!) but that would be pretty extreme.

If you write your app in scala, and unless you are deploying to a phone, it is rarely an issue (it doesn't bother me at all) - but Mirah as a drop in replacement for a java compiler, that is very compelling and something to watch.

Also - the fact that I use the scala libraries heavily means I need very few other deps myself - so for me it is a net win.

Seems like this would be really easy to use for Android development.

Anybody know if this would this work?

AFAIK, it _was_ a core aim.

This actually looks exciting.

Anyone have any idea when it will stop being "experimental"?

You can build "real" apps today. You can also just write Java whenever you feel Mirah is limiting you.

There are other questions one asks with a language dubbed "experimental".

Will it remain stable? Will it remain supported? Will more people write things in it so that it develops a community of users?

I'd love to see the video, but all I can see is embed code.