PilBox – Building Mobile Apps in PicoLisp
mail-archive.com
mail-archive.com
It's sad when a clean slate project for a platform created when we should have known better decades ago still ends up with this kind of overhead.
It perfectly matches my recollection of doing Android development too: a big uncontrollable mess of different kinds of files whose inter-dependencies and relations are completely opaque.
The initial build & deploy times for Clojure on Android are annoying, but using a REPL to edit your Android app dynamically is amazing.
Taking a picture, for example, needs only
(android~takePicture (tmp "img")
'((Intent)
(... do something with the image) ) )
The function 'takePicture' is in "lib/android.l": (de takePicture (Dst Fun)
(startActivityForResult Fun
"android.media.action.IMAGE_CAPTURE"
"output" (fileUri Dst) ) )It feels like a no-brainer for them to want straightforward usage with, say, Python. Is it really a performance thing?
I guess the Java bias is very strong among the ex-Sun and ex-Danger OS developers that have joined the team since the early days.
You see this on Google IO presentations from the last couple of years, when the audience asks about alternatives, including existing JVM languages.
And since I expect someone to post Jonathan Schwartz public comment commending them, here goes what Gosling had to say about it.
"We were all really disturbed, even Jonathan just decided to put on a happy face and tried to turn lemons into lemonade, which annoyed a lot of folks at Sun."
https://www.wired.com/2012/04/gosling-slime/
Companies like IBM, Atego, Aicas, MicroEJ, Excelsior, ARM among many others, never had known issues dealing with either Sun or Oracle licenses, while providing their own Java compilation chains and hardware specific APIs.
So trying to use this to build an Android app with the native (i.e. Java) Android UI libraries probably wouldn't work too well, right?