There were a couple of threads talking about a general purpose AST, here's a short paper on that: https://cs.wellesley.edu//~blocksplus/2018/pdfs/blocksplus18-paper21.pdf "Building an Accessible Block Environment Multi-Language, Fully-Accessible AST-based Editing in the Browser" (maybe I got the paper from here, not sure, I found the tab open 😄)
👍 2
s
Scott Anderson
02/08/2019, 11:32 PM
The UncleGoose toolkit sounds cool but I can't find any reference to it outside of that paper
Scott Anderson
02/08/2019, 11:33 PM
They named it directly after popular wooden blocks with symbols which is clever, but also unfortunate for search
b
Brian Hempel
02/11/2019, 11:42 PM
@jkoppel had a paper at SPLASH on an approach to writing multi-language analyses and transforms. Not quite the same deal as structured editing, but presents a nifty approach for dealing with scenarios where a particular language doesn’t match the universal AST. Might be able to adapt some ideas for a structured editor scenario. https://arxiv.org/pdf/1707.04600.pdf
👍 1
j
jkoppel
02/12/2019, 1:12 AM
Thanks Brian!
jkoppel
02/12/2019, 1:13 AM
Universal ASTs are my favorite idea to hate. They sound great, but they’re actually a terrible idea for most PL tasks
jkoppel
02/12/2019, 1:14 AM
(Structured editing may be an exception; you can basically treat every node identically)
jkoppel
02/12/2019, 1:14 AM
A long time ago I wrote half a blog post on why. It actually has a fair bit of my argument written down