This is probably beyond the scope of the "Newbies" mailing list, but how did Orca handle non-local returns?
On Wed, Aug 3, 2022 at 10:29 AM Yoshiki Ohshima Yoshiki.Ohshima@acm.org wrote:
On Wed, Aug 3, 2022 at 10:20 AM Tobias Pape Das.Linux@gmx.de wrote:
Hi
On 2. Aug 2022, at 19:53, Yoshiki Ohshima yoshiki.ohshima@acm.org
wrote:
Hi
(I am working on a few croquet.io projects.) Trying to run the current
way of how Squeak works and making it "croquetified" has some challenges. One is that a Squeak image typically uses 50MB, 100MB, or more memory, and that means that the "snapshots" that Croquet takes and sends over the network now and then can be quite expensive. (At the same time, as we know, most of the contents in an image is read only; we can imagine to have a different image structure where the read only part is separate and only the new stuff is exchanged.)
Another issue is to ensure determinism. There may not be too many
issues around it, as the old Squeak based Croquet has shown; but the current Morphic itself probably does not work out of the box in this regard.
An alternative approach is to have a different execution engine, or a
translator from Smalltalk to JavaScript. I actually gave a talk recently on that idea (and other things):
https://youtu.be/E3og3l5kKes?t=1353
If you enable English subtitles on YouTube, you can get the gist of
what I was talking about. The implementation of it is available here:
This reminds me of the Orca idea from a few years back:
https://www.hpi.uni-potsdam.de/hirschfeld/projects/orca/index.html
Nice! Even better, It is a C5 paper with familiar names.
-- Yoshiki