4 ms·
Do you agree with them that "implementations which really compile Python have to be able to handle some thread suddenly modifying something out from under anoth
by igouy 2y ago
Do you agree with them that "implementations which really compile Python have to be able to handle some thread suddenly modifying something out from under another thread"?
Would you say that's also true of Smalltalk implementations?
- pjmlp 2y agoAs if there aren't Smalltalk and Lisp implementations with support for parallelism, and have naturally to handle such cases.
- igouy 2y agoBy "support for parallelism" do you mean invoke multiple VMs and use sockets or shared memory?
- pjmlp 2y agoYou know what I mean, no need to play lawyer, leading the witness.
- igouy 2y agoNo. As-far-as I can see those comments are unhelpful to the point of being evasive.
- pjmlp 2y agohttps://franz.com/support/documentation/10.1/doc/smp.htm https://franz.com/support/documentation/10.1/doc/smp.htm https://www.lispworks.com/documentation/lw80/lw/lw-multiprocessing-ug.htm https://www.lispworks.com/documentation/lw80/lw/lw-multiproc... https://www.cincomsmalltalk.com/main/products/foundation/matrix/ https://www.cincomsmalltalk.com/main/products/foundation/mat...
- igouy 2y agoAnd when we look at that (proprietary Cincom Smalltalk) approach we find "simple mechanisms for spawning multiple running copies of the image and using those to perform various tasks in parallel" and "Objects are marshaled between the master and drone as copies, not as references." That's not an example of comparable complexity to "… Python ha[s] to be able to handle some thread suddenly modifying something out from under another thread".