Previously, Co-dfns could only support ranks <= 4. Now, the architecture supports arbitrarily large array sizes, though this is only partially used in various primitives. See the release notes for more details. However, this is a bigger deal than it might seem, as it opens up a number of new features for implementation that were previously not possible, and paves the way for reducing the size of the runtime greatly, and reducing the C/C++ footprint of the whole system to a bare minimum.
You might also like...
Speaking at LambdaConf 2024 (May) [Limited Time Discount code]
I'll be speaking at LambdaConf 2024 in May. We haven't finalized the exact talks that I'll be giving (I submitted
Suggestivity and Idioms in APL
Two major principles emerge in APL programming, suggestivity and idioms. I recently encountered a nice illustration of how these two
12 min read
Flat Program Organization: An Email Response
A fellow APLer, Tort, started an interesting email discussion with me, and one of his points was good enough that
7 min read
Thesis Correction: Sec. 3.7, Guard Lifting
I was thankfully informed by Douglas Mennella that the before and after diagrams for section 3.7 of my thesis
1 min read
2023 APL Problem Solving Competition
The 2023 APL Problem Solving Competition is now open! Solve problems in APL for the chance to win cash prizes
1 min read