• 5 Posts
  • 14 Comments
Joined 2 years ago
cake
Cake day: June 18th, 2023

help-circle

  • You’ve actually convinced me that this was at least a reasonable idea but I don’t think a gopherhole shop rating system would be in theme with the closely knit community dynamic we have.

    The protocol was going to have a feature-rich, shopping-friendly set of extensions, but when this was proposed in 1993, there was a community backlash against commercialising the gopher; read Albert Rosetti’s booklet about what happened. This is roughly how the gopher protocol became both liberated and hyperstable in the early 90s.

    Imagine if a gopher wanted to sell or otherwise pass on some old computer hardware. It makes sense to share that on their gopherhole: And since payment handling is ~ always legally middle-manned by a payment handler who will handle the payment anyway, the protocol wouldn’t need to be extended to be a payment processer.

    On the other hand, I would be against paving paradise to put up a parking lot.











  • @rwxrwxrwx@mathstodon.xyz Juan M. Bello-Rivas https://sr.ht/~jmbr/cl-buchberger/ ; Common Lisp Object System based polynomial-ring math

    • great use of defmethod to handle different useful notions of * / + - , and very powerful LOOP facility.
    • wonderful to have this common lisp native math package
    • (implementing buchberger’s algorithm to compute a polynomial’s Groebner basis, which exposes many of the polynomials properties obviously)
    • package-wise… The native package is purely native, and uses the old fashioned packaging idiom (package.lisp) in a source module, but
      • the test package uses #:fiasco which extends uiop:define-package and (secretly, I guess) uses/implies the :package-inferred-system idiom for tests.
      • So because Juan is also using #:fiasco I think both sides could benefit from asdf’s :package-inferred-system, though this would mean lots of uiop:define-package :mix ing of separate files, and the usefulness of :module is subsumed into :package-inferred-system as well.

    Carlos Ungil’s usual-suspects deep learning FFIS https://gitlab.common-lisp.net/users/cungil/projects

    • Carlos Ungil has done exactly the greatest thing.
    • (Though once again with old fashioned packaging)
    • Depends on lisp’s common cffi package
    • Since R libraries are fundamentally actually C libraries exposed in the R scripting language, Carlos has simply captured the underlying C libraries and built them into lisp using #:cffi , and done this for a variety of C libraries including both vendor-ish ones like tensorflow, and R’s underlying Rmath package.