And Comments that are FAQS -- particularly PyPy and Cython.
Here are some more comments.
I want to clear the air for posts on parsing.
None of these posts are particularly actionable. I may refer to some of these concepts later in the blog.
Parsing Techniques You Will See In Practice -- 3 styles of parsing you will see in practice
Review of the "Science of insecurity" / langsec [langseg.org] -- CFGs are the wrong abstraction
Parsing C -- lexer hack, etc.
This one is probably the most important link.
What Oil is doing is not new: Converge and Mython. Not to mention PyPy!
Follow up on type checking vs. metaprogramming
Followup on oheap:
[python3] -- why I didn't use it
Go/Erlang with Security
Also: weren't there ones about bazel and config languages?
Languages Should Focus on Compatibility
Joe Duffy talk -- did Rust focus enough on compatibility?
Nobody ever rewrites anything -- Go. However Oil does do automatic translation
C is two languages -- shell is also two languages
TensorFlow is a Progrmaming Language and has Language Problems
TensorFlow Model -- related to Oil. A strict computational model.
translation units, type information
Python structs vs. slots -- don't have time for this right now
Kinds of metaprogramming, why compile-time compuing
Data Frame Use Cases: https://news.ycombinator.com/item?id=15595181
I argue you don't need distributed data frames. R Dialect.
I am spending a lot of time on Oil, so I want it to be adopted. I'm scratching my own itch, but it's more interesting if other people's itches are also scratched!
C++ Viral, COM
Programming Languages Shouldn't Nag You (with respect to GC and safety) -- language
- C doesn't nag you, but it allows your program to run wild - Python and Java don't nag you, but they impose large runtime costs - Rust and Haskell nag you
TODO: The problem of killing untrusted processes ?
You should learn R, because it passes the Perlis test? (lobste.rs)