Matt Godbolt and Ben Rady Talking About Data Compression


Matt and Ben talk about how compression works, specifically deflate, which is apparently everywhere. Ben gets particular about compression ratios. Matt explains how to compress /dev/random by sorting it first. See the transcript here: https://www.twoscomplement.org/podcast/compression.txt

At 37:20 Matt talks about domain-specific transforms, and at 40:40 he mentions some kind of supercombinator type compression scheme. See Total Functional Programming. Sadly David Turner died last 19th October. He really knew how to program!

Here is an idea about formalising the notion of top-down context that might be interesting. I wrote it in 2009, just after I moved to Bolivia. It is only one and a half pages long: On Tarski's Semantic Notion of Truth. 

See also this post: Freya Holmér - Why Can't You Multiply Vectors? and I didn't go for a walk today, but I watched some of this video from 27th October:



And it's funny, the other day I was thinking about the  Universidad Mayor de San Andrés, as you do when you think about earthquakes, ... oops! http://www.informatica.edu.bo/ and I just learned that the University of St Andrews is older than the one at Cambridge. A taught one-year Computer Science MSc course at St. Andrews will cost around £18,000 (BOB 180,000) for self-catering accommodation and teaching fees, for UK students. Foreign students will need around.£35,000 (BOB 350,000) for the year, plus travel and living expenses.

Comments

Post a Comment

Popular posts from this blog

David Hestenes - Tutorial on Geometric Calculus

Image
This is a very high-level overview, and not too technical, for the first fifteen minutes at least. It's about the relationship between mathematics and physics and development of scientific ideas. At 40:11 he shows how Structure and Interpretation of Classical Mechanics should have been written! See Lasenby, Doran & Gull (1998[2004])  Gravity, Gauge Theories and Geometric Algebra   and  David Hestenes - New foundations for classical mechanics . Subscribe to  Nomen Nominandum . GAME2020 - 1. Dr. Leo Dorst. Get Real! (new audio!) See  https://www.jeremyong.com/klein/  in this context SSE stands for  Streaming SIMD Extensions . See also  https://enkimute.github.io/ganja.js . Subscribe to BiVector .
Read more

David Turner Obituary by Sarah Nicholas Fri 24 Nov 2023

Image
This was in The Guardian: David Turner obituary , only eight days after I posted  David Turner Talking About Sixty Years of Functional Programming History : This talk was given in London in 2017: See  Turner, D. A. "Some History of Functional Programming Languages" also  John Hughes - Why Functional Programming Matters  and David MacQueen's talk at ICFP 2015 in  Numberphile - Sophie Maclean on the Catalan Numbers . At 10:30 This whole discussion abut combinator reduction is especially interesting. I didn't know Arthur Norman had tried building hardware for combinator machines. I'll look that up: maybe start here A.C. Norman  Faster combinator reduction using stock hardware  in LFP '88: Proceedings of the 1988 ACM conference on LISP and functional programming. At 26:25 on the ISWIM virtual machine implemented in the PAL compacting garbage collector?! This work Reynolds and others did was at MIT in Masecheusetts and Argonne National Laboratory Illinois. Did
Read more

Modeling Probability Distributions and Solving Differential Equations

Image
See  Spread Polynomials : 27:48 On factorization of spread-polynomials, see  Lewis Carroll - Crocodile Story Subscribe to Norman Wildberger . JPL Orbits & Ephemerides: https://ssd.jpl.nasa.gov/orbits.html See also this very impressive student project, which is a digital orrery that uses the JPL Ephemerides: https://www.robots.ox.ac.uk/~mjc/Papers/3YP_Report.pdf  This seems to be the basis for a way to make a space-flight simulator for planets. See  Trying to Make a Sci-Fi Movie  for the context. See also  Laplace Transform : Algebraic construction The Laplace transform can be alternatively defined in a purely algebraic manner by applying a field of fractions construction to the convolution ring of functions on the positive half-line. The resulting space of abstract operators is exactly equivalent to Laplace space, but in this construction the forward and reverse transforms never need to be explicitly defined (avoiding the related difficulties with proving convergence).  Mikusiński
Read more